CCD视觉检测系统的整体标定

Global calibration method for CCD based visual inspective system

为了实现粒度分析仪对复合肥粒度分布的准确检测,提出了CCD视觉检测系统的整体标定方法。首先,分析了检测系统的检测精度;对光源进行了伺服控制,以减小光源的不稳定性、不均匀性对视觉检测系统检测精度的影响。然后,确定CCD合适的扫描速度,还原了被检测物体的真实形状。最后,基于直纹曲面理论,通过对不同直径的钢球进行多次标定,获得其像素直径和像素坐标,并借助Excel、Matrox Imaging Library(MIL)等工具建立了直纹曲面的数学模型,实现了对CCD视觉检测系统的整体标定。实验结果表明,整体标定以后,CCD视觉检测系统响应的非均匀性得到了明显改善,检测误差在±3%以内,达到了粒度分析仪的设计精度要求。

To detect the size distribution of compound fertilizers accurately,an overall calibration method was proposed for the CCD based visual detective system of a particle size analyzer.Firstly,the detection accuracy of the inspective system was analyzed and the light source was controlled in servo to reduce the effect of instability and asymmetry of the light source on the detection accuracy of the visual detective system,Then,an appropriate scanning speed of CCD was determined to insure the real shape of the object to be detected.Finally,based on the theory of the ruled surface,the pixel diameters and pixel coordinates were obtained through measuring the balls with different diameters for several times.The overall calibration of the whole detective system was implemented by using Excel,MIL and other tools to establish mathematical model of ruled surfaces.Experiment results show that the non-uniformity response of the CCD vision inspection system has been significantly improved,and the detection error is within ± 3%.It is simple,intuitive,feasible,and can meet accuracy demand of the particle size analyzer.