基于帧差法与SURF匹配的异纤机棉流速度测量

Measurement of Cotton Flow Velocity of Different Fiber Machine Based on Frame Difference Method and SURF Matching

针对异纤机输棉通道内棉流速度的不确定性导致其打击精度无法提高的问题,课题组提出了一种基于帧间差分法和SURF特征匹配算法相结合的混合棉速测量方法。首先,对采集到的棉团图像进行预处理,采用帧间差分法提取棉团边缘信息,得到分离的二值化棉花图像;其次,基于形态学操作对二值化图像进行腐蚀、膨胀运算,得到完整的待检测棉团信息;最后,通过SURF特征点匹配算法对2幅图像中的特征点进行匹配,计算相邻帧间同名点之间的像素距离,将像素距离转换为实际距离后,计算得到棉团运动速度。研究结果表明:与直接匹配相比,特征点数量减少了67.8%,无效匹配点对减少了97.47%,算法计算所得速度与直接计算所得速度误差低于4%。课题组提出的算法可以满足实际工业应用要求。

In order to solve the problem that the impact accuracy couldn′t be improved due to the uncertainty of cotton flow velocity in the cotton conveying channel of the foreign fiber machine, a hybrid cotton speed measurement method was proposed based on the combination of inter frame difference method and SURF feature matching algorithm. Firstly, the cotton ball image was preprocessed, and the edge information of cotton ball was extracted by the inter frame difference method to obtain the separated binary cotton image. Secondly, the binary image was corroded and expanded based on morphological operation, and the complete cotton ball information to be detected was obtained. Finally, the feature points in the two images were matched by SURF feature point matching algorithm, and the pixel distance between adjacent frames with the same name was calculated. After converting the pixel distance into the actual distance, the cotton ball motion speed was calculated. The results show that compared with direct matching, the number of feature points is reduced by 67.80%, the number of invalid matching points is reduced by 97.47%, and the error between the speed calculated by the algorithm and the speed calculated by the direct method is basically less than 4%. The algorithm can meet the actual industrial requirements.