Font Size: a A A

Design And Implementation Of Rice Variety Recognition System Based On Android

Posted on:2021-12-30Degree:MasterType:Thesis
Country:ChinaCandidate:S Y JiaFull Text:PDF
GTID:2481306467471734Subject:Master of Engineering
Abstract/Summary:PDF Full Text Request
Because rice is the most important food crop eaten by our people,it is of far-reaching significance to identify its varieties.In the past,the artificial method of hand checking or eye looking was usually used to achieve the purpose of identifying rice varieties.This type of method is more susceptible to subjective influence,is inefficient,and the results of the identified rice are not credible,Affecting accuracy.Nowadays,under the development of mobile Internet,research on mobile portable devices has become the mainstream.Regarding the identification of rice varieties,various PC-based identification systems have been researched and developed,but there is little research on identification system software based on mobile devices.Therefore,the development of rice recognition application software that can be used on mobile devices has high practical significance.This paper designs a rice variety recognition system based on the complex background of Android.The work done by the paper is as follows:(1)Image collection,Select three representative types of rice samples: indica,japonica and glutinous rice,and use mobile terminals to obtain rice images.(2)Image preprocessing and feature extraction,OpenCV function library is used to preprocess images.For example,grayscale the image,binarize the image,detect edges,and segment images.Through several common image preprocessing algorithms,Canny algorithm is used to detect the edge of the image,and Grab Cut algorithm is used to segment the image.After these methods are processed,the best preprocessing effect can be obtained.Extract the characteristic parameters of the processed image,extract the width-to-length ratio,gray average and roundness of the rice image according to the morphological characteristics of the rice,and input them into the classification model to complete the recognition.(3)Identification of rice varieties,K-means clustering algorithm and BP neural network algorithm are used to classify characteristic parameters to realize identification of rice varieties.At the same time,the experimental data of rice samples is divided into training samples and test samples to verify the accuracy of the system identification.The test results show that this algorithm can basically identify the rice varieties.(4)Designed a rice identification system based on Android,using TCP / IP protocol to achieve communication between the client and the server,the server returns the recognition result to the Android 'client,and the mobile client accurately receives the recognition result.
Keywords/Search Tags:Rice, Android, Feature Extraction, OpenCV, BP Neural Network, Grab Cut
PDF Full Text Request
Related items