The computer case assembly has lots of features such as many positions and complicated standard component types, thus extracting great presentation image feature is the key of computer case assembly quality inspection in machine vision. The image feature extraction method for standard component of computer case based on multi-corner combination was proposed in this paper. The three kinds of standard component image characteristics and corner features of computer case were analyzed, and the features of standard components of cover computer case with excellent corner features were pointed out. Then, extraction mechanism and determination condition of Harris, Shi-Tomasi, and Fast corner features were analyzed, and three methods for feature extraction of standard image of computer case were proposed. The results show that the algorithm can identify the error corner feature in the background as false corner feature and distinguishing feature of standard components of computer case as true corner feature, which can avoid the problem that corner features are wrongly extracted under the influence of interference.%机箱装配具有部位多、标准件类型复杂特点,采用机器视觉进行机箱装配质量检测,提取具有较强表示能力图像特征是关键,该文提出基于多角点结合的机箱标准件图像特征提取方法.首先,分析机箱3大类标准件图像特性及角点特征,指出角点特征较好地覆盖机箱标准件特性;然后,分析Harris、Shi-Tomasi、Fast方法角点特征提取机理与判定条件,并结合3种方法提出多角点结合的机箱标准件图像特征提取方法.结果表明:该方法能将背景中错误角点特征识别为伪角点特征,机箱标准件显著特征识别为真角点特征,较好地解决受干扰影响错误提取角点特征问题.
展开▼
