Remote Sensing Monitoring And Environmental Pollution Load Assessment Of Coastal Aquaculture Area Based On GF-2

Coastal aquaculture surveys play an important role in the development and utilization of coastal resources.Recently,with the development of satellite remote sensing technology in China,coastal aquaculture monitoring with high-resolution remote sensing technology will greatly reduce survey costs and improve survey efficiency.However,traditional pixel-based information extraction methods have been unable to meet the requirements of precision for the tremendous spatial and texture information contained in the high-resolution satellite images,it is increasingly important to explore effective high-precision remote sensing information extraction methods.In this study,we proposed an object-based classification method based on the index standard deviation and then adopted to extract the Houtouwan aquaculture area with high suspended sediment concentration in Gouqi Island with GF-2 on January 16,2016.Three traditional pixel-based classification methods were also performed on the same image to compare the classification accuracy,factors affecting the accuracy of each method were analyzed finally.Then,the optimal remote sensing extraction results were used to evaluate the environmental pollution load of nitrogen and phosphorus nutrients in the mussel aquaculture area.Results and understanding are as follows:1.The object-based classification method based on index standard deviation proposed in this paper is to establish the index RWI to highlight the features of the aquaculture area.By analyzing the characteristics of the index image,the index standard deviation is used as the discriminant index for the extraction of the aquaculture area.According to the distribution of suspended sediment concentration in the study area obtained by water color inversion,the extraction threshold of the aquaculture area under different turbidity is determined respectively,so as to obtain the overall optimal extraction effect.The advantage is that according to different suspended sediment concentration conditions,the local effects of each level are optimized by the rule limitation,so that the combined overall extraction result has the highest precision and the best effect.2.Although the study area exhibit serious "same spectrum with different objects" for the highly turbid water,the object-based method performs well with accuracy of 94.10%.Among the three pixel-based classification methods,the classification accuracy of the support vector machine with the highest classification accuracy is 80.78%,and the neural network method was indicated with lowest classification accuracy of only 71.17%.The results of precision evaluation show that the object-oriented classification method proposed in this paper can extract the aquaculture area in the turbid waters more accurately so as to obtain the information such as the area and distribution of the aquaculture area.3.Through the assessment of nitrogen and phosphorus environmental pollution load,it is found that from summer to autumn Houtouwan mussel aquaculture area has a higher nutrient load per day.This leads to greater pressure on the eutrophication of the waters.Therefore,the scale and density of the aquaculture should be reasonably controlled,and the pollution of the water should be monitored in a normalized manner to reduce or even eliminate the eutrophication damage caused by the aquaculture area.The results of this study can provide effective technical support for the management department of decision-making and scientifically plan when managing coastal aquaculture areas.