Image micro-target recognition method based on multi-layer empirical mode decomposition algorithm

Image micro-target recognition method based on multi-layer empirical mode decomposition algorithm

연구 분야: Artificial Intelligence

학회: 2022 14th International Conference on Measuring Technology and Mechatronics Automation (ICMTMA)

In order to improve the recognition accuracy of image tiny targets, a recognition method of image tiny targets based on multi layer empirical mode decomposition algorithm is proposed. Firstly, the spectral characteristics of the image are collected and frequency-divided to decompose the image into different sub-frequency domains, and the tiny targets are roughly extracted. Then, multi-layer empirical mode decomposition algorithm is used to decompose the input function of the original dim target image into two-dimensional characteristic mode functions. Finally, the difference between the small target area and the background area is obtained by differential calculation, and then the small target area is segmented. Experiments show that the plane image weak and small target recognition method based on depth learning algorithm can extract weak and small targets efficiently and accurately, better suppress complex background, improve image signal-to-noise ratio and ensure image recognition accuracy.