Abstract: When the silkworm cocoons count, in the face of cocoon adhesion, the common method of swelling and corrosion can not separate the silkworm cocoons. In this case, a method of segmentation of silkworm cocoons based on distance transformation is proposed. It processes, and then uses the maximum inter-class variance method to binarize the image, then transforms it, normalizes it, and performs thresholding binary segmentation, then uses morphological expansion processing, and finally performs connected domain labeling. The number of connected domains is statistically analyzed. The experimental results show that the algorithm can effectively segment the cocoons, and can accurately count the cocoons, and the counting accuracy is over 99%.

Key words: segmentation of silkworm cocoons, image sharpening, Otsu, distance transformation

摘要： 蚕茧计数时，面对蚕茧粘连情况，常见的膨胀腐蚀方法并不能很好地将蚕茧分割开来，针对这一情况，提出了一种基于距离变换的粘连蚕茧分割方法。对蚕茧图像进行锐化处理，利用最大类间方差法对图像二值化，对其进行距离变换，再进行归一化，同时进行阈值化二值分割，采用形态学膨胀处理，进行连通域标记，统计连通域个数。实验结果表明，该算法能有效地分割粘连的蚕茧，能够实现对蚕茧的准确计数，计数准确率达99%以上。

关键词: 蚕茧分割, 图像锐化, 最大类间方差法, 距离变换