畸变图像的目标区域自动提取及校正算法研究

摘要

The research on the algorithms of perspective distortion is increasing, which not only improves the reliability of the algorithms, but also improves the feasibility of the experiment.However, there are not many researches on how to extract the useful information of a distorted image.In this paper, a new algorithm is proposed to combine Hough transform and perspective transform.A useful target image region can be automatically selected and extracted from the distorted image, and the image is corrected and restored by perspective transformation.MATLAB simulation results show that this method has high precision, high fault tolerance and good stability, and can be widely used in pattern recognition.%研究畸变图像处理的方法较多, 这些方法提高了算法可靠性、完善了算法可行性, 但如何提取畸变图像的有用信息进行再校正的理论和方法却不多.为此提出一个新的算法, 结合Hough变换和透视变换, 在畸变图像中自动选取有用的目标图像区域并提取出来, 最后通过透视变换将其进行图像校正与还原.MATLAB仿真实验结果表明, 该方法精度高、容错性高、稳定性好, 可广泛应用到模式识别等领域.