Accurate and efficient dimensional measurement technique based on point cloud models and its application in the automotive industry

The rapid advancement in the modern automotive industry has put forward higher requirements for the quality of various parts. This paper explores automated dimensional measurement technique based on point cloud models and develops a novel digital inspection system to enhance measurement accuracy and efficiency. By working out specialized inspection schemes, creating local cross-sections at the corresponding inspection positions on the scanned point cloud model of the part to be measured, and adopting geometric analysis algorithms, such as random sample consensus and progressive iterative optimization, the proposed digital inspection method constructs key points for measuring dimensions. This process combines heuristic information from the design or theoretical model with local features of the actual part, enabling rapid and accurate acquisition of inspection results and errors. Besides, to meet the actual production needs for achieving efficient product inspection and quality evaluation, an open system development framework has been designed. It integrates independently developed core functionalities for collecting, sampling point cloud data, performing dimensional measurement and error analysis, with auxiliary functions developed on the basis of commercial software by applying automated scripts for the preprocessing of scanned data and visualizing dimensional measurement and error analysis results. Experimental and application results on actual components demonstrate that, compared to traditional inspection methods, the digital virtual inspection method maintains high inspection accuracy, enhances the overall intelligence level of quality evaluation, and significantly shortens measurement time. Additionally, it exhibits universality and is especially suitable for dimensional measurement where traditional methods are not applicable.