【摘要】 提出了一种基于内窥镜图像的孔内表面双向反射分布函数（BRDF）测量与建模方法，为实现基于机器视觉的孔内表面缺陷检测提供理论依据。建立内窥成像反射模型，分析内窥图像亮度及其三维形貌关系。以不锈钢和镀锌表面螺纹孔为例，通过图像亮度和法线方向测量孔内表面BRDF值。在混合反射模型基础上建立三种孔内表面反射特性模型，采用遗传算法确定不同材料的最优模型参数并分析模型误差。实验结果表明，基于Cook-Torrance模型的inBRDF模型能够较好地反映孔内表面反射特性，模型计算结果与测量结果吻合度高，两种材料平均拟合误差为6.22%，能够准确描述孔内表面反射特性并可应用于孔内表面缺陷检测及形貌重建。更多还原

【Abstract】 A method for measuring and modeling a bidirectional reflection distribution function(BRDF) of the inner surfaces of holes based on endoscope images is proposed, which provides a theoretical basis for hole inner surface defect detection based on machine vision. In this study, an endoscopic imaging reflection model was first established to analyze the relationship between the endoscopic image brightness and its three-dimensional topography. With stainless steel and galvanized threaded holes used as examples, BRDF values of hole inner surfaces were measured by image brightness in the normal direction. Based on the hybrid reflection model, three types of reflection models of the inner surfaces of holes were established. A genetic algorithm was then used to determine the optimal model parameters of different materials and to analyze the model errors. Experimental results show that the inBRDF model based on the Cook-Torrance model can effectively reveal the reflection characteristics of the inner surfaces of holes, and the calculation results of the model are in good agreement with the measurement results. The average fitting error of the two materials is 6. 22%, which can accurately describe the reflection characteristics of hole inner surfaces and can be applied to the detection and morphological reconstruction of inner surface defects.更多还原