SAR image classification method based on improved pcanet

Abstract

The invention discloses an improved PCAnet-based SAR image classification method, which mainly solves the high complexity of classification calculation caused by low efficiency of filter initialization and slow update and learning in the process of synthetic aperture radar high-resolution SAR image classification in the prior art , the problem of low efficiency. Concrete steps of the present invention are as follows: (1) read data; (2) slice processing; (3) normalization pretreatment; (4) extract image low-frequency component; (5) train principal component analysis network PCAnet; (6) Obtain the feature vector of the test set; (7) calculate the classification accuracy; (8) output the classification result. The invention has the advantages of short classification time and high classification accuracy for SAR image classification.

Description

technical field [0001] The invention belongs to the technical field of image processing, and further relates to a synthetic aperture radar SAR (Synthetic Aperture Radar) image classification based on an improved principal component analysis network PCAnet (Principal Component Analysisnet) in the technical field of high-resolution synthetic aperture radar image classification method. The invention proposes an improved PCAnet-based SAR image classification method, which effectively improves the problems of complex calculation and low calculation efficiency in SAR image classification. Background technique [0002] Synthetic aperture radar can work all day and all day long, and the image resolution it obtains is comparable to that of optical images. The classification of SAR images is an important branch in the field of synthetic aperture radar imaging. In the classification technology of SAR image, the feature extraction of the target is the most critical. Typical features ...

AI Technical Summary

Problems solved by technology

This method uses the data information of the image itself, the correlation information between images and the relevant information of images on multiple scales to extract relatively important information from a small amount of weak information, that is, from the training data marked with keywords. Learning the classification model of SAR images can greatly reduce the difficulty of obtaining accurate training data, and overcome some local uncertainties in SAR image classification. However, the disadvantage of this method is that at the same time In the process of obtaining multiple related information between images, the information between the data is cut too quickly, and a small amount of weak information used for training loses many important details, resulting in low multiple average classification accuracy of the classification method

However, the disadvantage of this method is that it takes a lot of time to train the filters in the deep RBF network, and the network parameters need to be adjusted by the method of back propagation error rate, and the computational complexity of the network training process is extremely high. , the training time is too long, and the network robustness is not strong

Images