A Semantic Segmentation Method of SAR Image Based on Contour Structure Learning Model

Abstract

The invention discloses a SAR image semantic segmentation method based on a contour structure learning model. The implementation steps are as follows: (1) sketching the SAR image to obtain a sketch map; (2) regionalizing the sketch map to obtain a region map, and dividing the SAR image according to the region map (3) For the extremely heterogeneous regions in the pixel subspace of the mixed aggregation structure ground objects, the uniform sampling method is used to obtain the sample sets of each area; (4) The learning model of the contour structure of the ground objects is established; (5) Through The object contour structure learning model obtains the sample structure features, and divides the mixed and aggregated structure object pixel subspace; (6) divides the homogeneous pixel subspace and the structural pixel subspace; (7) merges the segmentation results of the three parts to obtain the final Segmentation diagram of SAR image. The segmentation result of the invention has better regional consistency and segmentation accuracy, and can be used for semantic segmentation of SAR images. The invention mainly solves the problem of inaccurate SAR image segmentation in the prior art.

Description

technical field [0001] The invention belongs to the technical field of image processing, and in particular relates to a SAR image semantic segmentation method based on a contour structure learning model, which can be further used for SAR image target recognition and detection. Background technique [0002] Synthetic Aperture Radar (SAR) is an important direction in the field of remote sensing technology. In extremely low-visibility weather conditions, high-resolution radar images similar to optical photography can be obtained. SAR has the advantages of all-day and all-weather, and has a wide range of applications in military, geological, surveying and mapping and other fields. With the continuous accumulation of SAR data and the continuous development of SAR technology, automatic interpretation of SAR images has become a research direction in the technical field. SAR image segmentation is the key to SAR image understanding and interpretation, and the accuracy of segmentati...

AI Technical Summary

Problems solved by technology

The disadvantage of this method is that when representing multiple extremely heterogeneous regions in the pixel subspace of mixed aggregated structure objects, compared with the learned features, it is difficult for the artificially designed features to be very good. Describe each extremely heterogeneous region of each image accurately; thus, it will affect the segmentation result of the pixel subspace of the mixed aggregation structure composed of multiple extremely heterogeneous regions

Images