基于直方图均衡化的Robinson图像边缘检测算法

针对图像边缘检测,现有Robinson相关算法存在效率低、阈值设定随机性大、易出现伪边缘等问题,提出一种改进Robinson的图像边缘检测算法。该算法利用直方图均衡化对图像进行增强,然后将Robinson算子原有的八方向梯度依照两两垂直原则组合为八组,分别计算每组梯度的范数,并取其最大值作为该像素点的梯度。最后取整幅图像的灰度均值作为阈值来识别图像的边缘像素和背景像素。实验表明,相对于现有相关算法,该算法检测结果更加清晰完整,同时避免了传统算子人为设定阈值随机性大的问题。

Edge Detection Algorithm Robinson Based on Histogram Equalization

Aiming at the defects of low efficiency, high randomicity of manual threshold and the false edge of the existing Robinson related algorithms, this paper proposes an edge detection algorithm based on the improved Robinson. It adopts histogram equalization to enhance the image. Then it groups the traditional eight direction gradients of single pixel point into eight groups in accordance with the paired vertical principle, calculates the norm of each gradient group respectively, and takes the maximum as the gradient of the pixel. Finally, it takes the mean gray value of the whole image as the threshold to identify the edge pixels and background pixels. The experimental results show that, comparing with the existing related algorithms, the detection result of improved algorithm is more complete and clearer. In addition, it avoids high randomicity of traditional manual threshold.