【作者】 陈西广；

【导师】 徐乐年；

【作者基本信息】 山东科技大学 ， 测试计量技术及仪器， 2008， 硕士

【摘要】 目前啤酒生产企业越来越重视产品质量,越来越多的企业在罐装生产线上使用了各种自动检测设备,以期提高检测的可靠性。传统的罐装线空瓶检验方式是人工荧光灯箱检测。但在现代高速罐装线上,由于视觉疲劳等原因,仅靠传统的人工荧光灯箱检测不能保证可靠性,而且啤酒瓶的两个重要部位一瓶口和瓶底极易出现漏检的情况,造成部分漏气瓶和瓶底赃物瓶流入市场,给企业形象造成损害,在国外“对清洗之后的空瓶进行严格而精确的检测已经成为一项十分重要的规范”,欧洲的“HACCP”标准已经是啤酒生产企业必须履行的义务。目前国外基于图像处理技术的空瓶在线检测系统已相对成熟且有大量产品投放市场,如:德国的HEUFT、KRONES、Miho、和日本KTS等。我国在这方面则比较落后,现在尚无完善可靠的系统投入市场。本课题主要对空瓶检测机涉及的图像处理的软硬件部分进行了以下几个方面的研究:图像采集设备的选型,为保证在高速运行的生产线上瞬间完成对锁定目标的成像,选用高分辨率、高对比度和高帧速率的面阵CCD相机。相机工作在触发模式和异步重置方式,可以在1/60秒的时间内完成全象素的输出。结合高质量的光学照明技术,可以做到对啤酒瓶瓶口、瓶底的清晰成像,为后续的图像处理和分析打下良好的基础;图像处理分为图像预处理、边缘检测、二值化处理、腐蚀和膨胀、轮廓跟踪算法和图像定位六步。采用中值滤波滤除噪声、采取增益补偿增强图像,边缘检测采用了Canny算子方法,可以得到清晰的连续边缘,采用采用灰度扫描方法得到图像边界,用最小二乘法拟和得到圆心和半径,采用连通域搜索算法检测缺损,综合决策去除非污点干扰。瓶口和瓶底缺陷检测。瓶口和瓶底图像的准确定位是检测缺陷的前提条件,根据瓶口圆环边缘比较清晰,干扰边缘比较少的特点,对瓶口圆心的定位采用了边缘检测结合最小二乘法拟合圆的方法来得到圆心和半径。瓶底圆环边缘不明显干扰边缘比较多且存在防滑纹的影响,通过检测防滑纹的内边缘灰度突变点来检测边界,通过最小二乘法拟合圆心和半径,对于瓶底通过检测是否有超过限定长度的黑色圆周或白色圆周来判断瓶底是否合格。更多还原

【Abstract】 At present the beer enterprises think more and more of the product quality, and more and more enterprises use various automatic detection devices on the filling production line to improve the reliability of inspection. The traditional Empty Bottle Inspection method on filling line is human light detection, but on the modern high-speed filling line , human eyes have series of recessive in the detection link unavoidably, such as tiredness. Therefore, just rely on the traditional manual light detection can not guarantee the reliability of the product, especially the two important parts of the beer bottles-finish and base. It can lead to part blow-by beer and base sully beer inflow the market, and damage the enterprise figure. In aboard "detecting the cleaned empty bottles strictly and accurately has become a very important regulation" , European "HACCP" standard has been the obligation must to be performed by the beer enterprises. Now In aboard Empty Bottle Inspector based on machine vision technology has been mature and entering into the market largely, such as HEUFTHEUFT、KRONES、Miho and KTS. There are no mature equipments in our country.The key image processing technologies about empty bottle inspector researched in this paper are as following:The type electing of image gathering equipment. To guarantee getting image from the locked target on the high speed movement’s production line, high resolution、high contrast gradient and the high frame speed rate area array CCD cameras are Selected. The camera works in trigger and asynchronous reset mode and can provide all-pixel-data readout in less than 1/60 seconds. The camera can grasp high quality images of the beer bottle finish and base for sequential processes at the presence of good optics illumination.Inspecting the defects of bottle finishs and bases. The image localization, the key of bottle finish and bottle base flaw inspection lies in the localization of the circle center. The ring edge of the bottle finish is clear and the interfered edge is few, for the localization of the bottle finish circle center, the edge detection with least squares analysis curve fitting is used to search the circle center and the radius. Because the ring edge of the bottle base is not obvious and has knurlings for the localization of the bottle base, circle center the edge detection with finding the gray points changing rapidly locating inside of the knurlings is used. The edge detection with least squares analysis curve fitting is used to search the circle center and the radius. For bottle bases, the quality of the bottles is determined by the length of the white or black circumference.更多还原