【摘要】 为提高烟草打叶复烤工艺中烟叶叶片结构的检测效率,研究了用于烟叶叶片结构在线检测的叶片面积快速测量方法。采用空变模糊矩阵还原了线阵相机的场曲模糊,利用张正友标定法校正了图像畸变;提出了对噪声、照明环境不敏感的R、G分量差值阈值分割法,实现了叶片图像的高精度分割;提出了计算效率较高的叶片面积拼接法,解决了在线检测时单个叶片被分割在2幅图像中的问题。采用多线程技术,编写了并行图像处理软件,开展了重复性实验,结果表明:图像处理速度达22帧/s,能够满足在线测量要求,在100～2 500 mm2量程内,相对误差小于0.50%,变异系数小于1.01%。更多还原

【Abstract】 To improve detection efficiency of tobacco strip size distribution in the process of tobacco leaf threshing and redrying,a quick on-line measuring method of tobacco leaf area for analyzing tobacco strip size distribution was investigated.Leafs area were measured on-line by analyzing the leaf images on the conveyor.The field curvature blur of line camera was restored by means of space variant blur matrix,and the image distortion was rectified by Zhang Zhengyou calibration method.A threshold segmenting method which uses the difference of R and G component was proposed.The segmentation method which is insensitive to noise and illumination can segment the leaf image form background accurately.An efficient method of leaf area accumulating was proposed to replace the image stitching and the problem of single leaf partitioned in two images was solved.The parallel image processing software was written by using multi-threading technology.The repeatability tests results show that the image processing speed reaches to 22 frame in each second,the relative error is less than 0.50% and the coefficient of variation is less than 1.01% which is in the 100~2 500 mm2 measuring range.更多还原