摘要
清分机在高速采集纸币图像过程中,送入的纸币会不可避免地存在各种倾斜,导致采集的纸币图像存在不同程度倾斜。在对纸币进行识别之前,为不影响图像的识别效果,需要对采集获得的图像进行倾斜度校正预处理。本文研究一种首先利用Sobel算子找到纸币的边缘界线,再利用Hough变换方法计算出纸币边缘线倾斜角,进而完成对纸币倾斜度校正的方法。实验证明了本方法的有效性。
It is necessary to preprocess the banknote image before it is recognized. Besides, it is inevita- ble to cause skew when banknote quick feeding in image collection, which makes it difficult for the following image segmentation and recognition, and it is an important link to correct the banknote tilt. In this paper, a kind of method is proposed, in which the Sobel operator is used to pick up the edge of the incline RMB image and then a Hough transform is used to calculate its angle. The results verify that the tilt angle can be corrected accurately.
出处
《辽宁科技大学学报》
CAS
2012年第3期281-284,共4页
Journal of University of Science and Technology Liaoning