Font Size: a A A

Study On Identification Of The Hook On Catenary Based On Machine Vision

Posted on:2013-11-10Degree:MasterType:Thesis
Country:ChinaCandidate:J YuFull Text:PDF
GTID:2232330395965655Subject:Mechanical and electrical engineering
Abstract/Summary:PDF Full Text Request
The catenary production line has been widely used in industrial production, such aspainting, ball blast, drying and livestock slaughter, etc. Currently, loading and unloading theworkpiece on catenary is still done by manual, rely on the physical strength of workers. It’slabor-intensive for workers, low productivity, and painting and other operations will haveserious effects for human health. It needs to use automatic handling device to complete therequirements of improving production quality and efficiency on catenary. In order to loadingand unloading workpieces automaticly, it needs to find the location of hooks and workpiecesbofore it places the workpiece on the hook. According to this characteristic, we introducedmachine vision to the recognition of hook on catenary. We adopt machine vision to ensure thelocation of the hook, and the automated loading and unloading equipment adjust trajectoryplaced workpiece accurately. The content of this article includes the following several aspects:Firstly, we has designed hardware system of the hook’s image acquisition, includinglighting system, camera, the acquisition card and others. we selected bumblebee2binocularcamera made by Canada Point Grey Company to acquire the hook image on the catenary. Wehave completed the calibration of internal and external parameters of the bumblebee2binocular camera adopting ZhangZhengYou plane calibration, and we have done the erroranalysis for the parameters.Secondly, we put the hooks on poultry slaughtering line as the research object, accordingto the characteristics of this kind of hooks, first we used the median filtering algorithm to dealwith the noise. On the basis of comparing every feature matching method, this paper puts fora word using SIFT to realize image registration, and combined with estimated projectiontransformation matrix to get the matching points between two images, and finally got the3Dinformation of the hook according to the triangular law. It provided reliable recognition imagefor robot identify hooks automaticly.Thirdly, put the hooks on catenary of investment casting workshop as the research object,after the pretreatment for the acquisition of the image, we adopt Hough transform to reducethe recognition region to target range. It identify the target area clearly through the horizontaland vertical direction gray-level projection. Do images matching, and ultimately we get the spatial information of target recognition regional of the hook.
Keywords/Search Tags:catenary, machine vision, image processing, automatic identification
PDF Full Text Request
Related items