Research on radar target radial length extraction algorithm based on U-Net

The main problem with the current radar target radial length extraction algorithm is its susceptibility to interference signals, which makes it difficult to optimize the boundaries of the target support area, especially the impact of noise at positions that tend to be farther away from the target support. Therefore, based on deep learning networks, this article trains and analyzes different pixel points, completes the segmentation of the target support area and background area through image semantic segmentation algorithms, obtains the target support area, and estimates the radial length of the target based on the boundary of the target support area. Finally, validate the effectiveness of the algorithm using simulation data.