基于曲率约束和改进Hausdorff距离 匹配的轮毂识别算法

Wheel Indentify Algorithm Based on Curvature Restriction and Modified Hausdorff Distance Matching

  • 摘要: 针对在轮毂的自动化生产过程中对轮毂型号自动识别的需求,提出一种基于曲率的轮毂辐孔轮廓线角点特征提取和改进Hausdorff距离(MHD)轮毂型号自动匹配的方法. 对待识别的轮毂型号样本进行特征提取并建立标准样本库,在对实时采集的轮毂图像预处理后,利用曲率约束的最小二乘法(CRLSM)进行轮毂外轮廓提取,并用曲率阈值进行角点特征提取,进而采用改进的Hausdorff距离匹配实现轮毂自动识别. 仿真实验证明,该识别算法能够对建立样本库的轮毂类型进行正确识别.

     

    Abstract: For the automatic recognition of wheel hub type during automated production process, the corner feature extraction based on wheel hub curvature and modified Hausdorff distance (MHD) automatic matching method is proposed. First, the wheel hub sample features are used, and is established the standard sample database, after pre-processing the real-time wheel hub, curvature constrained least squares method(CRLSM)is used to extract the wheel contour, and extract the corner feature by curvature thresholds, and the MHD matching method is used to realize the wheel automatic recognition. The simulation experiments show that the proposed algorithm can identify the wheel hub types correctly for the sample database.

     

/

返回文章
返回
Baidu
map