ALI-TM: A moving objects detection algorithm for infrared images with dynamic background

作者：Zhao, Yufei; Song, Yong^*; Zhao, Shangnan; Li, Yun; Li, Xu; Hao, Qun; Guo, Zhengkun

来源：Infrared Physics & Technology, 2018, 93: 205-212.

DOI：10.1016/j.infrared.2018.08.003

• Moving objects detection
• Dynamic background
• Infrared images
• Lateral Inhibition
• Adaptive threshold segmentation

Moving objects detection is an important precursor of stable tracking and recognition, especially for the infrared scenes with dynamic background. In this paper, we proposed an Algorithmic Lateral Inhibition-Template Matching (ALI-TM) algorithm. This algorithm simulates the Lateral Inhibition (LI) mechanism, and achieves a good detection ability in the infrared scenes with dynamic background. First, the input infrared images are segmented into several binary images using an adaptive threshold segmentation. Then, for each binary image, the moving part can be detected by Algorithmic Lateral Inhibition (ALI) method, and the moving part includes moving objects and dynamic background. Finally, the moving objects can be separated from the dynamic background with a modified Template Matching method. Experimental validation of the ALI-TM algorithm is demonstrated on a diverse set of infrared images with dynamic background, the results showed that the proposed ALI-TM algorithm has higher detection precision, higher recognition rate, and better detection ability. Therefore the ALI-TM algorithm is capable to cope with the infrared scenes with dynamic background.

• 出版日期2018-9
• 单位北京理工大学