On-Line Detection of Hami Big Jujubes´ Size and Shape Based on Machine Vision

In this paper , we present a on-line machine vision system for Hami big jujubes´ size and shape detection, which solved the problems such as fast processing the large amount of image information, improving system performance for real time dynamic image capture and processing capability, increasing precision of detection etc. Based on the application of Visual Basic 6.0 designed the user´s interface . SOLIOS images gathering card linked with the CCD camera to collect image data, Which consisted united application of MIL9.0 and VB on Auto-Detecting of the Hami big jujubes´ size and shape. With image processing control of ActiveMil 9.0, comparison of the algorithm and verified, we can find the feature extraction to distinguish the rapid defects about the size and shape. During the image processing, many methods were used to identify the quality defects as soon as possible, such as application of wavelet denoising, threshold segmentation, mathematical morphology processing methods etc. A Hierarchical Clustering Analysis method was used to realize the recognition and to determine the classification , which the recognition rate can research 83%.