Research on road and bridge detection method in high resolution SAR image based on multi-feature fusion

Abstract

Synthetic Aperture Radar (SAR) is a microwave remote sensing imaging radar that reduces interference from imaging constraints such as climate. It has all-weather, all-weather imaging characteristics, and can obtain rich information representing the physical properties of real objects on the surface. Widely used in the field of modern economic development. With the rapid development of modern urbanization, the types and numbers of supporting roads and bridges are constantly increasing, and they are constantly developing and changing every year. Therefore, it is very important to detect highway bridges with high-resolution SAR images that can cover a large area. Development is of great significance and has broad prospects for development. In this paper, the detection of roads and bridges in single-polarization high-resolution SAR images and the connection of global road networks are studied. The scattering characteristics of urban roads, bridges and nearby objects under the condition of single polarization SAR imaging are analyzed, and the effects of different environments and materials on the results of the target appearing on the image are studied. According to the inherent properties of the road and bridge targets and the imaging features in the SAR image, the multi-angle feature extraction of the target to be detected is further carried out, which lays the foundation for the subsequent road and bridge detection.