To real-time identify anomaly small targets on the scraper,ensure the normal and safe operation of the scraper,the anomaly monitoring system of scraper based on machine vision is designed.The industrial camera acquisition unit in the data acquisition layer acquires real-time monitoring images of the scraper based on the principle of machine vision and transmits images to the data processing layer through the universal serial bus(USB)interface.After noise reduction and enhancement of the collected scraper image,the image is uploaded through the Ethernet communication module based on field programmable gate array(FPGA)in the data transmission layer.The anomaly state monitoring module in the data monitoring layer innovatively calls the improved mask region-convolutional neural network(Mask R-CNN)model based on the received images,sends alarm information by the anomaly alarm module,and realizes the anomaly monitoring of the scraper by presenting the anomaly monitoring results and alarm prompt messages through the data display layer.The test results show that:the peak signal-to-noise ratio of the scraper image processed by the system is significantly improved,the root mean square error is significantly reduced;the anomaly recognition loss of the enhanced scraper image is lower.The system can realize the recognition of different types of anomalies in the scraper and mark the abnormal targets.
NETL
NSTL
万方数据
