【摘要】 针对木材中存在的表面死结缺陷问题,提出一种基于中智集的图像检测算法。首先将木材彩色RGB图像输入,将输入图像的RGB通道图提取出来,得到R、G、B三个分量的图像。然后选取R通道图像,将该通道图像分别通过T、I、F三种通道的图像处理得到TIF通道的图像。同理,选取G、B通道的图像经过TIF通道处理,得到共9张关于T、I、F通道图像。将所得的TIF通道图像分别进行Otsu、Kapur、主动轮廓三种算法处理,对处理的图像进行jaccard、dice和bfscore三种评价得出JAC、Dice和BFScore三个评价参数,再对各个图像的评价参数进行比较,得出最佳分割图像。更多还原

【Abstract】 In view of the problem of wood surface dead knots,an image detection algorithm based on neutrosophic set has been proposed.Firstly,the color RGB image of wood is input,the RGB channel image of the input image is extracted,and the R,G,B three component images are obtained.Then select the image of R channel,and get the image of TIF channel through the image processing of T,I and f channels.Similarly,if the images of channel g and channel B are selected for TIF channel processing,a total of nine images of Channel T,I and F can be obtained.Then,the TIF channel image is processed by Otsu,Kapur and active contour algorithm respectively.The processed image is evaluated by Jaccard,dice and bfscore,and three evaluation parameters,JAC,Dice and bfscore,are obtained.Then the evaluation parameters of each image are compared to get the best segmentation image.更多还原