Research On The Detection Technology Of Charge Under The Detonating Tube Based On Real-time Image Processing

For plastic detonating tube,in order to maintain the normal shockwave transmitting forward,the detonating tube powder on its inner wall must be evenly distributed and the content cannot be too high and too low.In the production of plastic detonating tube,the distance between place where the powder is added and place where the finished products is detected is 10 meters.Traditional detection methods have a great lag and cannot meet the requirements of real-time information feedback which may cause waste of row material and producing time if there is unqualified powder content.In this paper,we use non-contact image processing detection methods to detect the real-time data of powder adding process by using high-frame rate camera.We use the method of stitching two consecutive frames of image to obtain the powder-imaging information contained in each frame of image so that we can calculate the actual content of powder according the area of powder 's image after removing duplicate parts from two consecutive frames.The system uses the ZYNQ 7020 development platform to design the image preprocessing IP core using Vivado HLS tools.Firstly through customized hardware IP core image preprocessing module based on programmable logic device,the image is preproeessed to intercept the sensitive area and transmitted to processing system.Based on the OpenCV library function in ARM,two consecutive frames of images are feature-extracted and match-analyzed by respectively applied SIFT,SURF and FAST algorithm.After comparing the accuracy of feature extraction and matching and the time spent,we choose feature extraction and match based on FAST and ORB.Finally we calculate the amount of powder added per second for the machine according the area of powder image of target.This design adopts mode of software cooperating with hardware.Combining the advantage of parallel high-speed of hardware and flexibility of software,this design can detect the amount of powder added per second for the powder-filling machine with good real-time and stable detection and other advantages.