Research On River Segmentation Method Of Remote Sensing Image Based On Morphology

Surface water is one of the necessary resources for human survival,and rivers are an important part of surface water.With the significant increase in the rate of human consumption of natural resources,finding an efficient and statistical detection method of water resource has become a current research hotspot for scholars.In fact,statistics on surface rivers and water resources are an effective way to analyze surface water resources.For the hydrological survey and area statistics of rivers,the traditional method is to conduct field measurement manually.This survey method has a long period,labor-intensive and difficult to ensure accuracy.With the successful implementation of China’s high-score special project(one of China’s 16 major scientific and technological projects),our country has initially established a Chinese high-score satellite remote sensing system that can operate stably.This high-quality satellite data has become the trend of my country’s governance An important source of modern information.Because satellites can detect all the time,a series of problems in the past survey methods have been solved.Because the river area in the remote sensing river image has complex topography and various types of features,this paper uses mathematical morphology as the theoretical basis to segment the river area in the remote sensing image.(1)The thesis analyzes the current research status of remote sensing image river segmentation,and conducts simulation experiments and comparisons related to morphological image processing.When the traditional morphological method is used to detect the edge of a remote sensing image,it cannot accurately retain the edge of the river and will cause the river itself to be blurred.This paper proposes a morphological edge detection method that uses structural elements with multiple morphologies to detect remotely sensed river images(This method is called in the following content: the edge detection method of this article).By designing structural elements of various scales and shapes and making linear structural elements carry directional information,the edge detection performance of the operator is effectively improved;a pre-processing step is added before the simulation to enhance the edge information of the river while simultaneously improving the edge information of the remote sensing image.Existing multiple noises are suppressed;Using this method to detect river edge images with various edge features,the comparison of simulation results and data analysis proves that this method can accurately extract river edge information,and the algorithm has high detection accuracy and strong anti-noise ability.(2)In theory,it is feasible to apply morphological methods to segmenting river images,but the use of fixed structural elements for segmentation will change the characteristic information of the river and reduce the accuracy of image segmentation.In order to accurately segment the river region in the remote sensing image,this paper proposes a method for segmenting rivers based on the combination of morphological edge detection and region growing method.After preprocessing and morphological edge detection,one pixel in the river area is automatically selected as the seed point,and the growth double threshold is determined according to the gray value distribution displayed by the gray histogram,and the river is segmented according to specific growth constraints.Through simulation comparison with various methods and experimental data analysis,it is proved that the method in this paper can not only accurately locate the river area,but also can completely retain the river edge information,and has strong anti-noise ability and high segmentation accuracy.