This paper presents a VIN (Vehicle Identification Number) recognition system based on machine vision. By capturing the VIN code on the car through the camera, the management system of unlicensed merchandise car garage is realized. The system achieves locally adaptive non-linear interference detection through improved threshold partition and morphological change pretreatment methods. Experiments show that the algorithm can be well adapted to dynamic environment and effectively improve the recognition accuracy. The algorithm fully considers the problems of backlighting and low illumination, and is simple in calculation and short in processing time. It is suitable for embedded devices with low computational speed.
引用
收藏
页码:368 / 371
页数:4
相关论文
共 5 条
[1]
Bulan O, 2015, IEEE COMP SOC C COMP, P758
[2]
Cordel MO II, 2017, INT J KNOWL-BASED IN, V21, P15, DOI 10.3233/KES-160343
[3]
Dash SS, 2016, INT C ART INT EV COM, V517, P653