基于深搜算法的手机外壳刮擦痕迹检测

Scratches Detection of Mobile Phone Shell Based on Deep Searching Algorithm

运用形态学滤波算法和自适应二值化算法对手机外壳图像进行预处理,同时提出了一种深搜算法,沿着手机外壳刮痕的方向进行搜索,增加了检测的鲁棒性和敏感性,在具有复杂表面纹理干扰的情况下仍然能够有效地对手机外壳微小刮痕进行检测.经过实验验证,刮痕检测精度可以达到0. 05 mm,漏检率和轮廓检测缺失概率分别为4. 0%和5. 3%.

Morphological filtering algorithm and adaptive binarization algorithm were used to preprocess the mobile phone shell image. At the same time, a deep searching algorithm was proposed,which can search in the direction of scratching the mobile phone shell. It further increases the robustness and sensitivity of detection and effectively detect the tiny scratches of the mobile phone shell even with the interference of the complex surface texture. The experimental results show that the precision of scratch detection can reach 0.05 mm, the missed detection rate and the missing probability of contour detection are 4% and 5.3%.