Study Of The Online Straightener Detection Technology

Straightener plays an important role in manufactural industry. In general, the popularstraightener has only straightening feature in domestic and overseas. The units need to besampled and measured by other equipments off-line after the straightening process. Using thisapproach, it will not only waste the labors, but also cannot guarantee all the units or productsare qualified. Therefore, straightener with on-line testing will be much valuable andsignificant.Based on one given flat, the main methods are Optical collimating, straightedge and soon. And in terms of large work piece straight testing, there are several approaches such asthree coordinating points, lining etc. However, those ways have relatively low accuracy andphysical testing weakness which only applied to off-line detection. Based on the art ofmanufacture and mechanism, this dissertation applies computer vision method to straightenerwhich has high accuracy and realizes real-time non-contact measure.This dissertation theory designs an optical image system of bars straight detection basedon already known research method, image phase difference theory and optical imaging systemfundamental theory. The designed system can calculate the required the parameters based onthe optical system. To be specific, firstly, the system will capture the bars’ images and thencapture the up-edge and down-edge image to build the coordinate system based on thecontrast of grey value. Secondly, according to central axle algorithm, this dissertation willcalculate the up-edge and down-edge central axle coordinates. However, as the detected unitis relatively large, one visual sensor is not enough, here is introducing two theodolites andmarked target to standardize all the sensors. In this way, the single visual sensor coordinateswill be systemized to one coordinate system. By transferring the every single coordinates tothe integrated system, the error will be computed by using Least Square Method (LSM).Generally, combining with straighter producing art and on-line measuring techniquesrequirements, designing the circuits’ structure of this straighter on-line measure system andbuilding the circuits principle diagram according to the required electronic unit technicalparameters are the main four steps to build the whole system. The manipulating programadopts the Siemens Step7software design approaches. All the algorithms use the OpenCV to simulate, exam and evaluate.This dissertation proves the straight error evaluating algorithms after the simulating test. Itcan control the on-line measure system with PLC programming. Overall, it basically achievesthe expectation for the original target.