基于扫描图像几何畸变的算法研究

Algorithms Based on the Geometric Distortion of Scanned Images

目的为了改进和优化适用于扫描图像的几何畸变校正算法。方法通过分析扫描仪产生几何畸变的原因,总结现有图像畸变校正方法,并与现有的图像处理软件——Photoshop软件相比较,建立合适的数学模型和算法流程。经由图像预处理,利用直方图均衡化法使图像增强,并用凯尼算子进行边缘检测,然后用雷顿算子提取图像轮廓,最后采用连接点法校正畸变图像,并用Matlab仿真加以实现。结果文中算法处理迅速,不需手动处理,校正后平均相似度可达0.95,效果优良。结论预处理后的扫描图像经点连接法进行几何畸变校正,作为一种可以快速处理扫描几何畸变校正方法的补充,算法适用于扫描图像的几何畸变,具有重要的实用价值和研究意义。

Objective To improve and optimize the geometric distortion correction algorithms for scanned images. Methods Through analysis of causes for geometric distortion in the scanner, summarization of the existing methods for image distortion correction, and comparison with the existing image processing software Photoshop, an appropriate mathematical model and algorithm flow were established. Through image preprocessing, images were enhanced by histogram equalization, edges were detected by Canny operator, and then the image contours were extracted by Radon operator, subsequently, image distortion was corrected by dot connection, and finally it was realized by Matlab simulation. Results A large number of experiments verified that this algorithm was suitable for scanned images, the algorithm processed quickly without manual processing. The similarity reached up to 0.95 after correction. Conclusion As a supplement of quick processing method for correction of scanning geometric distortion, this algorithm corrected geometric distortion of the preprocessed scanned images by dot connection, which was suitable for the geometric distortion of scanned images, with important theoretical and practical values.