Image Recognition Algorithm Verification Technology Based on Intelligent Unmanned Systems

With the rapid development of artificial intelligence technology, the application scope of intelligent unmanned systems is expanding. To evaluate the perception ability of intelligent unmanned systems, it is important to solve the problem of insufficient test methods and tools related to image recognition, so it is necessary to study the verification technology of image recognition algorithms. We have developed a testing tool, which uses containerization technology to package the image recognition model and the running environment into image files and run them on the same hardware platform. We choose two image recognition algorithms, build a benchmark data set, and use testing tools to verify the performance of the algorithms.