SAR object identifying method based on multiple feature fusion bag-of-words model

Abstract

The invention belongs to the technical field of radar object identifying, and discloses a SAR object identifying method based on multiple feature fusion bag-of-words (BoW) mode. The method includes the following steps: 1. acquiring multiple local features corresponding to each image from training slice images and test slice images; 2. obtaining a vision dictionary corresponding to the multiple local features; 3. coding low-level features of the multiple local features to obtain histogram features of a BoW model of the training slice images and test slice images so as to further obtain the global features of the fused training slice images and test slice images; and 4. using a histogram suprachiasmatic nucleus support vector machine to identify the global features of the test slices. According to the invention, the method increases identifying properties, and can be applied to SAR object identification in complex scenarios.

Description

technical field [0001] The invention belongs to the technical field of radar target identification, and in particular relates to a SAR target identification method based on a multi-feature fusion word bag model, which can be used to provide important information for identification, identification and classification of vehicle targets. Background technique [0002] With the continuous development of SAR imaging technology, it has become one of the indispensable remote sensing platforms for earth observation. SAR image automatic target recognition (Automatic Target Recognition, ATR), especially vehicle target recognition, is one of the important applications of SAR images, and has attracted more and more attention from researchers. A typical SAR ATR system generally includes three stages: target detection, target identification and target classification. In the three-level processing flow of SAR ATR, target identification is in the middle link, which has the function of conne...

AI Technical Summary

Problems solved by technology

Considering the difference characteristics between the local structure of the target and the clutter in the high-resolution SAR image, the traditional features only provide a rough and partial description of the target, and they cannot describe the detailed local shape and structure information of the target and the clutter, which indicates that the discrimination Cannot make full use of the rich detail information of high-resolution images

When the target and clutter have no obvious difference in texture, size and contrast, traditional features cannot show good discriminative performance

[0007] To sum up, with the continuous improvement of SAR image resolution, traditional features have great limitations for target identification in complex scenes.

Images