基于图像分析的籽棉含杂率检测与实验

Detection of and experiment on seed cotton impurity rate based on image analysis

籽棉中杂质的存在是难以避免的,若不及时进行清除,会影响后续的加工效益,也会降低棉制品品质。本文基于图像处理技术,实现了籽棉图像含杂率的检测。运用最大类间方差法和多种算子结合的边缘检测两种方法分别对图像进行分割,得出杂质区域的像素面积占比,作为最终含杂率。同时,为了对比,本文还进行了质量法检测籽棉含杂率的实验,杂质与籽棉的质量比即可作为含杂率。实验结果表明,运用多种算子结合的边缘检测方法分割图像后得出的含杂率稍高于最大类间方差法,但结果比较接近。总体上,两种分割方法均实现了较准确的籽棉图像含杂率检测,不同于质量法的含杂率检测,本文提出的基于图像的含杂率检测方法,解决了棉花加工行业籽棉含杂率在线检测的难题,可为棉花加工过程提供含杂信息。

For the impurity in seed cotton not to be removed in time,the subsequent processing efficiency and the quality of cotton products will be affected.In this paper,the method for detecting impurity ratio in seed cotton image was realized by the image processing technology.The maximum inter-class variance method and edge detection method combining with multiple operators were used to segment the images,with the pixel area ratio of the impurity region obtained as the final impurity ratio.At the same time,in order to compare the results,this paper also carried out the mass method to detect seed cotton impurity ratio experiment,with the mass ratio of impurities in seed cotton used as impurity ratio.The experimental results showed that the impurity ratio obtained by the edge detection method combined with multiple operators was slightly higher than that obtained by the maximum inter-class variance method,but the results were relatively close.Overall,both two image segmentation methods achieved accurate detection of the impurity ratio for seed cotton image.Different from the detection of impurity ratio by mass method,the proposed image-based impurity ratio detection method can cope with online detection for seed cotton impurity ratio in the cotton processing industry,and can provide impurity information for the cotton processing process.