Extraction And Evaluation Of Remote Sensing Information Of Surface Water In China

Surface water is an irreplaceable strategic resources for human survival and social development,the information of surface water will help to fully grasp the characteristics of spatial distribution,monitoring the regional differences,promoting the study of the water cycle,but also on the rational utilization of water resources,the implementation of strategic measures has important significance.With the wide application of remote sensing technology in many fields such as national economy and social development,remote sensing technology has become an effective way to monitor the surface water efficiently and accurately.In the past,due to the fact that a large number of training data was difficult to obtain and the high-precision computing power was limited,remote sensing technology mainly focused on the extraction of small water bodies by single sensor,which was difficult to meet the efficient extraction of multi-source data on large scale in large data era.With the development of deep learning,a series of machine learning methods such as BP neural network,artificial neural network and Multi-layer perceptron have been developed rapidly.Therefore,this paper uses the Multi-layer perceptron method to extract information of the large scale surface water body.The main contents and conclusions are as follows:(1)This paper overviews the information products of large scale water body,summarizes the method of water extraction,and analyses the advantages and disadvantages of the method.In this article,the preprocessing steps of multi-source remote sensing data and the post processing method of water body are introduced in detail,which will be prepared for the study of surface water extraction based on multi-source remote sensing data in China.(2)By extracting and comparing the difference of water index,Multi-layer perceptron and supervised classification in the running time,storage space and product accuracy of the water body,we choose the method of Multi-layer perceptron,which is suitable for large scale water extraction.The method is not limited by satellite sensors and can be used to extract water directly by modifying the bands involved in the calculation.The learning algorithm of the multi hidden layer fully excavate the original data information,constantly the low layer features to the high level,and extract it.It not only accelerates the computation speed of a large number of data,but also improves the recognition rate of sample and the accuracy of the prediction.The selection of samples is beneficial to the improvement of the accuracy of the method,and the accuracy can be improved to a certain extent by iterative calculation of the objects which are similar to the water spectrum as non-water samples.(3)The method of Multi-layer perceptron is used to extract the surface water of the region of China.Using Landsat series images and domestic GF satellite images,the effective extraction of surface water body with 30m resolution in the land area of China in 2015 has been realized,and the overall precision is 90.56%.And spatial distribution of China regional water can be expressed with water rate and water area.Among them,the area of surface water in China is mainly concentrated in the 800mm-1600mm(49966km2)interval?inland river basin(51883km2)and Tibet autonomous region(33534 km2).The largest water rate is distributed in the 1600mm-4050mm(2.42%)interval?the Huaihe River Basin(3.74%)and Jiangsu Province(10.74%).According to the space distribution of land area in China,we can locate the southeast coast with more precipitation and the Qinghai Tibet Plateau with a large amount of snow water.Therefore,according to the application requirement and technology,and the use of a variety of domestic and international medium and high resolution satellite remote sensing images to study the method of surface water extraction in China.Large scale water body can effectively reveal the distribution law of surface water in China,and provide data support for the study of water resources and environmental changes in China and even in the world.