An Algorithm for High-precision Edge Location of Cannon Barrel Image Based on Human Visual Characteristics
Hongzhao Jiang, Baojun Han, Jinchun Fu, LangNing Huang · 2005
A novel algorithm for high-precision edge location based on human visual properties is proposed considering characteristics of cannon barrel images, which improves the observation effect of spying bore images and diminishes the target image ambiguity caused by background and noise. The variable gray scale area method is applied according to the image feature diversity between the target and background/noise. The accuracy of edge detection and location is higher than 0.01 pixel when being applied for engineering. The imperfect images are correspondingly enhanced, and the target is thus shown clearly, which is convenient for observation by human eyes and actual measurement. The algorithm is more general-purpose and adaptive to different barrel image processing, and it can significantly inhibit noise for its glossy and enhanced effect.