Rapid Detection of Characters on Automobile Electronic Gear Lever.

Character display of automotive electronic gear lever is affected by LED lamp beads and transparent cover trim panel. The problems in the manufacturing processes include abnormal color coordinates, abnormal luminance value, and incomplete characters. The algorithm for rapid detection based on CMOS industrial cameras was proposed to replace manual visual inspection and improve abnormal recognition rate and detection efficiency. The color coordinates and luminance model for transient detection of the electronic gear lever were established through the temperature drift. The improved method of circle projection feature was used to achieve recognition. Curvature detection was used for incomplete detection of characters. The effectiveness was verified by experiments. Results show that high and low luminance weight traceability for color coordinate calibration has good consistency, and the luminance and gray values have an exponential relationship after the temperature drift of the characters. The method of circular projection feature has high robustness for the recognition of inclined and reduced character outlines. Curvature detection is effective for incomplete contours. The curvature steps of characters at different positions are six, four, and three, and the threshold is uniformly set to 0.15. The outflow rate of the problem electronic gear lever is reduced, and the productivity is increased by 0.3 percent with this detection method. This study provides a reference for the test platform of other types of electronic gear lever and rapid detection of characters. [ABSTRACT FROM AUTHOR]