基于K-L变换和模糊集理论的彩色字符图像分割

根据彩色印刷字符图像的特点，在Lab颜色空间下提取a分量，将彩色图像转换为灰度图像。根据模糊逻辑和阈值分割方法将图像分为目标区域、背景区域以及模糊区域。用K-L变换组合邻域的区域隶属信息和灰度信息，将灰度域换成模糊域，在该模糊域上进行分割。经实践,该算法在工业环境中对复杂背景的彩色印刷图像可以得到较好的分割效果，其时间复杂度不高于传统的阈值分割算法，并且在分割的精确度上要优于传统的阈值分割算法。

Color printing character image segmentation based on K-L transform and fuzzy set theory

According to the features of color printing character images, a* channel of Lab* color space was selected for the further process. Using fuzzy logic method and threshold segmentation, the image could be divided into three regions: target region, background region, and fuzzy region. Combined with neighborhood membership information and gray information by K-L transaction, the gray level field was changed into fuzzy field in fuzzy region. Segmentation could be done based on fuzzy field. The proposed algorithm can get better segmentation for the color printing images in the industrial environment, the time complexity of the presented algorithm is not higher than the traditional threshold segmentation algorithm, and the segmentation accuracy is superior to the traditional algorithm.