摘要
针对篮球使用的天然皮革质量检测,提出了一种基于视觉显著度的皮革表面缺陷可视化检测框架,利用视觉显著度模型对球皮皮革图像的颜色和亮度特征等进行提取,在计算皮革颜色和亮度特征显著图以后,利用区域生长方法对篮球球皮表面的瑕疵进行定位。与传统皮革表面缺陷检测方法对比后发现,该方法能够有效解决皮革自然纹路对缺陷检测的影响,检测精度与检测速度都能够得到有效保障。
Aiming at the quality detection of natural leather used in basketball, a visual detection framework of leather surface defects based on visual saliency was proposed. The color and brightness features of the leather image were extracted by using the visual saliency model. After calculating the leather color and brightness feature saliency map, the defects on the surface of basketball leather were located by using the region growth method. Compared with the traditional leather surface defect detection method, it is found that this method can effectively solve the influence of the natural texture of leather on the defect detection, and the detection accuracy and detection speed can be effectively guaranteed.
作者
黄一平
HUANG Yi-ping(Chang'an University,Xi'an 710064,China)
出处
《中国皮革》
CAS
2022年第10期36-39,43,共5页
China Leather
基金
国家级大学生创新训练课题(GXX14238)。
