摘要
针对矿仓中矿石块度尺寸不一、光照不匀而导致图像分割失真的问题,提出了一种基于分块处理的矿石图像多阈值二值化算法。该算法对矿石图像做全局自适应HDR与双边滤波处理,均等分块并应用提出的多阈值算法分割,根据信息熵分布将子图分类,分别应用Ostu、改进Niblack与加权阈值方法分割,将子图二值图像合并后进行形态学处理,得到最终的分割结果,并将该算法与传统算法进行比较,结果表明,新算法错分割率低,二值化效果明显改善,准确率提升了50%。
To solve the problem of image segmentation distortion caused by different sizes and uneven illumination of ore blocks in the ore silo,a multi-threshold binarization algorithm of ore image based on block processing was proposed.The algorithm performs global adaptive HDR and bilateral filtering on the ore image,divides it into equal blocks,and applies the proposed multi-threshold algorithm to segment.The sub-graphs are classified according to the information entropy distribution.What’s more,Ostu,improved Niblack and weighted threshold method are used for segmentation.The final segmentation result is obtained by combining the binary images of the subgraph and morphological processing.The experimental results show that the new algorithm has a low misclassification rate,and improve the binarization effect and the accuracy by 50%.
作者
蔡改贫
汪龙
罗小燕
姜志宏
CAI Gaipin;WANG Long;LUO Xiaoyan;JIANG Zhihong(Jiangxi University of Science and Technology,Ganzhou,Jiangxi 341000,China)
出处
《矿业研究与开发》
CAS
北大核心
2020年第12期153-157,共5页
Mining Research and Development
基金
国家自然科学基金项目(51464017)。
