摘要
针对水下特殊环境的目标检测与识别,提出一种基于霍夫变换与几何特征相结合的特征匹配算法。通过对Canny边缘检测和迭代阈值分割的有机融合,实现了对图像的后处理式联合分割,有效地减小了光照不均等噪声干扰对目标提取的影响。通过霍夫变换直线匹配和几何特征匹配算法对规则的几何目标进行识别,具有较好的抗旋转和抗缩放性能。实验结果表明了该算法的准确性和稳定性。
Aiming at the detection and recognition of underwater target with special environment, a feature matching algo rithm based on Hough transform and geometrical feature is proposed. Combined the advantages of Canny edge detection and iterative threshold segmentation, the post processing combined segmentation algorithm greatly reduced the noise effects on target extraction such as non uniform illumination. Hough straight line matching and geometrical feature matching algorithm realized regular geometrical shape target＇s recognition, showed the performance of anti rotation and anti scaling. Experimental result shows the algorithm＇s accuracy and stability
出处
《现代电子技术》
2011年第4期73-76,80,共5页
Modern Electronics Technique
基金
水下智能机器人技术国防科技重点实验室研究基金资助(2009-1)
关键词
水下目标
检测与识别
后处理式联合分割
霍夫变换
几何特征匹配
underwater target
detection and recognition
post processing combined segmentation
Hough transform
geometrical feature matching
