Study On Chestnuts Sorting Equipment Based On Machine Vision

The main objective of this paper is to design an on-line precision chestnuts sorting machine based on machine vision and develop image processing and recognition algorithms for real-time chestnuts sorting. This dissertation has achieved the following results.(1) The experimental machine consists of three parts. The first part is control system, which includes of PC computer and windows 2000 operation system. The second one is image processing system, which is made up of Panasonic B/W CCD, BS602 image grabbing card and photoelectricity position sensors. The third one is sorting machine, which is driven by stepping motor. The control system transmits image processing information and instructions to motor-driven circuit by standard LPT.(2) High-speed dynamic image processing algorithm is developed. And the degeneration models and restoring methods are established for blurred image. The optimizing gray scale histogram is adopted in this paper to threshold and draw the chestnuts from background. The Gauss- Laplacian operator is applied to edge extraction. To meet the demand of characteristic extraction, this paper makes use of a specific method for contour extraction and track.(3) This paper extracts the feature of chestnuts and takes them as sorting criteria. The minimal external rectangle (MER) is taken to measure the long and short radius of approximate elliptic chestnuts, which are computed to actual size of chestnuts. This paper also put forward adopting long and short radius and their ratio as the sorting criteria.(4) This paper designed the grading driving circuits based on the L297, L298 and NE555. And the circuit can control the start and stop and corner of the step-motor. Under the platform of Windows 2000 and WinlO driver, programs written by Visual C++ 6.0 is compiled to drive sorting machine.(5) The chestnuts sorting machine based on machine vision is tested. It is proved that this system meets the requirements on sorting speed and accuracy of dynamic image grabbing and real-time processing. The chestnuts sorting velocity and precision are up to 40-50 pieces per minute and 95.33%, respectively.