基于线结构光的发动机缸盖表面字符识别

Character Recognition of Engine Cylinder Head Surface Based on Linear Structured Light

为解决发动机缸盖因油渍较多导致表面字符拍不清和识别难的问题,设计出一套基于线结构光的缸盖表面字符识别系统。该系统使用线结构光扫描字符表面,结合激光三角测量法得到缸盖表面字符的深度图,并对该深度图进行图像降噪、图像增强和倾斜校正处理;最后,使用YOLOv4算法进行训练并识别出字符。对缸盖表面油渍较多情况下所获得的图像(相邻字符之间粘连严重)识别字符平均正确率能达到99.6%以上,识别时间约为50 ms。实验结果表明所设计的系统在生产节拍内可准确识别出缸盖的型号和铸件号等信息,具有抗干扰性强、识别准确率高等特点。

In order to solve the problem that the engine cylinder head is not clear and difficult to recognize because of the oil stain,this paper designs a set of cylinder head surface character recognition system based on the line structure light.The system uses linear structured light to scan the character surface and laser triangulation to obtain the depth map of the characters on the cylinder head surface.The depth map is processed with image noise reduction,image enhancement and tilt correction.Finally,the YOLOv4 algorithm was used to train and recognize the characters.For the image obtained under the condition of more oil stains on the cylinder head surface(serious adhesion between adjacent characters),the average character recognition accuracy can reach more than 99.6%,and the recognition time is about 50 ms.The experimental results show that the designed system can accurately identify the cylinder head model and casting number in the production beat,and has the characteristics of strong anti-interference and high identification accuracy.