A novel LBP based algorithm for small target detection in infrared image

作者：Guo, Tong; Sun, Xiechang^*; Li, Meng; Xiao, Weidong

来源：Optoelectronics and Advanced Materials-Rapid Communications, 2013, 7(9-10): 672-675.

• Small target detection
• Local binary patterns
• Fusion image
• Target preservation

Small target detection is a crucial problem in Infrared Search and Track (IRST) systems. Conventional algorithms encounter a high false alarm ratio under situations with cluttered background. Here a novel algorithm based on modified local binary pattern (LBP) model is proposed to deal with this problem. The modified LBP process will generate a specific uniform pattern at the position where potential small targets exist, while background and other image components will generate different patterns because of their different intensity distribution from target. After potential small target is determined a post process using fusion image of two LBP images is incorporated to preserve target information. The algorithm is real-time. Comparative experimental results demonstrate its efficiency and robustness.

• 出版日期2013-10
• 单位中国人民解放军国防科学技术大学; 中国空间技术研究院