基于Prewitt算子的红外图像边缘检测改进算法

Improved Algorithm for Infrared Image Edge Detection Based on Prewitt Operator

传统Prewitt算法存在人为选取阈值造成边缘结构缺失较粗、产生伪边缘、定位不精准等缺点,为此,提出了一种基于Prewitt算子,结合人眼视觉特性的动态阈值的改进算法。首先增加方向模板数量使边缘检测更加精确,然后结合人眼视觉特性对图像中平坦区进行分割,最后采用动态阈值进行边缘检测,并根据粗边缘特性进行边缘细化。实验结果表明,与经典算法相比,改进算法有效解决了Prewitt算子边缘过粗的缺点,检测出的图像边缘精度提高了6.0%以上,边缘结构更加完整,定位更加精准。

The traditional Prewitt algorithm has some disadvantages such as the missing of edge structure,the appearance of rough edge,false edges and inaccurate positioning caused by selecting threshold artificially.Thus,an improved algorithm for dynamic threshold of the human visual properties based on Prewitt operator was proposed.Firstly,the improved algorithm increases the number of directional templates to make edge detection more accurate.Secondly,this paper splits the flat areas in the image based on the human visual properties.Lastly,detects edge used dynamic threshold,thins edge based on the rough edge features at the same time.Results show that,compared with the classical algorithm,the proposed method can solve the shortcoming of the Prewitt operator rough edge effectively,the precision improved by 6%and detect edge structure more completely,and can position more accurately.