Research On The Method Of Classifying The Perimeter Of Roads Around Roads Based On SAR Polarization Technology

The development of the human race requires a road,and the environment around the road is affecting the use of the roads,so the needs of the land around the roads are increasing.As the telemetry microwave technology continues to evolve,people are increasingly concerned about the image of the microwave field.However,compared with the abundant image data and the increasing enthusiasm of people for the full polarimetric SAR technology,the processing power of microwave remote sensing image is not satisfactory.But full polarization SAR image processing technology is not affected by time and environment is a factor of optical image can’t have,this is the reason why people cannot give it up.This is a study of how to classify the land around a full polarization SAR.Based on the four classification algorithms,wishart iterative algorithm and watershed segmentation algorithm are used to assist in the classification of ground objects within 5kilometers of the road.The specific research contents are as follows:(1)The four methods of decomposition of Pauli decomposition,cloude decomposition,Freeman decomposition and Yamaguchi decomposition were comprehensively analyzed in qionglai area and maoxian area,and compares and analyzes the advantages and disadvantages of species classification within 5 kilometers.(2)In view of that problem of the classification of the land and the cause of speckle noise in the fully polarized SAR,the cloude decomposition and wishart algorithm are combined to make a preliminary classification of the processed images by using wishart classification,and the validity of semi-supervised classification is verified by an example.By combining the watershed segmentation algorithm and Yamaguchi algorithm,a large number of speckle noise problems appear in the image.(3)The way to analyze a multiple dissolving method in a support vector in a support vector in a support vector in the way that the property is divided into the surface of the road,it shows that the characteristics of the Pauli dissociation method are good for the higher accuracy.Design a watershed segmentation algorithm and Pauli decomposition,the combination of texture feature and its take into simple multi-core vector machine(SVM)classification algorithm.And on this basis,put forward a the same kernel function,the transform weighting coefficient to the polarizationcharacteristic matrix,and then into the calculation method of support vector machine(SVM),show that the method of precision has a reference value.(4)The higher the polarization characteristics of the support vector machines,the higher the accuracy of the classification results.The polarization characteristics of the three different classification methods are substituted into the support vector machines to classify,to a certain extent,to improve the classification accuracy and avoid the repeated use of polarized information.