基于引力定律与移动准则的图像边缘检测算法

Image edge detection algorithm based on law of universal gravitation and moving criterion

针对当前边缘检测算法的边缘定位低,在同质区域难以精确检测图像边缘细节等问题,提出一种基于万有引力定律与探测器移动准则的图像边缘检测算法。基于万有引力定律,建立图像空间模型;根据空间模型,设计基于万有引力的边缘检测模型的探测器,分别计算探测器上的力、速度、变异算子、位移以及引力势,建立探测器移动准则,根据移动准则,更新探测器位置和引力势能矩阵,精确跟踪边缘像素;设置迭代停止条件,通过多次迭代完成图像边缘检测。实验结果表明,与目前常用的边缘检测算法相比较,在视觉感知和定量评价的比较中,该算法具有更高的边缘定位精度和噪声抑制能力,能够有效对同质区域边缘完成检测。

In view of the low edge location accuracy of the traditional edge detection algorithm and that using the traditional algorithm is difficult to effectively detect the edge details in homogeneous regions,an image edge detection algorithm based on adaptive law of universal gravitation was proposed.Based on the law of universal gravitation,the image space model was established and initialized.The detector of edge detection model was introduced,according to the model,force,velocity,mutation operator,displacement and gravitational potential of the detector were calculated respectively,and the principle of the detector movement was established.The position of the detector and the potential energy matrix were updated according to the moving criterion.The image edge detection was done through several iterations according to the iterative stopping condition.Experimental results show that compared with the commonly used edge detection algorithm,in the comparison of visual perception and quantitative evaluation,the proposed algorithm has better edge location accuracy and noise suppression ability.It can be effective for the homogeneous region edge detection.