一种基于Harris特征点检测的改进算法  

Improved Algorithm Based on Harris Feature Point Detection

作　　者：卢伟家 刘缠牢[1] 

机构地区：[1]西安工业大学光电工程学院,陕西西安710021

出　　处：《仪表技术与传感器》2017年第12期107-109,104,共4页Instrument Technique and Sensor

摘　　要：针对Harris角点检测算法存在对含有噪声的图像角点检测不准确的问题,为降低噪点对角点检测的干扰和提高角点检测的精准度,提出了一种改进算法,通过在Harris算法设置的局部移动窗口中加入一种加权系数自适应的加权均值滤波,其加权系数由像素灰度值决定。通过实验证明,该方法可有效减少伪角点的形成,同时具有良好精确度和检测速度。In view of the Harris corner detection algorithm has an inaccurate problem in detecting corners of the image whichhas noise, in order to reduce the interference of noise when detecting corners and improve the accuracy of corner detection, animproved algorithm was proposed by adding a weighting coefficient adaptive weighted mean filter to the local motion window set bythe Harris algorithm, the weighting coefficient is determined by the pixel gray scale value. Experiments show that this method caneffectively reduce the formation of pseudo？corner points and has good accuracy and detection speed.

关 键 词：HARRIS算法 特征点检测 角点检测 加权系数自适应 图像处理 

分 类 号：TN911[电子电信—通信与信息系统]