一种基于中值滤波的局部阈值分割算法

在实际的工业现场,光照环境无法非常均匀,在部分位置可能会出现目标区域与背景无法分割,这就给后期的BLOB分析带来了非常大的困难.针对这种光照不均匀、背景复杂的情况,提出了这种基于中值滤波的局部阈值分割算法.在得到原始图像之后,使用中值滤波得到一个大于原始图像宽度两倍的掩膜,再用原始图像与掩膜图像做差得到一个区域,最后选择一个合适的阈值对这个区域进行分水岭阈值分割,就可以得到期望的区域,为之后的连通性分析打下基础.通过在halcon中的实验结果可以说明,提出的算法优于全局Otsu分割算法和分水岭分割算法,更加适用于缺陷检测,字符识别等工业应用的现场.

Local threshold segmentation algorithm based on median filter

In the industrial scene, the light environment usually can not be uniform.It may appears background can not be divided in some locations, which Makes it difficult for later BLOB analysis.For this kind of situation, this paper puts forward the local threshold segmentation algorithm based on median filter.After getting the original image, we can get the mask which is twice as much as a width greater than the original image using median filter.Then, Original image minus the mask image to get an area.Finally, segment the area choosing a suitable global threshold.The region is the desired area.Through the experimental results in Halcon, shows that the proposed algorithm is better than the global Otsu segmentation algorithm and watershed segmentation algorithm, which is more suitable for defect detection, character recognition and other industrial applications.