表面贴装LED全自动编带机视觉检测系统研制

发布时间：2018-05-02 02:50

本文选题：编带机 + 视觉检测　；参考：《西安工业大学》2013年硕士论文

【摘要】：编带机是将半导体芯片编入载带以方便后续设备对芯片的处理的自动化设备。为保证编入载带的芯片质量,编带机视觉检测系统要以快速度和更高精度来捕捉到各种缺陷,并自动分类,且利用新的理论和分析技术来检测各种缺陷,因此,视觉检测系统是编带机非常重要的组成部分。传统的全自动编带机视觉检测系统多使用国外通用检测系统,存在成本高、缺乏核心技术、操作复杂等问题,因此研制拥有自主知识产权的表面贴装LED全自动编带机视觉检测系统具有非常重要的现实意义。本文设计了一种编带机视觉检测系统,本设计首先对表面贴片LED进行了需求分析,确定了视觉检测系统所要完成的工作,并对编带机视觉检测系统进行了总体结构设计。其次根据需求分析完成了编带机视觉检测系统的硬件选型,主要是CCD摄像机、图像采集卡及摄相机镜头的选型。最后根据编带机视觉检测系统所要完成的工作确定并实现了各个检测项目的算法。本系统主要完成了贴片LED的方向及缺失检测,在进行缺失或方向检测前首先对图像进行了预处理,图像预处理就是要对图像受到的外界干扰进行消除,以求达到最好的检测结果。编带机工作过程中主要会受到椒盐噪声和脉冲噪声的干扰。经过对比实验,本文采用了中值滤波算法对图像进行预处理。其次是对图像进行位置补正,算法是通过指定补正窗口,使补正窗口的位置偏移数据自动反映在其它检测范围。再次是对贴片LED的方向及缺失进行检测,方向检测采用的算法是对二值化后的图像进行面积检测,统计贴片LED各个角白色像素或黑色像素的面积,通过对比各个角的面积偏差找到其缺角位置,从而确定贴片LED的方向。缺失检测采用的算法是对图像进行明度检测,即检测载带里有料时和无料时的明度值,通过比较载带里有料和无料时的明度偏差值判断载带里是否有料缺失。最后提出了采用MeanShift法首先对0603型绿色贴片LED图像进行分割,然后提取其边缘特征,通过判断芯片的内部引脚个数的方法检测其方向。
[Abstract]:Braiding machine is an automatic device that integrates semiconductor chip into carrier band to facilitate the processing of the chip by subsequent equipment. In order to ensure the quality of the chip which is programmed to the tape, the vision inspection system of the braiding machine should capture all kinds of defects with high speed and higher precision, and automatically classify them, and use new theory and analysis technology to detect the defects, so, Visual inspection system is a very important part of braiding machine. The traditional vision inspection system of automatic braiding machine mostly uses foreign universal inspection system, which has many problems, such as high cost, lack of core technology, complicated operation and so on. Therefore, it is of great practical significance to develop a visual inspection system of LED automatic taping machine with independent intellectual property rights. In this paper, a kind of vision inspection system for braiding machine is designed. Firstly, the requirements of the surface patch LED are analyzed, and the work to be accomplished is determined, and the overall structure of the vision detection system of the braiding machine is designed. Secondly, according to the requirement analysis, the hardware selection of the vision detection system of the braiding machine is completed, mainly the selection of the CCD camera, the image acquisition card and the camera lens. Finally, according to the work of the Tape Machine Visual Inspection system, the algorithm of each detection item is determined and realized. The system mainly completes the orientation and deletion detection of the patch LED. The image is preprocessed before the deletion or direction detection. The image preprocessing is to eliminate the external interference of the image. In order to achieve the best test results. Salt and pepper noise and pulse noise will interfere with the working process of the braiding machine. Through the contrast experiment, the median filter algorithm is used to preprocess the image. The second is to correct the position of the image, the algorithm is to specify the correction window, so that the position offset data of the correction window can be automatically reflected in other detection areas. The third is to detect the orientation and missing of patch LED. The algorithm of direction detection is to detect the area of binary image, and to calculate the area of white or black pixels in each angle of patch LED. By comparing the area deviation of each angle to find the position of the missing angle, the orientation of the patch LED is determined. The algorithm used in the missing detection is to detect the brightness of the image, that is, to detect the brightness of the material in the load band when there is material and when there is no material, and to judge whether the material is missing in the load band by comparing the deviation value of the brightness between the material in the load band and the material without material. Finally, the MeanShift method is proposed to segment the 0603 green patch LED image first, then extract the edge features of the image, and detect the direction by judging the number of internal pins of the chip.
【学位授予单位】：西安工业大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP391.41

【参考文献】