Polarized SAR Image Classification Method Based on Weighted Dense Network

Abstract

A polarimetric SAR image classification method based on a weighted dense network, the steps of which are: (1) constructing a weighted dense network; (2) selecting a polarimetric SAR image to be classified; (3) filtering the polarimetric SAR image to be classified; (4) Obtain scattering features; (5) Use the scattering eigenvalues ​​of polarimetric SAR images to be classified to form a three-dimensional feature matrix; (6) Generate training data sets and test data sets; (7) Use weighted dense network to train data sets Carry out classification; (8) classify the test data set, and obtain the classification result. The present invention calculates the weight value of the feature map of the polarimetric SAR image to be classified, and only reserves the feature map with a weight value greater than 0.5 for classification, fully utilizes the most important features for classification, and suppresses unimportant features at the same time, improving the classification Accuracy, speeding up the training speed of the network.

Description

technical field [0001] The invention belongs to the technical field of image processing, and further relates to a polarization synthetic aperture radar (SAR) image classification method based on a weighted dense network in the technical field of radar image object classification. The invention can be used to classify ground objects in polarimetric SAR images. Background technique [0002] Polarization synthetic aperture radar is a high-resolution active microwave remote sensing imaging radar, which has the advantages of all-weather, all-time, high resolution, side-view imaging, etc., and can obtain richer information on targets. The purpose of polarimetric SAR image classification is to use the polarization measurement data obtained by airborne or spaceborne polarimetric SAR sensors to determine the category to which each pixel belongs. It has a wide range of applications in agriculture, forestry, military, geology, hydrology, and ocean research and application value. [0...

AI Technical Summary

Problems solved by technology

Although this method obtains the multi-type scattering features of the ground objects in the polarimetric SAR image and combines the texture features of the ground objects in the polarimetric SAR image, the disadvantage of this method is that due to the The image is obtained based on the principle of coherent imaging, so coherent speckle noise inevitably exists in polarimetric SAR images, so there are many features interfered by coherent speckle noise in the joint features, which are information of little significance for classification. The presence of information of little significance for classification can lead to low classification accuracy

However, the disadvantage of this method is that the training time of the deep SVM network is too long due to the need for multiple data exchanges during the training process of the collaborative training algorithm.

Images