Design of Intelligent Measurement System of Vehicle Dimensions Based on Structured Light Imaging and Machine Vision

Abstract

At present, manual measurement is the main method of vehicle dimensions detection, which has some defects such as low efficiency, high cost, low accuracy, and non-standard manual operation. In order to solve these problems, this study proposed an intelligent measurement system of vehicle dimensions based on depth of field reconstruction and machine vision. The design of the depth of field sensor was based on the principle of structured light imaging. The structured light was decoded by two-dimensional cross-correlation technology to acquire the depth information of the vehicle surface. Then, hybrid Gaussian background modeling algorithm was used to extract the foreground signal of the vehicle, and machine vision was used to realize the coarse registration of point cloud information. Finally, vehicle accessories were extracted and removed according to the national standard requirements. Meanwhile, a prototype system was built to measure 964 vehicles with various types, and experimental results showed that the proportion of vehicle dimension error within 10cm is more than 90%, and the error of repeated measurement for the same vehicle is less than 0.5%. In this paper, the intelligent measurement method of vehicle dimensions meets the requirements of GB 21861–2014, and ensures the accuracy and objectivity of the detection process, as well as the fairness and openness of the results.