Application of remote sensing image in water environment testing / Aplicação de imagem

de sensoriamento remoto em testes de ambiente de água

Chen ling-xia, Zhang JunliCollege of Tourism and Resources Environment, Xianyang Normal University, xianyang 712000,China

chen1lingxia@163.com

ABSTRACT: Remote sensing image has been widely used in water environment testing. This paper briefly introduces the basic principles and methods of remote sensing image in water environment testing, its application in water environment pollution test and control is selectively analyzed and the tendency of the development is summarized. The aim is to analyze and summarize the application of remote sensing image in the water environment testing, which is of reference significance to the research.

KEYWORDS: Remote sensing image; Water environment; Test; Application;

RESUMO: A imagem de sensoriamento remoto tem sido amplamente utilizada no teste de ambiente aquático. Este artigo apresenta brevemente os princípios e métodos básicos da imagem de sensoriamento remoto em testes de ambiente de água, sua aplicação no teste e controle de poluição do meio ambiente da água é analisada seletivamente e a tendência do desenvolvimento é resumida. O objetivo é analisar e resumir a aplicação da imagem de sensoriamento remoto no teste do ambiente aquático, que é de referência para a pesquisa.

PALAVRAS-CHAVE: imagem de sensoriamento remoto; Ambiente aquático; Teste; Aplicação;

BOLETIM TECNICO DA PETROBRAS

VOL. 61, 2017Guest Editors: Hao WANG, Min DUANCopyright © 2017, PB PublishingISBN: 978-1-948012-00-3

http://www.pbpublishing.com.br

ISSN:0006-6117 EISSN:1676-6385

1. INTRODUCTION

Remote sensing refers to the process of remote detection, identification, and information of the object in the process of non-direct contact with target. The electromagnetic wave, acoustic wave and gravity field can be used in remote sensing, however, it is generally said that the remote sensing is the electromagnetic wave remote sensing of the target information by the electromagnetic wave[1-3]. Remote sensing technology is widely used in military reconnaissance, missile warning, military surveying and mapping, ocean monitoring,

meteorological observation and inter agent detection etc.In civil terms,Remote sensing technology is widely used in the field of earth resources, vegetation classification, land use planning[4-6], crop diseases and insect pests and crop yield survey, environmental pollution monitoring, marine development, seismic monitoring, etc.The general development trend of remote sensing technology is: improve the resolution of remote sensors and the ability of comprehensive utilization of information,advanced remote sensor、information transmission and Treatment equipment to remote sensing system to achieve all-weather

BOLETIM TECNICO DA PETROBRAS 88

work and Real time information acquisition[7.8],then enhancing the anti-jamming ability of remote sensing system. Because all the objects can be emitted, reflected or absorbed by the absolute zero temperature, and different objects have different properties, the radiation characteristics of different objects are unique. In the water environment testing, the radiation characteristics of different temperature[9-11], sediment concentration, algae quantity and pollution degree of water are different. Under normal circumstances, the characteristics of different water bodies can be reflected by remote sensing images. Spectral characteristics of polluted water body and clean water body are different, the spectral characteristics are reflected in the absorption or reflection of specific wavelength, while the spectral characteristics can be captured by remote sensing image in the same time[12-15]. The water quality parameters or water pollution of the water body can be obtained by the recognition of the remote sensing image. Based on this, remote sensing images can be used in water environment testing[16,17].With the development of social economy, the water pollution of rivers and lakes in our country is becoming more and more serious, which mainly includes the suspended solid pollution, oil pollution, eutrophication pollution and heat pollution and so on[18-20]. The China environmental testing station provides the data shows that the water environment of our country has three main problems: ① The main pollutant discharge amount is more than the water environment capacity; ①The rivers and lakes generally suffer from pollution; ①the ecological water is shortage, water environment is deterioration.According to the present situation of water pollution, water pollution in China is serious, so it is imperative to improve the efficiency of water environment testing[21-23]. The test of water environment in the traditional way is usually ground sampling, and then through the laboratory analysis concluded, the way affected by factors such as natural conditions and space-time, there are some limitations[24,25]. For example, in the process of testing the large area of water, only through the test station and the traditional test method, it can not meet the requirements of real-time, fast, macro and accurate test, and thus cannot fully reflect the situation of the water environment. Compared with the traditional testing method, remote sensing image technology has the characteristics of macro, comprehensive[26-28], dynamic and fast, and can get the information of which cannot be obtained by other test methods. The deterioration of the water environment and the shortage of the traditional test methods will promote the wide application of remote sensing images in the water environment testing[29-32].

2. THE DEVELOPMENT OF REMOTE SENSINGTECHNOLOGY AND THE STATUS IN CHINA

Remote sensing image data can reflect the electromagnetic wave radiation of the surface features in imaging area, and has a clear physical meaning, and the ability of the surface features to reflect and transmit electromagnetic wave is directly influenced by the attribute and state of object itself.The basic principle of remote sensing is based on the analysis of the size and variation rule of the remote sensing image data for analyzing and recognizing the type of surface features. Remote sensing digital image processing as an important means of remote sensing image processing, is the use of computer through the digital processing method to improve and collect the professional information of remote sensing image.Remote sensing is a kind of modern air to ground observation technology, has the advantages of all-weather, multi phases and different space observation scale and so on, it can conduct a comprehensive monitoring for air pollution, water pollution, changes in the ecological environment, urban heat island effect and urban greenbelt, at the same time can provide the required remote sensing image.From the beginning of twentieth Century, Wright brothers invented the first plane in the history of human beings, aerial remote sensing began to be used in military applications and since then, aerial remote sensing has been widely used in the fields of geology, engineering construction, mapping, agricultural land survey and so on. With the application of remote sensing in hydrology and water resources, including the water resources survey, watershed planning, change of water area distribution、runoff estimation、water temperature, water depth, snow cover, soil moisture detection, estuarine coastal zone and shallow sea topography survey, Marine investigation and research. Especially in desolate areas where human footprints are inaccessible. Especially in desolate areas where human footprints are inaccessible. Remote sensing technology can be used as an effective means of hydrology and water resources investigation. In the Second World War, because camouflage technology gradually increased, the color, infrared and spectral band camera technology appeared in remote sensing. Multi spectral imaging technology is the key development of airborne remote sensing, from 1960s, the earliest use of multi-image sensor is based on multi spectral photography, to the multispectral image acquisition by multi camera sensor, the spectral characteristics will be used in the terrain, terrain identification. Satellite remote sensing has been promoted the remote sensing technology to a new stage of comprehensive development and wide application.

89 SEPTEMBER 2017

From the first earth resources satellite launched in 1972, China, the United States, France, Russia, the European Space Agency, Japan, India and other countries have launched a lot of earth observation satellite. At present, the satellite remote sensing technology has been able to fully cover all parts of the atmospheric window, optical remote sensing mainly includes visible, near-infrared and shortwave infrared region, to detect the target’s reflection and scattering, wavelength of thermal infrared remote sensing reaches 14μm, to detect the target’s emissivity and temperature, the wavelength range of microwave remote sensing is from 1mm to 100cm, passive microwave remote sensing is usually used to detect the target’s emission rate and temperature, and active microwave remote sensing is used to detect the reverse scattering characteristics of a synthetic aperture radar detecting target. The microwave remote sensing has realized the earth observation of the whole day. Aerospace remote sensing is one of the main tasks for the imaging and remote sensing, which is not dependent on the ground control and the position of the image target. In aerial photography conditions it can achieve decimeter level precision, in satellite remote sensing conditions, accuracy can reach 5m~10m. The wide application of this technology will change the current photographic survey and remote sensing operation process, so as to complete the real-time mapping and real-time database update. If combined with high precision laser scanner, it can complete real-time 3D measurement. With the development of sensor technology, aerospace technology and data communication technology, modern remote sensing technology has entered a new phase of earth observation data with dynamic, fast, multi-platform, multi temporal and high resolution. At present, the research of our country is mainly including the comparison between the new and the old DOM. The new image and the old digital line graph (DLG) comparison automatically discovery the changes and update database, the most ideal method is used for the combination of three-dimensional reconstruction of the image target and change detection instrument to complete the three-dimensional changes testing and automatic update. Further development of the data processing is achieved through smart sensors, the data processing is achieved on line, and the returned directly is the information, but is not the image data. Remote sensing technology has developed into one of the earth observation technology with multi-platform, multi band, multi resolution and all-weather, and the hot spot is high resolution image acquisition and recognition in the future. Although RS technology is applied alone in many fields, RS is usually used in conjunction with GIS and GPS.The development of remote sensing technology is in the 70’s, and accumulated rich experience in 20 years, has been at the forefront of the world. With the rapid development of

remote sensing technology, remote sensing application has been deep into all areas of national economic construction, containing the following several aspects: to provide the dynamic basic data and scientific decision-making basis for the sustainable and stable development of national economy; to provide timely the accurate assessment test data and maps for major natural disasters in the country; Test, prediction and evaluation of renewable resources; survey and evaluation of geological and mineral resources and large scale engineering; weather and climate forecasting; ocean testing and ocean development. In December 1979, Institute of remote sensing technology application in the Chinese Academy of Sciences was formally established, multi-phase dynamic observation of the urban environment and marine pollution was started, while the forest fires, glaciers, ice floating on the Bay and estuary sediment migration are tested, to analyze the diffusion of urban pollution. In 1998, in the catastrophic flood, remote sensing image is as the important basis for hydrological analysis, predict and responding, has made great contribution to maximally reduce the flood damage. And in the specific policy decided of “west gas east transmission” and other western development, the exploration data of western mining, development capacity and others of remote sensing institute are as reference. At present, a multi-function, multi - purpose application of satellite system is being formed in china.The future of remote sensing technology will move towards the development direction of a variety of sensors, multi-level resolution, multi spectral and multi temporal, so as to provide the ground observation data of economic, real-time, positioning with faster speed, higher accuracy and more information.

Table 1 The main parameTers of fpso

Lpp (m) 310.0

Breadth at waterline (m) 47.17

Draft (m) 18.90

Vertical CoG above keel (m) 13.32

Longitudinal CoG in front of Lpp/2 (m) 6.83

Displacement (kg) 2.3862E+08

Roll radius of gyration (m) 14.77

Pitch radius of gyration (m) 77.47

Yaw radius of gyration (m) 79.30

Turret position (in front of Longitudinal CoG)(m) 84.62

Turret elevation below FPSO keel (m) 1.52

Turret diameter (m) 15.85

Front wind area(m2) 1012

Transverse wind area(m2) 3772

BOLETIM TECNICO DA PETROBRAS 90

3. THE PRINCIPLE AND METHOD OF WATERQUALITY PARAMETERS TESTING ACCORDING TO REMOTE SENSING IMAGEThe principle of water quality parameters by using remote sensing image is mainly that the water body containing a certain material has specific spectral signature and which is difference to the clear water body. The remote sensing spectral signal obtained from the water body is a complex of many signals, which includes the scattering of atmosphere, the reflection of the water surface and the bottom water, and the scattering of multiple factors in the water. The propagation process of the remote sensing spectral signal is shown in the following Figure 1:

Table 2 The sTaTisTic of 6Dof moTions of fpso

Motion Mean Std Max Min

Surge(m) -0.10062E+03 0.10083E+02 -0.68745E+02 -0.13326E+03

Sway(m) 0.32006E-07 0.17342E-06 0.61843E-06 -0.42536E-06

Heave(m) -0.56944E+01 0.11908E+01 -0.13261E+01 -0.10015E+02

Roll(deg) 0.20025E-09 0.88863E-08 0.58196E-07 -0.61182E-07

Pitch(deg) 0.55386E-01 0.12915E+01 0.45927E+01 -0.46059E+01

Yaw(edg) -0.38013E-07 0.17102E-06 0.44091E-06 -0.47798E-06

Lp

Lg

Lw

Lb

Figure 1 . Propagation of remote sensing spectral signal in water environment

Such as suspended solids, algae, chemical substances, dissolved organic matter and other water components, because of the impact of light reflection, absorption and backscattering, which are reflected in the remote sensing image, thus we can conclude the water quality parameters of water body the reflection of them in the image. At the same time, the water quality parameters of the water body Fig. 1 model of FPSO

Fig. 1 shows the model of FPSO in this research.

Fig. 2 The layout configuration of turret mooring system

91 SEPTEMBER 2017

can be inferred from the scattering function of remote sensing spectrum:

1

y1( )m

cb

c

Tf y R

eme∧

=

− =

∑ (1)

The minimum mean square error (MISE) of the kernel mean function Epanechnikow (Pam Nechkov):

0 0( , ) ( , )

yx

i jI x y I i jΣ = =

= Σ Σ (2)

Figure 2 shows the reflection spectra of different types of inland water bodies. It can be seen that the difference in the content of water generates significantly different for reflectivity of certain wavelength range, which is the basis of quantitative measurement of material content.

Figure 2 . Reflectance spectra curve of different water bodies

The water quality is estimated by kernel mean density function:

( ) ( )( )

2 52h

f xhM xd f x

∧

∧=+

(3)

Mooring1

Mooring5

Mooring8

Mooring11Fig.3 the mooring force comparison between our program and

Ariane software

BOLETIM TECNICO DA PETROBRAS 92

Table 3 sTaTisTic of fpso’s 6Dof moTions

Motion Mean Std Max Min

Surge(m) -1.01E+02 9.25E+00 -7.24E+01 -1.30E+02

Sway(m) 2.82E-08 1.41E-07 6.05E-07 -3.32E-07

Heave(m) -5.69E+00 1.18E+00 -9.56E-01 -1.03E+01

Roll(deg) 2.19E-10 7.35E-09 6.17E-08 -4.69E-08

Pitch(deg) 5.61E-02 1.29E+00 5.03E+00 -4.53E+00

Yaw(deg) -3.66E-08 1.33E-07 3.00E-07 -4.15E-07

Table 4 The sTaTisTic of mooring 2sTaTic analysis of mooring 2

F_X(N)

F_Y(N)

F_Z(N)

M_X(N·m)

M_Y(N·m)

M_Z(N·m)

AxialTn(N)

Mean 923861.4 4.5E-06 -1058123 0.007299 77604534 0.042495 1405096

Std 98610.4 0.000131 60939.28 0.03279 4511886 0.190903 110853.1

Max 1292124 0.000374 -888148 0.102888 97731850 0.599015 1819079

Min 666522.4 -0.00037 -1280669 -0.09132 63758422 -0.53164 1110593

Dynamic analysis of mooring 2

F_X(N)

F_Y(N)

F_Z(N)

M_X(N·m)

M_Y(N·m)

M_Z(N·m)

AxialTn(N)

Mean 812026.3 0.000859 -997083 0.012448 74013755 0.072474 1285937

Std 90642.98 0.019399 103311.5 0.287824 7573709 1.67572 137170.8

Max 1051843 0.437108 -805420 6.485114 96949361 37.75656 1651546

Min 650726.1 -0.00275 -1278645 -0.04223 60041487 -0.24585 1035445

Table 5 The sTaTisTic of mooring 5

sTaTic analysis of mooring 5

F_X(N)

F_Y(N)

F_Z(N)

M_X(N·m)

M_Y(N·m)

M_Z(N·m)

AxialTn(N)

Mean 6292.811 782581.1 -969324 11613503 83622505 67614223 1245831

Std 4135.45 15059.07 12421.77 223476.5 1387579 1301088 19124.04

Max 19413.27 838028.2 -927225 12436338 88460944 72404796 1316143

Min -6656.44 731777.6 -1014840 10859580 78487256 63224853 1181226

Dynamic analysis of mooring 5

F_X(N)

F_Y(N)

F_Z(N)

M_X(N·m)

M_Y(N·m)

M_Z(N·m)

AxialTn(N)

Mean 1736.631 779980.6 -966252 11574912 83446018 67389543 1241802

Std 7111.958 71565.83 88317.96 1062037 7721270 6183216 113645.3

Max 40820.68 974443.9 -738368 14460748 1.04E+08 84190982 1549099

Min -26800.2 596673 -1204549 8854627 63903322 51551947 949321.8

The water quality parameters can be divided into the following 4 categories: turbidity, phytoplankton, dissolved organic matter, chemical water quality index. Three methods: theoretical method, semi empirical method, and empirical method are used, and the following are analyzed in detail.

3.1 Theoretical MethodThis method is based on the theoretical model of the light field in the water. The relationship between the absorption coefficient and backscattering coefficient is determined. Then the reflectivity is measured by using this relation, and which is

93 SEPTEMBER 2017

Table 6 The sTaTisTic of mooring 8sTaTic analysis of mooring 8

F_X(N)

F_Y(N)

F_Z(N)

M_X(N·m)

M_Y(N·m)

M_Z(N·m)

AxialTn(N)

Mean -669075 1.72E-05 -892420 -0.00718 87019072 -0.04179 1115735

Std 73428.29 0.000101 51105.04 0.022889 5266036 0.133262 84930.38

Max -447023 0.000492 -727864 0.053329 1.05E+08 0.310481 1441743

Min -952639 -0.00018 -1082213 -0.10759 70947295 -0.62637 854175

Dynamic analysis of mooring 8

F_X(N)

F_Y(N)

F_Z(N)

M_X(N·m)

M_Y(N·m)

M_Z(N·m) AxialTn(N)

Mean -763986 0.000474 -952906 0.007309 93649022 0.042551 1221521

Std 127094.8 0.010575 117702.9 0.157037 11690709 0.914278 172035.4

Max -562398 0.151539 -699537 2.252462 2.92E+08 13.11391 4272386

Min -3177532 -0.03517 -2855971 -0.52115 69600853 -3.03418 912921.1

Table 7 The sTaTisTic of mooring 11sTaTic analysis of mooring 11

F_X(N)

F_Y(N)

F_Z(N)

M_X(N·m)

M_Y(N·m)

M_Z(N·m)

AxialTn(N)

Mean 6292.811 -782581 -969324 -1.2E+07 83622505 -6.8E+07 1245831

Std 4135.45 15059.07 12421.77 223476.6 1387579 1301088 19124.04

Max 19413.27 -731778 -927225 -1.1E+07 88460944 -6.3E+07 1316143

Min -6656.44 -838028 -1014840 -1.2E+07 78487256 -7.2E+07 1181226

Dynamic analysis of mooring 11

F_X(N)

F_Y(N)

F_Z(N)

M_X(N·m)

M_Y(N·m)

M_Z(N·m)

AxialTn(N)

Mean 1736.63 -779981 -966252 -1.2E+07 83446017 -6.7E+07 1241802

Std 7111.957 71565.8 88317.93 1062037 7721268 6183214 113645.2

Max 40820.66 -596673 -738368 -8854627 1.04E+08 -5.2E+07 1549099

Min -26800.2 -974444 -1204549 -1.4E+07 63903322 -8.4E+07 949321.8

associated with the characteristic absorption coefficient and the backscattering coefficient, to calculate the ratio of the actual absorption coefficient and the backscattering coefficient in the water, finally, the content of the components can be obtained. The model based on the theory of light field is still small, and the model is simplified, which makes the forecasting value cannot meet the accuracy requirements.

3.2 Empirical MethodThis method is a method of using image information to carry out water pollution test, and it is a kind of method which is developed with multi spectral remote sensing data applied in water body test. It is based on the statistical analysis of the correlation between the remote sensing band data and ground

measured data, and the correlation model is obtained by statistical analysis of the optimal band or band combination data and the ground water quality parameters, and then the water quality parameters are made inversion. The defects of this method cannot guarantee the fact correlation between the water quality parameters and remote sensing data, and the accuracy of the model is usually not high and has the special characteristics of time and space. In recent years, some scholars have done some research on this method. Such as Raey and other used SPOT multi spectral image to test and evaluate the pollutants in the coastal waters of egypt. Wang Xuejun and others used remote sensing data and field test data to establish the prediction model of water quality parameters in Taihu. The spatial distribution

BOLETIM TECNICO DA PETROBRAS 94

characteristics of water pollution in Taihu was well reflected. By using the internal average relative reflectivity method, the TM images of the Xiaoqing estuary in the Yellow River Delta were made atmospheric correction to obtain the apparent reflectance, and the apparent reflectance of 1~4 band (R1,R2, R3, R4) was sensitive to different water quality. R2/Rl>1 can be used to distinguish the more suspended sediment area, and R4/R3 can be used as an indicator of organic pollution in the water body. The physical processes of the reflection spectra of the polluted water body were studied by Deng Ruru et al. The function relationship between the reflectance of remote sensing data and the pollutant concentration was established. The comprehensive concentration of pollutants in the Pearl River Estuary was extracted by using the TM satellite data in 2002.

3.3. Semi Empirical MethodAccording to the spectral characteristics of the water body in the airborne imaging spectrometer or in the field, the best band or band combination is selected to estimate the water quality parameters, and then a suitable mathematical method is used to establish the quantitative empirical algorithm for the remote sensing data and water quality parameters. It is the most commonly used method of water body remote sensing since 1990s.Many scholars at home and abroad use this method to test and evaluate the lake and reservoir, and get higher test precision. Shu Xiaozhou uses OMIS-II airborne imaging spectrometer to make the surface water quality remote sensing experiment in area of Taihu Lake, to establish the correlation model between OMIS- 11 band reflectance ratio R (21) / R (18) and algal chlorophyll concentration. Both empirical and semi empirical methods are conducted to properly make statistical analysis of the airborne remote sensing data, the spectral data of the ground water body and the laboratory water quality analysis data, then to invert the water quality parameters. The main factors affecting the accuracy of the algorithm are the band setting and statistical analysis techniques of remote sensing image data. A lot of researchers combine the two methods together. Such as Fu Jiang using domestic CHJ- B type of color infrared film make color infrared aerial photography of water pollution remote sensing study on the Grand Canal in southern Jiangsu, and along the canal it was distributed. Water quality sampling and field water reflectance spectra are measured synchronously and quasi synchronously with aerial photo. Using the method of mathematical statistics, the relationship between the transmission density of the color film and the organic pollution parameters of water quality is described quantitatively. Ji Gengshan, et al. made a comprehensive survey and evaluation of the water pollution in the Grande Canale of South of Jiangsu by using the synchronization test pattern with combination of surface synchronization test and spectral measurements, satellite remote sensing and color infrared remote sensing. Ma Gang preliminary explored the test method for water pollution

test parameters of Daliaohe estuary by using CBFRS-1, TM and field spectral measurement data.

3. APPLICATION OF REMOTE SENSING IMAGE INWATER ENVIRONMENT TEST

Spectral characteristics of water body and its pollutant is the basis of the water environment test and evaluation using remote sensing information. Different types and concentrations of pollutants, make the water body in the color, density, transparency and temperature differences, resulting in a change in the reflected spectrum energy of water body. According to the difference of the reflected image information between remote sensing images in color, gray scale, texture features, the sources of pollution, pollution scope, area and concentration are identified. The chemical constituents of pollutants in the water body are complex and varied, and the shape is different, and the remote sensing sensor records the electromagnetic radiation characteristics of the target object. Therefore, not all of the polluting substances can be distinguished by remote sensing technology. Many scholars at home and abroad use remote sensing method to estimate the parameters of water pollution and the change of water quality. At present, the water eutrophication, suspended solids, oil pollution and heat pollution, urban waste water pollution and other pollution types are mainly analyzed, the details are as follows:4.1. Water Eutrophication TestThe concentration of chlorophyll in water is an indicator of the distribution of plankton, and the main factors reflected the eutrophication of water bodies, among them, the chlorophyll a is particularly prominent. Through the sampling of the chlorophyll biomass and other data, the greenness index of water body reflected by sampling data and remote sensing image data, the regression model of remote sensing can be established, to obtain spatial distribution information of chlorophyll and biomass in the water body, so as to realize the water eutrophication test .The test formula is as follows:

( ) 2 2, 2 | | | |

s r

s r

y h ps rs r

C x xK x k kh hh h

=

(4)

( ) ( )( )( ) 1/2

l ll

l

d E dU d

Var d

−=

(5)

At the beginning of 1970s, the researchers found that the chlorophyll content of the rich plants could be detected from the satellite. In 1978, the nimbus 7 satellite emitted by NASA loaded the first coast color scanner (CZCS) in the world,

95 SEPTEMBER 2017

the sensor has six bandwidth for 20 nm band, specifically designed for the chlorophyll quantitative remote sensing of sea surface. China in 1987 and 1989 respectively launched two satellites of FY-1A and FY-1B configurated with ocean color channel high resolution scanning radiometer, and get the high quality ocean’s chlorophyll distribution map. The development and application of new generation of Sea color remote sensing sensor (Sea WiFS) and Moderate Resolution Imaging Spectroradiometer (MODIS), has more bands and higher spectral resolution. In 2002, China launched the first ocean color remote sensing satellite HY-1, which carries two remote sensors, ten channels of ocean color scanner (COCTS) and CCD camera, which is more suitable for in the ocean color environment test and management.The scanning principle can be expressed by the following formula:

( ) [ ]( )

[ ]( )

1 1

i i hh x

h i ix s x s xx x

M x x x x xn n∈ ∈

= − = −∑ ∑ (6)

Using remote sensing method to estimate chlorophyll concentration of water body, many scholars at home and abroad have done a lot of work. Lathrop Landsat et al. using the R.G. 5 thematic mapping (TM) data evaluate the water quality of the Green Bay and the central lake. Shu Shourong et al. carried out spectral measurements in the pond water with different concentrations of chlorophyll by using multi spectral scanner (MSS). According to the experimental results, the mathematical model of the concentration of chlorophyll in fresh water was established. S.Ekstrand using TM data and measured data established a regression model to estimate the concentration of sea water. In “Nine five” period, Shanghai Institute of technical physics, Chinese Academy of Sciences in the support of National 863 plan, established the water quality remote sensing models and methods based on spectral measurement and sampling analysis, ground spectral measurement and remote sensing coordination. And successfully applied the remote sensing interpretation in the changes of eutrophication and the distribution in Taihu and Dianchi with the year, from 1980s to 90 years, to provide a quick way to master the situation of water quality, and a reliable technical method for the water quality remote sensing test.

In recent years, high spectral remote sensing technology has begun to pay attention to the analysis of spectral characteristics of water bodies. AISA were used to test the lake in Finland by K.Kallio and S.Koponen, and the practical algorithm of concentration of chlorophyll a was established, and the band of 685 ~ 691 nm was found to be in favor of the test of the poor nutrition and middle nutrient lakes. P.F link using principal component analysis of two lakes in Sweden obtained a small airborne imaging spectrometer (CASI) data, and the concentration diagram of chlorophyll was plotted, and the best band position and band width of the chlorophyll test were also pointed out.

4.2. Test of Suspended SolidsSuspended solid (SS) content in water is one of the important parameters of water quality. SS SS can not only be as a pollutant tracking agent, the sediment concentration of it has a direct impact on the water transparency and color of optical properties. Generally speaking, in the visible light remote sensing, different sediment concentrations in the 0.58~ 0.68 m appear to be the peak of radiation, which is the best band in the remote sensing of suspended material in the water, and selected by the land satellite, NOAA, the weather satellite and ocean satellite. In the practical test, the band is usually selected due to the better correlation with the concentration of suspended material, and combing with the practical test data of suspended material, which is analyzed, so as to establish the relationship model of the radiation value in specific band and the suspended solid concentration. Then the radiation of the band is retrieved and the concentration of suspended solids is obtained.

A series of models were proposed for the quantitative remote sensing study of solid suspended matter in water, and the quantitative relationship between the water body reflectance and SS was determined. In 1973, A.N.Williams et al used L.andsat-1 satellite to carry out remote sensing of suspended sediment in Chesepeake Bay, and found that the content of suspended sediment was linear with satellite remote sensing image data. V.Klemas et al applied remote sensing data in Delaware Bay study, the results show that the suspended sediment concentration and MSS brightness values are a logarithmic relationship, and a linear statistical model for estimating suspended sediment concentration is proposed. The relationship between water sediment and sea surface reflectance, and the relationship between sea surface and spectral reflectance of remote sensor in the atmosphere is the two important aspects of the establishment of the quantitative relationship between water reflectance and suspended solids. At the same time, some studies have considered that the logarithmic model is better than the linear model. Li Xia derived a unified formula that contains the Gordon relationship and negative exponential relationship, and uses TIM data to successfully apply the formula to the remote sensing quantitative analysis of SS in the Pearl River estuary.To obtain the linear model, according to the principle of least squares:

( )

( )

2

20 1 1 2 2

ˆa aa

a a a k aka

Q y y

y b b x b x b x

= −

= − + + + +

∑

∑

(7)

In the use of NoaA/AVHRR data to obtain SS content, Li Jing set up a negative exponential relationship between the reflectivity and the content of SS, and determined the relationship with the empirical formula, the results of theoretical analysis and practical application show that the

BOLETIM TECNICO DA PETROBRAS 96

formula has higher precision. Based on the Gordon model and Gordon atmospheric correction method, the utility system to obtain the content of SS is established by NOAA/AVHRR data by R.P.Stumpf. However, the atmospheric correction of II class water body still depends on the field synchronous sampling and measurement to eliminate the error. Although the NOAA satellite has a short scan interval, the resolution of the 1.lkm is still a bit rough on the mouth of the estuary, and these studies generally need the measured data on ground, and many areas are often lack of measured data, making these methods difficult to effectively apply. Spectral reflectance characteristics of different suspended solids concentration in the laboratory are tested by L.A.K.Merles. Spectral mixture analysis is also used in the study of the suspended material concentration in Minjiang, Fujian. The method can make full use of the multi band data, which does not require a large number of experimental data. It is suitable for the area of lack of measured data and has better practicability.

4.3. Oil Pollution TestAs we all know, the pollution of the marine environment, oil pollution is a large number, wide range, a deep pollution. It is reported that around 84% of Japan’s pollution is caused by oil pollution. Oil pollution not only in different degrees of growth and living environment of marine aquatic resources of coastal residents harm, but also damage to coastal facilities, the impact of climate change, reduce the self purification ability. Therefore, effective monitoring and control of marine oil pollution has become an urgent research topic.

It is difficult to realize the monitoring of oil pollution in a large area with the conventional method of point by point measurement. However, with the development of remote sensing technology, especially due to the development of various measuring instruments, make it possible. In fact, the remote sensing technology can make people from various phenomena in large range, fast speed, the space periodic observation on the ocean, especially for the detection of oil pollution is the most effective, can monitor the vast waters in a relatively short period of time, won the forecast time for people, for the protection of marine environment service. At present, foreign oil pollution monitoring using remote sensing technology, has made significant progress. Its main features are: a single application to the combined application of several methods of the original, from the monitoring by weather day and night to limit the development of all-weather, all day long time monitoring, from single parameter monitoring to multi parameter monitoring, monitoring range from small to large range monitoring. Therefore, the use of remote sensing test oil pollution can not only find the sources of pollution, pollution of the area and determine the estimation of oil content, and through continuous testing, can obtain the diffusion direction and speed of the oil spill, will affect the regional forecast. Remote sensing testing of oil pollution not only can find the pollution sources, determine the regional scope of the

pollution and estimate the oil content, but also can get the direction and speed of the oil spill through continues test, to predict the area which will be affected . The Xingang wharf oil pollution photos is shown as Figure 3.

Figure 3. Xingang wharf oil pollution photos

The application of remote sensing technology in offshore oil pollution testing began in 1969. In the United States, the offshore oil pollution area in California has been carried out by using airborne multi band visible light scanner, and it has better effect. Due to the advantages of flexible and optional remote sensing, the airborne remote sensing system is mainly used in the emergency treatment of oil spill. In recent years, the application of satellite remote sensing to the ocean petroleum pollution has been paid much attention by many countries. Li Qiqi et al. used the complementary advantages of the land satellite precision and the NOAA satellite, through the special processing of the image, to separate the oil pollution area and the surrounding sea area. Through the image data of ERS-2 satellite, M.Gade et al. analyzed parts of Mediterranean Sea, and revealed that the sea oil pollution is impacted by the different seasons of the wind speed. The satellite contained synthetic aperture radar, which has the advantages of all-weather and all day time because its working band is microwave and it uses the active working pattern. In 2002 “prestige” oil spill accident, the European Space Agency (ESA) using ERS-2 and ENVISAT-1 satellite made continuous oil spill test for the waters near the province of Galicia, Spain, to overcome the influence of weather conditions, and provide important technical support for emergency response decision-making of oil spill.

4.4. Heat Pollution TestBecause of human activities to discharge the waste heat to water body, the pollution of caused by warming effect of environment water body is called water heat pollution. The thermal pollution of water body can directly affect the diversity of aquatic organisms, leading to the destruction of

97 SEPTEMBER 2017

local ecosystems, which affects the production and life of human beings. Remote sensing image is an effective method to detect the thermal pollution of water body, and the main detection method is thermal infrared remote sensing and microwave remote sensing.

As the representative of the sensor, the high resolution radiation meter (AVHRR) can be accurately mapped out of the sea with a resolution of LKM, sea surface temperature images in which temperature accuracy is better than 1 degree Celsius. Based on the different multi-angle atmospheric path, A.Chedin eliminate the effect of the atmospheric effect by using the difference of the target absorption thermal infrared radiation, to made inversion of sea surface temperature. Using thermal infrared remote sensing, the experimental research results of water quality and ecological environment of Tangshan Douhe reservoir show that NIR can effectively test thermal pollution of water in the reservoir area. The effect of thermal infrared band of multi-phase TM data is used in Dayawan Guangzhou nuclear power station to make water temperature inversion. The effect of water temperature and water temperature, diffusion range and environmental impact of nuclear power station is evaluated. According to the principle of satellite remote sensing and image interpretation method, through the study of water body thermal pollution test method of thermal infrared and microwave remote sensing by Wang Jian, which provides a practical method for measuring the thermal pollution of water body.The use of remote sensing images to test the thermal pollution of water body has not caused the error caused by the need of tested hull into the polluted area, and can find many sources of pollution, and which has the advantages of low cost and high efficiency.

4.5 Urban Waste Water Pollution TestBecause of the different nature of wastewater, the reflection position and intensity of the characteristic curve is different, the ratio between the pollutant content and which bands has good correlation which is determined by the pollution and pollution degree of water body. Wastewater pollution is generally used by multi - spectral image testing, and can be determined by the thermal infrared method according to temperature difference.

The emissions of industrial wastewater and domestic sewage of City have large quantities of organic matter, their decomposition consumes a large amount of oxygen, and sewage is black and stink. Pollution status in the color infrared aerial photograph has very good display, it can not only directly observe the migration of pollutants, but also trace the source of pollution with the suspended sediment and phytoplankton of water as interpret indicator. The data of TM image in Pearl River of Guangzhou reach are made log transformation, HIS transform, KL transform and then density segmentation and image classification by Wang Yunpeng, et al. It is found that which is better used in the water pollution

distinguish and recognition. Yao Jun et al. interpreted the information of the three different time phases color infrared remote sensing images and thermal infrared remote sensing images of the Suzhou River, and the status and historical reasons of the pollution of the water body in Suzhou River are analyzed. In June 2001, spectrum test is made for eight different degree of pollution sewage in Xi’an Moat and Xingqing Park, to establish the correlation model between reflectance spectroscopy and BOD and COD content, and achieved good results. Ma Yueling et al, using TM image data to test the water quality pollution of water environment quality in the Pearl River of Guangzhou reach, and to establish a remote sensing model for water pollution prediction.

4. SUMMARY AND OUTLOOK

The application of remote sensing images in water pollution test shows the great potential of water environment remote sensing test method and the advantages of which is not in conventional test method. By the use of research results and practical experience of remote sensing image in water environment test can provide useful experience for the river, rivers and large areas of water pollution test and control. However, due to the complexity of the spectral characteristics of the polluted water, remote sensing image information is influenced by the atmospheric and the immaturity technology of remote sensing data processing, and there are still many problems in the application of remote sensing image testing water environment in China. Mainly include:

(1) Most of them are limited to qualitative research, or to carry out the analysis of existing aviation and satellite remote sensing data, but few quantitative analysis.

(2) The test accuracy is not high, and the algorithm is based on experience and semi empirical method.

(3) The algorithm has poor locality, seasonality, applicability and portability.

(4) The water environment parameters of the test are few, mainly in the suspended sediment, chlorophyll and transparency, turbidity and other parameters.

(5) The range of the remote sensing water environment is small, mostly concentrated in the visible and near infrared range, and the spectral resolution is different, especially the lack of water environment in the microwave band.

The development trend of the application of remote sensing image in water environment is as follows: (1) The establishment of remote sensing testing technology system.

Studying on the key techniques and methods of using a new remote sensing image data to make quantitative test of water environment, to form a standard of water security quantitative remote sensing technology system. According to the different types of inland water, a variety of water environment

BOLETIM TECNICO DA PETROBRAS 98

parameters inversion algorithms are established, to realize experimental remote sensing and remote sensing quantitative leap, to obtain the original and created results.(2) To strengthen the basic research of water environment remote sensing.

To deepen the understanding of the mechanism of remote sensing, especially the influence of water environment on the optical and thermal characteristics of surface water, to further develop the physical model, the water environment parameters are better associated with the optical measurement value obtained by remote sensing; To deepen the research of visual interpretation and digital image processing, and improve the accuracy of remote sensing image interpretation; to enhance the study of hyperspectral remote sensing, and perfect the data processing technology of airborne imaging spectrometer.(3) Remote sensing test of water environment in microwave band.

The majority of the selection of conventional water environment remote sensing measurement range are in the visible or near infrared, especially lacking of microwave band surface water environment study. The inversion of the surface water environment parameters can be improved by the combination of microwave band and visible or near infrared.

(4) To broaden the remote sensing water environment test.

At the present stage, the water environment remote sensing is limited to some specific water environment parameters, and the parameters such as chlorophyll, suspended matter and water body transparency and turbidity. It is less to study the characteristics of the parameters such as soluble organic matter, COD, and the quantitative remote sensing. Study on spectral characteristics of other water environment parameters should be enhanced to expand the quantitative testing of water environment parameters, and to further establish the spectral characteristic database of different water environment parameters.(5) To improve water environment remote sensing test accuracy.The results show that the accuracy, stability and special expandability of the water environment parameters are affected by remote sensing band setting using the inversion of water environment parameters. By using the satellite borne hyperspectral data, the spectral response characteristics of the continuous wave band of the wave height of 10nm and the main water environment parameters are studied, To determine the diagnostic spectrum and band combination of water environmental parameters, to form the remote sensing model of structural water environment parameters and the core technology of inversion, the accuracy of water environment test is improved.(6) To extend water environment remote sensing test model space.The intrinsic optical properties of the water environment component are studied systematically and deeply. The

mechanism of water environment remote sensing quantitative measurement is carried out by using hyperspectral data and medium and low resolution multi spectral data. The water environment components are quantitatively extracted and the mixture information of the components are separated, to eliminate the mutual interference between the components of water environment, and to establish a water environment parameter inversion algorithm, which is not limited by time and region. A remote sensing quantitative model for large scale and dynamic testing of low resolution remote sensing data is formed by using the multi spectral remote sensing measurement method and technology of the inland water body.(7) improved statistical analysis techniques.The accuracy of the water environment parameters obtained by the wide band remote sensing data is not very high. Mixed spectral decomposition technique and artificial neural network classification technology can be used for reference in the fields of geology, ecology and so on, to make full use of the water environment information, and to establish a water environment parameter inversion algorithm, which is not limited by time and region.(8) Integrated use of the “3S” technology.Using the characteristics of wide vision and information quickly updated of remote sensing technology and the water environment information and various parameters of the large area drainage basin and its surrounding areas are extracted in real time and quickly; GPS provides real-time and fast spatial location for the acquisition of spatial object and attribute information, to realize the corresponding relation between the space and ground measured data; GIS is used to complete the huge water resources and environmental information storage, management and analysis. The application of “3S” technology in the water environment remote sensing test is integrated, and the water environment remote sensing test and evaluation system is set up, and the water environment quality information is accurately and quickly released, to promote the construction of national water safety warning system.

5. Acknowledgement

1.The specialized research fund for the xianyang normaluniversity project (14 xsyk020)2.The xianyang normal university college students’ innovativeentrepreneurial training program (2015054).

6. REFERENCES

Han M, Liu B. Ensemble of extreme learning machine for remote sensing image classification. Neurocomputing, 2015, 149(PA):65-70.

Gao J, Xu L. An efficient method to solve the classification problem for remote sensing image. AEU - International Journal of Electronics and Communications, 2015, 69(1):198-205.

99 SEPTEMBER 2017

Chen L, Ma Y, Liu P, et al. A review of parallel computing for large-scale remote sensing image mosaicking. Cluster Computing, 2015, 18(2):517-529.

Ma Y, Chen L, Liu P, et al. Parallel programing templates for remote sensing image processing on GPU architec-tures: design and implementation. Computing, 2016, 98(1):7-33.

Ghahremani M, Ghassemian H. Remote-sensing image fusion based on curvelets and ICA. International Journal of Remote Sensing, 2015, 36(16):4131-4143.

Xu Xiaojuan. Application of remote sensing technology in water environment testing . Technology innovation and application, 2013, 30 (30): 149-149.

Wang H, Wang J, Jin H, et al. High-order balanced M-band multiwavelet packet transform-based remote sensing image denoising. EURASIP Journal on Advances in Signal Processing, 2016, 2016(1):1-14.

Wang L, Song W, Liu P. Link the remote sensing big data to the image features via wavelet transformation. Cluster Computing, 2016, 19(2):793-810.

Li J, Du Q, Li Y. An efficient radial basis function neural net-work for hyperspectral remote sensing image classifica-tion. Soft Computing, 2015, 20(12):1-7.

LIU Ying,YUE Hui.Analysis of Hongjiannao Lake Area Based on SMMI. Science Technology and Engineering, 2016, 16(16):122-127.

Atasever U H, Kesikoglu M H, Ozkan C. A New Artificial Intelligence Optimization Method for PCA Based Unsupervised Change Detection of Remote Sensing Image Data. Neural Network World, 2016, 26(2):141-154.

Jia Y, Su Z, Shen W, et al. UAV Remote Sensing Image Mosaic and Its Application in Agriculture. International Journal of Smart Home, 2016, 10(5):159-170.

GUO Chongbing.CHEN Hongyu,HAN Xingbo,et al.Visual Simulation of the Multi-System Interactive Verification for Remote Sensing Observation Satellite.Computer Simulation, 2016, 33(4):144-149.

Luo X, Zhang Z, Wu X. A novel algorithm of remote sensing image fusion based on shift-invariant Shearlet transform and regional selection. AEU - International Journal of Electronics and Communications, 2016, 70(2):186-197.

Liu J. Quality assessment of remote sensing image fusion using feature-based fourth-order correlation coefficient. Journal of Applied Remote Sensing, 2016, 10(2):26-50.

LI Gang,WU Zengyi.Remote sensing image classifica-tion based on multi-scale integrated. Electronic Design Engineering , 2016, 24(9):180-182.

Ma Y, Chen L, Liu P, et al. Parallel programing templates for remote sensing image processing on GPU architec-tures: design and implementation. Computing, 2016, 98(1):7-33.

Ghahremani M, Ghassemian H. Remote-sensing image fusion based on curvelets and ICA. International Journal of Remote Sensing, 2015, 36(16):4131-4143.

Deng RuRu. Guangdong, Hongkong and Marco water yield and water environment remote sensing testing system and application . China achievements in science and technol-ogy, 2014, 22（23）: 20-21.

Yu Haixia. Remote sensing image change detection and its application in rare earth mining test in Gannan [D]. Jiangxi Science University, 2014.

Huang W, Bu M. Detecting shadows in high-resolution remote-sensing images of urban areas using spectral and spatial features. International Journal of Remote Sensing, 2015, 36(24):6224-6244.

Zhao Rongjuan. Application of high resolution remote sensing images in coal gangue extraction Based on ENVI .Science and Technology Plaza, 2015, 02（10）: 217-220.

Song Y, Wu Y, Dai Y. A new active contour remote sensing river image segmentation algorithm inspired from the cross entropy. Digital Signal Processing, 2016, 48(C):322-332.

Basaeed E, Bhaskar H, Hill P, et al. A supervised hierar-chical segmentation of remote-sensing images using a committee of multi-scale convolutional neural net-works. International Journal of Remote Sensing, 2016, 37(7):1671-1691.

Lv D, Jia Z, Yang J, et al. Remote sensing image enhance-ment based on the combination of nonsubsampled shear-let transform and guided filtering. Optical Engineering, 2016, 55(10):103-104.

Zhang Yan, Wu Baoguo, Wang Xiaohui, et al. Parallel clas-sification of remote sensing images based on Grid Environment . Computer applications and software, 2015, 2:194-197.

Diao W, Sun X, Dou F, et al. Object recognition in remote sensing images using sparse deep belief networks. Remote Sensing Letters, 2015, 6(10):745-754.

Tang C, Chen Y, Feng H, et al. Motion deblurring based on local temporal compressive sensing for remote sensing image. Optical Engineering, 2016, 55(9):93-106.

[29]. Harikumar A, Kumar A, Stein A, et al. An effective hybrid approach to remote-sensing image classifica-tion. International Journal of Remote Sensing, 2015, 36(11):2767-2785.

[30]. Zhang F, Hou Z, Du Z, et al. Peer-to-peer transmission of remote sensing image data using image subset block. Optical Engineering, 2015, 54(9):93-107.

[31]. Huang, H.P., S.D. Xiao and Z.B. Li, Walking gait plan-ning for a kind of humanoid robot. Journal of Mechanical Engineering Research and Developments, 2016. 39(1): p. 88-96.

[32]. Xu, S., et al., An improved manchu character recognition method. Journal of Mechanical Engineering Research and Developments, 2016. 39(2): p. 536-543.