Design And Implamentation Of Grain Condition Monitoring System Based On Mobile Augmented Reality

Since ancient times,food security has been one of the focuses of people all over the world.Food security is a major strategic issue concerning the country's economic development,social stability and national independence.However,because China's research on food security management started late compared with Western countries,and with the rapid development of China's economy,people's living standards are growing,and the demand for food is increasing year by year,resulting in an increasingly larger domestic grain warehouse.At present,the existing grain condition monitoring system deals these changes,which presents many problems,such as the cumbersome real-time detection procedure of granary,the imperfect analysis of grain storage security data,and the scientific and technological level and low efficiency of grain condition monitoring system.In recent years,with the rapid development of smart mobile terminals and mobile networks,mobile augmented reality technology has become more and more popular.It has been widely used in the fields of university education,medical research,urban planning,etc.However,the research on the combination of mobile augmented reality technology and grain situation monitoring system has not been involved in the domestic and foreign research.Therefore,this paper proposes to integrate mobile augmented reality technology into the grain situation measurement system,and carry out research work from the following points.1.Conducted field investigate the grain warehouse site,combined with the staffs suggestions and requirements for improvement of the grain condition monitoring system,analyzed and integrated,used the current mainstream Android system to develop the mobile client of the grain condition monitoring system,and the mobile augmented reality technology was integrated into it.Through the mobile augmented reality technology,the food security data and the granary information are rendered into 3D models,which are displayed in real scenes,thus improving the management efficiency and technology level of the grain condition monitoring system.2.Based on the analysis of the functional modules and main functional requirements of the existing grain condition monitoring system,the implementation process of the mobile augmented reality technology in the grain condition monitoring system is divided into five modules,wherein the scene sensing module is responsible for obtaining the positioning and position of the mobile terminal.The posture information or the target object feature information;the data interaction module is responsible for encapsulation,parsing and transmission of data transmitted by the mobile terminal and the server;the target recognition matching module is responsible for identifying the target object and calculating the positioning information;the tracking registration module is responsible for the video image sequence Tracking the target object and updating the positioning information in real time;the virtual and real fusion module is responsible for rendering the virtual information for enhancement and superimposing on the real scene according to the positioning information.Due to the modularization of the system,the system has the advantages of high reusability,low coupling and high scalability.3.Aiming at the problem of identification of grain varieties in the target recognition and matching module of the system,a food variety recognition algorithm based on Fuzzy C-means clustering support vector machine(SVNM)was proposed,the algorithm is introduced into the fuzzy c-means clustering method for preprocessing the training sample set,the training sample data set is divided into several independent sub clusters,then an SVM model is established on each sub-cluster,thereby correcting the abnormal value of the input data and reducing the redundancy of the classification model;and selecting Gaussian radial kernel function as a kernel function of SVM improves the learning ability of the model.The experimental results show that the proposed algorithm has better recognition effect than BP(Back Propagation)neural network and original SVM model,and the accuracy rate reaches 95.83%.4.Aiming at the problem of target tracking registration in the tracking registration module of the system,based on the feature-based target tracking algorithm,an online target tracking algorithm based on improved RANSAC(Random Sample Consensus)is proposed.Firstly,the online update target matching template strategy is adopted,using the ORB(Oriented FAST and Rotated BRIEF)algorithm to extract the local features of adjacent frames and feature matching;Then the RANSAC algorithm is introduced to remove the mismatch in the matching result,in order to solve the real-time problem caused by the introduction of RANSAC algorithm,the hamming distance was used to improve the RANSAC algorithm,and finally.the target tracking is performed according to the purified feature points.Since the adjacent frame is used as the target matching template,the accuracy and time-consuming stability of the target tracking are improved,and the improved RANSAC algorithm reduces the number of iterations,so the real-time performance of the algorithm is improved.The experimental results show that the proposed algorithm has better tracking success rate and accuracy than the existing target tracking algorithm,and the algorithm has better real-time performance.To sum up,the thesis designs and implements a grain condition monitoring system based on mobile augmented reality technology,which is of great significance and potential application value for improving the current grain storage management level in China.