Application And Research Of Image Stitching Technology In Rice Detection System

With the rapid development of computer information technology, machinevision-based online detection technology is used very widely. and rice-qualityinspection system is one of them. As the rice detection system process videoinformation with a mode of a sub-frame sampling. The Multi-frame images obtainedin this method either have overlapping part or are more separated, which increase thedifficulty to deal with rice image. At the same time, they can’t achieve real-timeprocessing.Using image mosaicing technique can mosaic multi-frame rice image into a riceimage of panoramic view, and then use rice detection system for processing. it caneffectively reduce the burden which come from the system to process imageoverlapping area, and it also can avoid the difficulties which come from the system todealing with rice image edge information. And it can improve the efficiency of therice detection system, and realize the rice detection system real-time detection.The main study of this paper is the application of image mosaicing technique inrice detection system. Firstly, I study the mage sampling of rice detection system,then determines a best sampling frequency on the rice image system. Determinedthe preconditioning method of the rice image.We used image gray histogramequalization, image denoising, image enhancement, image correction, and restoreimages blurred methods to preprocess the rice image by analyzed different factors tothe impact the image quality before image stitching, and use matlab for a largenumber of experimental analysis compared,then elect the most suitable measures forthe pretreatment of the rice image. Obtain the characteristics of the image of rice, thispaper put forward a limited range of template matching algorithm base on thetraditional template matching algorithm. The algorithm processing speed to more thantwice as fast than the traditional template matching algorithm. Then according to theedge of the rice characteristics and video image characteristics of sampling isproposed based on the edge of the phase of the characteristics of rice relatedalgorithm. According to the edge characteristics of the rice image and video image,this paper put forward a phase position related algorithm based on the edgecharacteristics of the rice.The algorithm of improvement is not only good onprocessing the rice image which effects by noise pollution, but also computation more efficient, and it suitable for real-time monitoring of treatment. The paper also putforward a weighted average fusion algorithm base on rice image. Analysis and studythe algorithm and process of image matching process, then I do many experiment andanalyzed the experimental results.