LED字符分割和识别， 在此项目中我们主要完成了对拍摄到的6位LED数字进行分割和识别的课题。识别的数字范围在0~9。由于拍摄环境比较复杂，需要解决复杂背景下对LED字符的分割问题。我们利用了图像的色彩信息，先进行基于色彩的分割，然后对分割结果进行了固定阈值的二值化，并利用投影修正完成了对LED数码管图像的分割，并使用基于模板匹配的算法进行了识别，在匹配过程中对图像进行了尺寸归一化处理。最终得到了较好的识别效果，对10幅样图的识别率达到100％-LED character segmentation and recognition, in this major project, we completed the shooting of the six LED figures for the segmentation and identification issues. Recognition of the range between 0 and 9. As the filming environment more complex, the need to address complex background of the character LED division. We took advantage of the color image information, the first color-based segmentation, and then divided the results of a fixed threshold value of the two values, and to use the projector that completed a pair of LED digital image segmentation possession, and use based on template matching algorithm for the identification, in the process of matching the right image In one size treatment. Finally a better identification effect of 10 samples the recognition rate of 100% 下载
|