本平台为互联网非涉密平台,严禁处理、传输国家秘密或工作秘密

卷烟小盒外观缺陷检测设备成像质量评价方法

Imaging quality evaluation method for cigarette packet appearance quality inspection apparatus

  • 摘要: 针对不同卷烟小盒外观缺陷检测设备间检测能力存在差异等问题,建立了一种烟盒外观缺陷检测设备成像质量评价方法。搭建了烟盒外观缺陷离线检测系统以构建不同成像环境,制作了白度和棋盘格标准块并建立了系统光照均匀性和成像畸变率的量化方法,设计了标准缺陷烟盒样本并在不同成像环境下对其进行检测,分别考察光照均匀性、成像畸变率、成像清晰度以及图像分辨率对缺陷识别率的影响。结果表明:成像畸变率对缺陷识别率影响不大,光照均匀性、成像清晰度以及图像分辨率与缺陷识别率呈正相关;为保证检测效率和识别准确性,检测系统中各参数阈值为光照均匀性系数<5.0%,成像清晰度T值>12.0,图像分辨率≥128 px×128 px。该方法可为提升卷烟产品包装品质提供支持。

     

    Abstract: To verify the inspection performance of various appearance quality inspection apparatuses for cigarette packets, an evaluation method for the imaging quality of the interested inspection apparatuses was established. An off-line appearance inspection system was configured to create different environments of image formation, methods for quantifying the illumination uniformity and imaging distortion rate of the system were established, the whiteness and checkerboard standard blocks were fabricated as well. A standard cigarette packet sample with appearance defects was designed and tested in different environments of image formation. The effects of illumination uniformity, imaging distortion rate, imaging clarity, and image resolution on the defect recognition rate were investigated. The results showed that the imaging distortion rate did not affect defect recognition rate significantly, however the illumination uniformity, imaging clarity, and image resolution were positively correlated with the defect recognition rate. To ensure detection efficiency and recognition accuracy, the recommended thresholds of the system are illumination uniformity coefficient < 5.0%, imaging clarity value T > 12.0, and image resolution≥128 px×128 px. This method provides support for the improvement of cigarette packing quality.

     

/

返回文章
返回