基于颜色分离与特征统计分析的工件图像表面异物检测算法

Workpiece image surface impurity detection algorithm based on color separation and statistical analysis of the characteristics

为了解决当前工件图像表面异物形态特征不确定且分布随机性,导致检出能力不足或者严误判过高的问题,分别从颜色分离和特征统计分析的角度出发,提出了基于颜色分离与特征统计分析的工件表面异物检测算法。首先,根据棕色与蓝色色彩差异,进行图像对数变换和线性变换,推导出颜色分离方程,达到分离出棕色异物目标的目的。然后,基于统计学分析原理,计算出目标颜色特征,筛除杂质干扰,进一步精确检测异物目标。最后,基于软件开发环境Visual Stadio实现算法,并系统集成。实验测试结果显示,与当前普通异物检测技术相比,该算法拥有更高的准确性与稳定性。

Foreign body in order to solve the current workpiece surface morphology distribution randomness,uncertainty and high detection ability is insufficient or yan misjudgment caused by the problem,this paper respectively from the perspective of color separation characteristics and the statistical analysis,was proposed based on color separation and characteristics of the statistical analysis of surface of workpiece foreign body detection algorithm.First of all,according to brown and blue color difference,image logarithmic transformation and linear transformation,color separation equation is deduced,achieve the goal of isolated brown foreign body target.Then,based on statistical analysis principle,calculate the target color features,screen impurities interference,further accurate foreign body target detection.Finally,the algorithm was test based on the software development environment of Visual Stadium algorithm and system integration.Test results show that compared with current common foreign body detection technology,this algorithm has higher accuracy and stability.