遗传算法在易拉罐罐盖喷码系统中的应用

目的为实现饮料易拉罐拉环背部激光打码的自动化,提出一种基于遗传算法的易拉罐罐盖图像识别新方法。方法首先搭建一套易拉罐盖激光自动打码机,基于所搭建的实验系统,利用CCD相机实时采集罐盖图像。对所采集到的图像进行中值滤波和灰度增强处理,在此基础上,研究基于遗传算法的罐盖图像阈值分割新方法,分析、确定算法的关键参数(个体数目、交叉率、变异率等),由此得到罐盖的二值化图像,并对算法处理结果进行误差分析。结果遗传算法经过约15代的迭代计算,能够收敛,获取到最优的图像阈值,整个算法的运行时间约30 ms,最终的图像精度约为7.9 pixel。结论基于遗传算法的图像阈值分割实时性好,分割后的图像精度高,与传统的Ostu阈值分割法相比,得到的信息更加丰厚,能抑制光线不均所造成的图像干扰。同时对遗传算法阈值分割后的图像进行了sobel边缘检测,得到了清晰的罐盖边缘,为激光打码的准确定位奠定了基础。

Application of Genetic Algorithm in the Code Printing System for Aluminium Can Cover

The work aims to propose a new method for the recognition of aluminium can cover image based on the genetic algorithm, so as to realize the automation of laser code printing at the back of aluminium can cover pull ring. A set of automatic laser code printing machine for aluminium can cover was firstly set up. Based on the setup experiment system, images of aluminium can cover were captured by CCD in real time. Then, the median filter and gray enhancement operations for the captured images were performed. On this basis, a new threshold segmentation method for the images of can cover based on the genetic algorithm was researched. Key parameters, such as the total number of individuals, the crossover rate and the variation rate were analyzed and determined. Therefore, the binary image of aluminium can cover was obtained, and the errors of the algorithm processing result were analyzed at the same time. After 15 times'' iterative computations, the genetic algorithm could be convergent and the optimal image threshold was obtained. The running time of the whole algorithm was about 30 ms, and the final image accuracy was about 7.9 pixel. The image threshold segmentation based on genetic algorithm has good real-time performance and the segmented image has high accuracy. Compared with the traditional Ostu threshold segmentation method, the image threshold segmentation method can obtain more information and restrain the image interference produced by the asymmetry of brightness. In the meantime, the sobel edge detection is performed on the images subject to threshold segmentation based on the genetic algorithm. Therefore, the clearer edge image of can cover can be acquired to lay a foundation for the accurate location of the laser code printing.