基于改进SUSAN算法的红外目标边缘检测  

Edge Detection of Infrared Object Based on Improved SUSAN Algorithm

作　　者：赵静[1] 刘兴淼[1] 李忠科[1] 胡波 

机构地区：[1]第二炮兵工程学院,陕西西安710025 [2]第二炮兵工程设计研究院,北京100011

出　　处：《现代防御技术》2011年第3期152-156,共5页Modern Defence Technology

摘　　要：从3个方面改进了传统的SUSAN算法:①自适应选取阈值,以提高算法自动化程度;②使用双几何阈值,增强算法的抗噪性;③加入目标点判断因子,减少计算时间。利用改进的算法进行了红外目标边缘检测,结果表明,改进算法比传统的SUSAN算法有更强的噪声抑制能力,更快的边缘检测速度,显示出更加优良的性能。对对比度比较低的红外图像,改进算法可以减少40%以上的计算时间,而对对比度较高的红外图像,改进算法可以减少70%以上的计算时间。The traditional SUSAN algorithm is improved from three aspects, including： （1）adaptive select threshold to increase the robotization of the algorithm ; （2）make use of double geometric threshold to enhance the restrain noise ability of the algorithm; （3） add the target judge gene to improve detect efficiency. The simulation results have shown that the improved algorithm outperforms the traditional SUSAN algorithm in terms of both the de-noising performance and the edge detects efficiency. For low-contrast infrared image the improved algorithm can reduce more than 40% calculate time and more than 70% calculate time for high-contrast infrared image.

关 键 词：边缘检测 SUSAN算法 自适应阈值 双几何阈值 目标判断因子 

分 类 号：TN21[电子电信—物理电子学] TN957.52