Font Size: a A A

Research Of The Silkworm Cocoon Automatic Counting System Based On Image Processing

Posted on:2014-01-09Degree:MasterType:Thesis
Country:ChinaCandidate:L L QiaoFull Text:PDF
GTID:2248330398494635Subject:Signal and Information Processing
Abstract/Summary:PDF Full Text Request
In the silkworm cocoons spinning process, the first-class silk is span by7-9cocoons. Thisrequires real-time count of the number of silkworm cocoons, traditional method is that workerskept walking back and forth to observe whether the number of silkworm cocoons is7to9, thisprocess is not only a waste of time, but also a cause of visual fatigue. It also has low workingefficiency and the accuracy is hard to guarantee. Therefore, the development of a fast, accurateautomatic counting system, not only can improve the productivity, but also can improve thequality of the product, and realize auto-counting of the number of silkworm cocoons.In order to achieve the goal of silkworm cocoons automatically counting, the paper is todevelop a silkworm cocoons automatic counting system that based on image processingtechnology. This system includes image acquisition module, image processing module, and theresults shown module. In the paper, the research of each module and innovation of the systemare as follows:(1) Research of development of image acquisition programming and image storage. Thepaper develops image acquisition program based on camera’ SDK.A computer-controlledcamera takes photos to memory buffers, and change picture data of this memory buffer into animage handle, which facilitates subsequent processing. The pictures are stored by SQL Sever,which can achieve fast query speed and is convenient to analyze.(2) Adapt two area threshold value improved corrosion algorithm based on prioriknowledge. In this paper, with setting two area threshold value S1, S2(S2<S1), the system hassome characteristics such as accurate judge, fast processing, and solves the mistaken of thenoise in traditional single area threshold value as a single target object, to achieve a fast,efficient segment of the target object.(3) Generate boundary with the method combined4-neighborhood with8-neighborhood.Inthe different corrosion process, it takes the8-neighborhood at even corrosion,4-neighborhoodat odd corrosion. It avoids of the long time taking while using one single method to generate theboundary, increases the boundary generating speed.(4) Build a cocoon automatic counting system. In order to verify the implementation of thekey technology, the paper builds a cocoon automatic counting system, of which the advantages is accurate, fast and stable, and it solves the problem of traditional manual counting slower,lower accuracy.Experiment testing of the silkworm cocoons automatic counting system shows that thesystem basically meets the needs of the automatic counting of the silkworm cocoons, and theaverage accuracy rate is more than95%.It solves the trouble of manually counting in thespinning process, and achieves a steady and rapid counting.
Keywords/Search Tags:image acquisition, median filtering, HSI model, binarization, corrosionsegmentation, finding center point
PDF Full Text Request
Related items