摘要
通过红外图像边缘检测实现激光红外目标识别。针对灰度值轮廓点标记边缘检测方法精度不高的缺点,提出一种基于组合优化理论下自适应权重均衡分割的红外图像边缘检测算法。首先对采集的红外图像进行小波降噪预处理,然后对降噪提纯的红外图像进行Harris角点检测,以邻域灰度值为信息素优化融合信息组合,采用自适应权重均衡分割方法提取红外图像的边缘特征,对提取的边缘特征通过最小二乘组合优化处理实现轮廓检测。最后通过仿真实验进行性能测试,结果表明,采用所提方法进行红外图像边缘检测的输出峰值信噪比较高,均方根误差较小,提高了边缘检测的特征点跟随能力,实现图像目标识别优化。
The objective is to achieve the laser infrared target recognition through the infrared image edge detec- tion. Consider the the low accuracy of the edge detection method based on the gray value contour point mark, this pa- per proposed a infrared image detection algorithm of adaptive weight equilibrium segmentation based on combinatorial optimization theory. First, wavelet denoise preprocessing the acquired infrared images. Then, go on the Harris corner detection of the denoised and purified infrared image. Optimize the information combination according to the neighbor- hood gray value information elements. Acquiring the edge features of the infrared image by adaptive weight equilibrium segmentation. To achieve the contour detection by the least squares combinatorial optimization of the acquired edge fea- tures. Finally, the performance analysis of the simulation shows that this method maked the output image a high peak SNR （noise-signal ratio）, a small RMS error. It improved the following ability of feature point in edge detection and realized the optimization of image target recognition.
作者
陈阳
CHEN Yang(Zhejiang University of Media and Communications, Hangzhon 618000, China)
出处
《激光杂志》
北大核心
2017年第4期105-108,共4页
Laser Journal
基金
教育部高校硕士点基金(200801120007)
关键词
红外图像
组合优化理论
边缘检测
识别
infrared image
combinatorial optimization theory
edge detection
recognition
