Font Size: a A A

Research On Sorting Method For Granular Agricultural Products

Posted on:2017-05-10Degree:MasterType:Thesis
Country:ChinaCandidate:J LuoFull Text:PDF
GTID:2283330485462245Subject:Information and Communication Engineering
Abstract/Summary:PDF Full Text Request
For sorting granular agricultural products, the operation by machine has gradually replaced that by manual work in recent years. Instead of the long sorting period, high error rate and low efficiency by traditional manual work, machine operation has been widely applied for its advantages such as high speed, processing of massive information and multi-function. This article is intended to Research on Sorting Method for Granular Agricultural Products. That is to acquire image of granular agriculture products through color liner array CCD imaging system, and save the acquired image on PC for further feature extraction and image filtering. Eventually, implement KNN algorithm on FPGA to sort granular agriculture products according to extracted optimal feature values. This article mainly aims at the search on:(1) Research on connected domain labeling algorithm. In order to trace and label moving objects, connected domain labeling is required for the objects in image. Connected domain labeling algorithm based on linear CCD will be discussed. Then the algorithm will be validated on PC and its implementation method through FPGA will also be stated.(2) Feature selection based on KNN algorithm. Execute feature extraction in RGB color space, mainly in one and two dimensional histograms. Make a further extraction on the basis of above feature by KNN algorithm and verify the KNN algorithm through obtained optimal feature values in return. The results show that sorting accuracy of peanuts and pistachios can reach more than 95% when a reasonable feature number and K values are applied. Experimental results show that 3-5 feature makes highest accuracy, which also indicate that the KNN algorithm is applicable for FPGA implementation.(3) Implementation of KNN algorithm based on FPGA with the utilization of HDL (Hardware Description Language), implement KNN algorithm on FPGA and validate it through simulation. Based on a linear array camera vision system, provide the algorithm ideas and verification.
Keywords/Search Tags:KNN, FPGA, Feature Selection, Sort, Connected Domain labeling algorithm
PDF Full Text Request
Related items