基于改进的Canny算子图像信息提取算法研究

Improved Canny Operator Image Information Extraction Algorithm

研究文本定位与特征提取问题。针对传统的Canny算子图像检测算法的不足,提出了一种改进的Canny算子图像信息特征提取算法。研究方法是:首先对彩色图像进行高斯金字塔分解,然后用Canny算子检测彩色图像,提取边缘图像,再经过通二值化,去噪方差投影定位文本区域。实验结果表明,本文提出的方法有效、实用。

the study of text location and feature extraction. According to the traditional Canny operator image detection algorithms, this paper proposes an improved Canny operator image feature extraction algorithm. Method based on color image by Gauss Pyramid decomposition, and then use the Canny operator to detect color image edge extraction, image, and then passes the two values, noise variance projection to locate text area, the experimental results indicate that the proposed method is effective and practical.