Study Of Quality Detection System For Agricultrual Products Based On ARM

With the rapid development of the embedded technology, the ability of processing for ARM (Advanced RISC Machine) processor is increased rapidly and the cost of ARM processor is also decreased. This paper studied the problem of design and implementation of agricultural products quality detection system based on embedded ARM technologies. It has important practical significance for reducing system cost and expanding scope of application in agricultural products processing areas.Apple processing algorithms are mainly researched for the problem of quality detection.Through simulation experiments, the pretreatment algorithm and the feature detection algorithm are conformed.four aspects of apple size, shape, colour and defect feature detection algorithms are studied. Using apple symmetric features, apple axis is confirmed.After extracting apple edge points, apple diameter is calculated. H component is used for apple surface red color segmentation In HIS model.Then red component percentage is calculated.the standard ball gray model method is used for defect recognition, by setting thresholds, the defect features are detected out.System implementation problem is systematically studied here.The hardware of system is discussed detailedly.After designed the acquisition device and lighting source and selected right system processor. The whole hardware of detection system is constructed at last; then the whole system software is conformed, and QT platform is used for application software development, Linux operation system and file system are transplanted, each software module is designed and implemented at last. After debugging successfully on host computer and transplanting it to ARM detecting system, the whole system software design is realized.System detection test is carried out at last. Experimental results show that the embedded quality detection system beased on ARM can complete apple feature detection.And the accuracy of this system for size and shape detection is over 90%, the consistent rate of color detection is at 80%, the recognition rate of defects is about 80%. The structure of system is simple. System is low cost and of high speed. Through extending the detection algorithm, the system can be used for other agricultural products feature detection.