目的 解决传统亮度、色彩等检测方法不能有效检测出烟包(尤其背景为白色的烟包)有无油封包裹的问题。方法 在传统检测方法的基础上，增加基于灰度共生矩阵算法的检测方法进行散包检测，通过对新型散包检测硬件架构设计、硬件隔离设计、相机优化、数学建模、灰度共生矩阵特征值提取、最优解等处理，采用基于灰度共生矩阵算法的新型散包检测方法能实现烟包散包检测精度的提升，有效检测出烟包(尤其浅色烟包)有无油封包裹的情况，降低无油封包裹烟包流入下道工序的风险。结果 基于灰度共生矩阵算法的新型散包检测方法能有效提升烟包的散包检测精度，散包烟包的检出剔除率≥99.9%。结论 基于灰度共生矩阵算法的新型散包检测方法能有效提升烟包的散包检测精度，尤其对浅色烟包有无油封包裹情况效果提升明显。

The work aims to solve the problem that traditional brightness, color and other detection methods can not effectively detect whether the cigarette pack (especially the white background of the cigarette pack) is oil-sealed.On the basis of traditional detection methods, the detection method based on gray scale co-occurrence matrix algorithm was added for bulk pack detection. Through the novel hardware architecture design of bulk packet detection, hardware isolation design, camera optimization, mathematical modeling, eigenvalue extraction of gray scale co-occurrence matrix, optimal solution and other processing, the novel bulk pack detection method based on gray scale co-occurrence matrix algorithm could improve the detection accuracy of cigarette pack and effectively detect whether the cigarette pack (especially the light-colored cigarette pack) was oil-sealed, and reduce the flow of unsealed packs into the next process.The novel detection method based on gray scale co-occurrence matrix algorithm could effectively improve the detection accuracy of cigarette pack, and the detection and rejection rate of cigarette pack was ≥99.9%. The novel detection method based on gray scale co-occurrence matrix algorithm can effectively improve the detection accuracy of cigarette pack, especially for light-colored cigarette pack with or without oil seal.