红外弱小目标TSA检测技术  

Infrared Small Target Threshold Segmentation Algorithm Detection Technology

作　　者：李佳来 

机构地区：[1]空军哈尔滨飞行学院,哈尔滨150001

出　　处：《四川兵工学报》2014年第5期89-92,共4页Journal of Sichuan Ordnance

基　　金：国家自然科学基金(51279036)

摘　　要：针对红外弱小目标检测,提出了一种改进的最大类间方差算法,利用遗传算法在寻求最优解方面的优势,采用图像像素点间的类间方差作为判决函数,实现了改进的算法设计方案,达到了对类间方差阈值分割检测算法的改进;通过对真实红外视频中采样得到的红外图像进行处理,将本文算法与传统的Otsu算法进行仿真比较,实验结果证明此算法在保证了分割算法检测精度的前提下,达到了对传统算法实时处理性能上的改进,具有工程实践价值。For infrared dim target detection, this paper proposes an improved Otsu algorithm based on threshold segmentation algorithm （TSA） , advantages in the search for the optimal solution using genetic algorithms. We used class variance between the image pixels as a judgment function, realized the improved algorithm, and reached the class variance detection threshold segmentation algorithm improvements. Infra- red image processing by real infrared video sampling, the proposed algorithm with traditional the Otsu algo- rithm simulation comparison, experimental results show that the proposed algorithm guarantees that, under the premise of the detection accuracy of the segmentation algorithm, it reaches the improvement of the per- formance of real-time processing of traditional algorithms, and excellent algorithm program with a value of engineering practice.

关 键 词：目标检测 阈值分割算法 遗传算法 红外目标 

分 类 号：TN976[电子电信—信号与信息处理]