论文部分内容阅读
为提高电路板热红外图像分割效果,解决电路板故障诊断的问题,提出了一种将自适应遗传算法和改进二维OTSU算法相结合的新算法。自适应遗传算法采取快速排序法,利用遗传代数的变化改进交叉、变异概率;根据邻域像素点与中心像素点之间的距离对二维OTSU邻域均值算法的比例系数进行加权,判断噪声点并对热红外图像进行降噪处理;引入类内方差法,对阈值进行优化。利用了遗传算法的并行性和强大的空间搜索能力,提高了二维OTSU的阈值查找速度,提高了热红外图像的分割效率。实验结果表明,该算法提高了热红外图像的分割准确度,有一定的应用前景。
In order to improve the thermal infrared image segmentation effect of the circuit board and solve the problem of circuit board fault diagnosis, a new algorithm combining the adaptive genetic algorithm and the improved two-dimensional OTSU algorithm is proposed. The adaptive genetic algorithm adopts the fast ranking method to improve the crossover and mutation probability by utilizing the changes of the genetic algebra. The proportion coefficients of the two-dimensional OTSU neighborhood mean value algorithm are weighted according to the distance between the neighborhood pixel and the central pixel, and the noise point And the thermal infrared image is denoised; the class variance method is introduced to optimize the threshold. Utilizing the parallelism and powerful spatial search ability of genetic algorithm, the threshold search speed of two-dimensional OTSU is improved and the segmentation efficiency of thermal infrared images is improved. Experimental results show that the algorithm improves the accuracy of thermal infrared image segmentation and has certain application prospects.