Abstract
The test quality estimation of an Unmanned Aerial Vehicle (UAV) scattering measurement system is important to target detection and remote sensing. The kernel issue is to obtain credible scattering results, which give a ground truth for the test quality estimation. However, it is quite difficult to obtain these characteristics due to the limitation of test facilities. To tackle the mentioned problem and give a test quality estimation criteria for the UAV system, the geometric model of the test region is first reconstructed using a physical-based Convolutional Neural Network (CNN) and a self-designed vehicle scattering measurement system. Then, Shooting and Bouncing Rays (SBR) method is used to obtain accurate scattering of the test region. Finally, test errors and qualities are evaluated. Simulations and experiments are conducted to verify the geometric reconstruction and scattering calculation algorithms. Results show that the key error sources of the UAV scattering measurement system are observation angle and position deviations. The maximum and the minimum test quality of the UAV measurement system are 4.393 and 88.780, respectively.
Funding Information
  • National Natural Science Foundation of China (61871457)

This publication has 15 references indexed in Scilit: