High-resolution SAR image classification method based on non-subsampled contourlet fully convolutional network

Abstract

A high-resolution SAR image classification method based on a non-subsampled contourlet full convolution network, including inputting a high-resolution SAR image to be classified, performing multi-layer non-subsampled contourlet transformation on each pixel in the image, and obtaining each pixel The low-frequency coefficients and high-frequency coefficients of points; the low-frequency coefficients and high-frequency coefficients are selected and fused to form a pixel-based feature matrix F; the element values ​​in the feature matrix F are normalized to obtain a normalized feature matrix F1; The normalized feature matrix F1 is cut into blocks to obtain the feature block matrix F2 as sample data; construct the feature matrix W1 of the training data set and the feature matrix W2 of the test data set; construct a classification model based on a fully convolutional neural network; train the classification model; use The trained model classifies the test data set T, obtains the category of each pixel in the test data set T, compares the obtained category of each pixel with the class label map, and calculates the classification accuracy, which improves the classification accuracy and speed .

Description

technical field [0001] The invention belongs to the field of image processing, and in particular relates to a high-resolution SAR image classification method based on a non-subsampling contourlet full convolution network, which can be applied to high-resolution SAR images and effectively improve target recognition accuracy. Background technique [0002] Synthetic Aperture Radar (SAR) is a remote sensing sensor that has been widely researched and applied in recent years. Compared with other sensors such as optics and infrared, SAR imaging is not limited by conditions such as weather and light, and can be used for interested The target conducts round-the-clock, round-the-clock reconnaissance. Moreover, SAR also has a certain penetration capability, and can detect targets under unfavorable conditions such as cloud interference, tree cover, or shallow buried targets. In addition, due to the special imaging mechanism of SAR, high-resolution SAR images contain different content f...

AI Technical Summary

Problems solved by technology

However, none of the above feature extraction methods take into account the multi-scale, multi-direction, and multi-resolution characteristics of high-resolution SAR images. Therefore, it is difficult to obtain high classification accuracy for high-resolution SAR images with complex backgrounds.

Images