A Review of Satellite Remote Sensing Monitoring Methods for Sea Surface Oil Spill

WANG Xinsheng, WANG Chenxu, ZHAO Yinan, LUO Qinghua, LIU Zhiyong, ZHU Zhiquan

Harbin Institute of Technology (Weihai), Weihai 264209

Abstract: Using satellite remote sensing to monitor oil spill on the sea is an advanced means of oil spill monitoring,and it has the characteristics of wide coverage, speediness and real time, synchronization, continuity, and low cost.Hence, accelerating the research on this technology and establishing a satellite remote sensing monitoring mechanism suitable for oil spill emergency situations is of great signi ficance to improve China’s oil spill monitoring capability and prevent or reduce the pollution damage caused by oil spill in the marine environment.This paper analyzes and studies the current situation using satellite remote sensing to monitor oil spills at home and abroad. Based on the basic principle of satellite remote sensing, this paper systematically studies the satellite remote sensing monitoring oil spill principles, satellite data processing methods and oil spill information ident ification, and summarizes an oil spill ident ification system that can realize oil spill information reproduction. This system provides an important means of support for the handling of oil spill accidents.

Key words: satellite remote sensing, oil spill, radar satellite, spectral satellite, image processing

1 INTRODUCTION

Frequent occurrence of oil spill accidents and the associated serious damage to the marine ecological environment[1]demands higher requirements for China’s marine oil spill response work. After the oil spill accidents occur, determining the oil pollution event, the scope of pollution, sources of pollution,predicting the trend of oil spill drift, etc., implementation of the corresponding oil spill emergency decision-making process has become an urgent task to be solved in China.

After the oil spill incident, from satellite remote sensing data,we can timely and accurately establish the oil spill, the range of oil pollution, and predict the oil spill dynamics, which greatly enhances the pertinence and effectiveness of emergency decisionmaking process, thus considerably reducing the operational blindness, improving the ef ficiency of emergency response, and reduces the pollution damage caused by oil spill. In addition,the use of satellite remote sensing to monitor the oil spill, as the satellite remote sensing data can accurately and objectively establish the scope of pollution and the degree of pollution, is an important assessment tool to determine the damage caused by oil spill and providing legal evidence for damage claims after the incident.

Remote sensing is now a comprehensive technology de-rived from Earth observation technology which was developed in the 1960s. Remote sensing detection is the comparison of the electromagnetic wave of atarget object received by the remote sensing instrument using the reflection spectrum of the object, so that the objects on the ground can be ident ified and class ified. This is the basic principle used by remote sensing[2].Satellite remote sensing refers to a remote sensing method in which sensors are placed on art ificial satellites. Satellite remote sensing can be divided into radar satellites and spectral satellites according to the sensor imaging mechanism of satellites. Internationally, the use of satellite remote sensing to monitor the marine environment began in the early 1970s. Some foreign countries have turned from visible and infrared light research in the 1980s to microwave remote sensing monitoring for oil spill research in the 1990s. The research using satellite remote sensing to monitor oil spills has also evolved from a single qualitative analysis that monitors the spread of the oil spill and predicts the direction of oil spill drift to the quantitative analysis which calculates the thickness and area of the oil film. The application research using satellite remote sensing to monitor oil spills is still in its infancy in China. Therefore, a feasibility study on the information extraction method for satellite remote sensing monitoring oil spills, the establishment of an application operating system, the satellite information acquisition method for continuous monitoring of oil spill accidents and the monitoring of the illegal discharge of ships is of great sign ificance to improve the emergency response capability of oil spill management in China and to prevent or reduce the pollution damage caused by oil spills in the marine environment.

2 THE PRINCIPLE OF RADAR SATELLITE REMOTE SENSING FOR MONITORING OIL SPILLS

Radar satellite is the collective name of the ground-sensing remote sensing satellite carrying synthetic aperture radars (SAR).It is a radar measurement that combines the brightness and phase on the received back-reflected signal of the electromagnetic wave information transmitted to the ground. SAR operational bands belong in the microwave, microwave bands and can penetrate cloud, rain and snow, hence the system has an all-weather working ability[4].

Radar satellite remote sensing is a kind of active remote sensing. As an active sensor using microwave frequencies to sense, the radar is impacted less by meteorological conditions than optical and multi-spectral sensors. It can work day and night and in all weather, regardless of time and climate conditions, and can detect oil spills in real time. It has become an important tool for oil spill pollution detection. Figure 1 illustrates the flow chart of oil spill detection in SAR images.

The radar imaging process is as follows: electromagnetic waves are transmitted from the radar antenna of the radar and propagate towards the Earth. Some electromagnetic waves are refracted, absorbed, and attenuated by one or more layers as the signals are propagated through the atmosphere. Other electromagnetic waves interact with the ground and then are reflected back to the radar.

Such phenomenon in radar satellite images can be explained with the “Radar Backscattering Coef ficient”. Generally, it can be understood that the rougher the surface, the brighter the image, and the smoother surface appears black because little or no electromagnetic wave energy is reflected back. The imaging of any ground object by a radar satellite is described by the backscattering mechanism of the electromagnetic wave.

When volume of the oil spill on the sea surface is relatively low, the sea surface can be seen as a rough surface, and when there is a larger volume oil spill on the sea surface, the tension of the sea surface is changed due to the action of oil, so the sea surface tends to appear as a smooth horizontal plane. The radar technology is based on the backscattered echo signal of the receiving target. The capillary wave enlarges the roughness of the sea surface, making the echo signal returned to the radar stronger, so that the corresponding sea surface in remote sensing image shows brighter. The oil film formed by oil diffusing will suppress the generation of capillary waves on the sea surface due to the viscous force and smoothness of the sea surface. Therefore, the reduction of the echo signal intensity makes the sea surface covered by the oil film appear darker on the radar image, and the oil pollution is distinguished from the background sea water. This is the mechanism of radar detection of oil pollution on the sea[5].Through the study, we found that not all radar satellite can successfully monitor the oil spill under different conditions. One of the main factors is decided by an important technical parameter of radar, that is, the resolution of satellite detection. The other is decided by sea conditions and environmental factors. Satellite detection resolution refers to the minimum distance that a satellite can distinguish between two objects. The satellite detection resolution usually has three parameters: spatial resolution, gray resolution, and temporal resolution. The SAR satellite’s ability to measure depends on weather conditions and the thickness of the oil film. At low wind speeds, naturally occurring marine oil spills occur and may cause false alarms. At high wind speeds, oil may be mixed into the water and SAR satellites cannot detect surface suppression.

Figure 1 The flow chart of oil spill detection in SAR image

3 THE PRINCIPLE OF SPECTRAL SATELLITE MONITORING FOR OIL SPILL

Spectral satellite is the collective name of the satellite carrying spectral sensors. It uses the spectral sensor to detect the emitted solar radiation and reflected radiation of a ground object, so it is limited by solar radiation and can only work in day light[6].

The information content of an ocean remote sensing image is mainly spectral feature information, which is the scient ific basis for interpreting the remote sensing image. On the premise of excluding atmospheric in fluence, the luminance value of a single pixel in a remote sensing image reflects the average reflected radiation value of the object, which is called the spectral char-acteristics of the ground object. Marine oil pollution will change the physical properties of sea water. The difference in temperature, thermal radiation, and reflection, scattering, and absorption of sunlight between the oil film and sea water has led to changes in the gray value of satellite image data, which makes a difference in color, texture and brightness of satellite image[7].The spectral characteristics of oil film are the basis of remote sensing detection of oil pollution at sea. Mastering the spectral characteristics of oil diffusion, drift, melting and evaporation after the oil pollution enter the sea provides a reliable basis for judging the types of oil pollution, calculating the diffusion area of oil pollution, analyzing the thickness of oil film and estimating the amount of oil spill. It also has important guiding sign ificance for remote sensing detection of oil pollution at sea, as well as for oil pollution removal and pollution assessment after pollution occurs.

As ground objects have both the characteristics of reflecting sunlight and emitting heat, the characteristics of reflecting radiation and emitting radiation can be used to analyze the same spectral satellite data at the same time. The characteristics of reflected radiation are mainly manifested in the visible light band,different ground objects exhibit different reflectivity on satellite images, while the characteristics of emissive radiation is mainly exhibited in the infrared band, different ground objects with different thermal radiation characteristics can be determined on the satellite images. Figure 2 illustrates the framework of spectral image class ification.

In nature, the composition, structure, shape, and even the nature of the ground objects will be affected by various factors, which are often deduced together with the surrounding environment. The spectral characteristics of an oil spill area are affected by three factors:

1) Oil spill’s chemical properties, such as composition and fusion degree with sea water and other physical properties, such as spec ific gravity, solubility, temperature;

2) External environment’s intervention, such as sea temperature, sea waves, sea color, and regional sea area characteristics;

3) Instrument error and deviation, such as solar elevation angle, satellite attitude, electromagnetic interference in transmission. In addition, if the satellite passes over the oil pollution area which may be covered with cloud or thick fog, the satellite data captured under these meteorological conditions will not be useful. Another important factor is that spectral satellites can only be utilized during the day. The above physics cannot be changed by human beings.

4 SATELLITE IMAGE PROCESSING TECHNOLOGY

The raw data of the remote sensing image provided by the satellite contains a large amount of information, including remote sensing information of various ground objects, as well as interference and noise. It is necessary to eliminate interference,correct distorted signals, suppress unnecessary information and highlight useful information in order to achieve practical results.All these tasks are part of the remote sensing image processing.

Figure 2 The framework of spectral image classification

The image processing software includes PCI image processing software and ENVI image processing software. The remote sensing image is a kind of information processing that objectively and truly measures and reflects the electromagnetic radiation intensity of ground objects on a certain scale. It is a form of information obtained by remote sensing exploration. As a result, the key problem of remote sensing technology is to judge and recognize the class ification and distribution characteristics of ground objects according to the characteristics of the electromagnetic wave radiated by ground objects in remote sensing images.

There are many methods to extract oil spill information from radar satellite images, and the appropriate methods can be selected according to the actual situation. Image processing refers to various technical processing of images or data acquired by remote sensing. The purpose of processing is to make the images or data more suitable for easy use, such as enhancement, filtering, class ification, correction, mosaic[8]. In fact, each method is completed by a combination of computer image processing and arti ficial image interpretation. In addition, there is atmospheric correction algorithm and image fusion algorithm.The atmospheric correction algorithm is suitable for spectral satellites, and the image fusion algorithm is based on the radar satellite and spectral satellite data, both of which are key technologies for image processing.

4.1 Oil Spill Recognition Technology Based on Radar Satellite Images

According to the radar monitoring oil spill principle, we know that the existence of the oil film on the ocean surface smooths the ocean surface, which reduces the backscattering echo received by radar. In the radar image, the brightness of the oil film is lower than the brightness of the characteristics of the surrounding sea surface, so it shows the characteristics of black patches. However, it does not mean that as long as there are black patches on the radar satellite image, it is absolute that there is an oil spill on the sea. Due to other marine features as well as atmospheric phenomena, such as sea areas with relatively low surface wind speed, raindrops and oily sea ice can also damp short surface waves and form black patches on radar satellite images. The image features with black patches formed by these non-surface oil films can be collectively referred to as“oil-like film” phenomenon[9].

Oil-like film ident ification methods include using: geometric features to distinguish, geographical features to distinguish, environmental features to distinguish, “complexity” and using “contrast”.

There are two difficulties in calculating the amount of oil spill:

1) Oil film thickness is dif ficult to measure. The oil film on the sea surface is not continuous. It is affected by wind and flow. Over time, there will be different shapes and thicknesses of oil films, or oil bands, or fragments, or small tar balls;

2) The concentration of oil dispersed in water is not easy to measure. The thickness of oil film directly affects the concentration of oil dispersed in water, and different thickness determines different concentrations[10].

4.2 Spectral Satellite Image Processing Technology

There are two main methods for spectral satellite image processing: the sensitive channel color synthesis method and the channel operation enhancement method[11,12].

4.2.1 Sensitive channel color synthesis method

There are usually two methods for sensitive color synthesis,namely false-color synthesis and pseudo-color synthesis. Falsecolor synthesis is to reproduce a natural color image using a new tricolor component obtained by mapping function transformation, while pseudo-color synthesis is to transform each gray level of a gray image into different colors according to a mapping function. Both of these image enhancement methods are aimed at enabling the human eye to better distinguish the detailed features on the image.

4.2.2 Channel operation enhancement method

The multi-channel image combination method combines different channels on an image according to certain rules, which not only synthesizes the characteristics of each channel, but also enlarges the dynamic range for naked eye observation, so that different types and shapes of ground objects in the image have a better visual effect. The basic starting point of multi-channel is to extend the spectrum difference and display the dynamic range of ground objects as far as possible.

5 CONCLUSIONS

At present, the ability to use satellites to monitor oil spills is basically available in China. Satellite resources are relatively abundant. Although the acquisition time has a certain delay, the satellites can be quickly utilized through a certain cooperation mechanism.

Satellite remote sensing technology for monitoring oil spills on the sea is basically mature. On the basis of understanding the basic principle of satellite remote sensing, the theory of monitoring oil spills by radar satellites and spectral satellites has been studied. Through the study, we believe that the technology of monitoring oil spills by radar satellite is mature, and there is a mature application mechanism available. Monitoring oil spills by spectral satellite is limited due to the spec ific remote sensing band, but some spectral satellites can also be used to monitor oil spill.

On the basis of theoretical research, the technology for oil spill information processing in satellite imagery has been studied by using the professional image processing software PCI equipped by the center. In particular, the radar satellite is further studied and a set of standardized radar satellite oil spill information has been summarized. Through the application with several accidents, the ident ification method has been proved to be normative, scient ific and practical, and the ident ification method is further improved through such accidents.