Non-contact detection of yarn tension based on embedded system
In order to improve the accuracy of yarn tension detection and avoid the contact error caused by contact measurement,a non-contact real-time yarn tension detection method based on machine vision and embedded system was proposed.Firstly,in order to ensure both the speed of image processing algorithm and the accuracy of yarn tension detection in real-time system,a fast feature extraction algorithm based on threshold segmentation was proposed.The image dimension reduction and edge feature extraction were realized at the same time,which can extract the moving yarn waveform in real time and reduce the interference of the external environment on the yarn image.Secondly,an image acquisition system composed of high-speed linear array camera and FPGA high-speed storage system and a detection system composed of image processor with TMS320C6678 as the main chip were built.Through real-time acquisition of images of moving yarns,yarn vibration frequency was extracted from yarn vibration waveform in unit period,and real-time tension of yarns was calculated by using string vibration model.The experimental results show that the error between the detection value of the non-contact yarn detection system based on embedded machine vision and the resistance sensor was about 1.5%,which meets the control requirements of the real-time system.
万方数据
维普
