Research On Supervision Method And Application Of Digital And Cloud Maps Of Grain Reserves

Grain storage is an important link to ensure the safety of grain production and circulation,which directly affects the overall level of grain security in China.Since the 21 st century,with the application of the "four-in-one" grain storage technology,China has established the world's largest grain Internet of Things system,so that the warehouses in different locations of our country is closely connected,which initially formed a "new infrastructure" for grain security.At present,the topic about management of grain warehouses has become a new issue that needs to be resolved in grain storage.One of the technical difficulties to be solve in this topic is the supervision of grain reserves.The characteristic of grain storage in China is that grain storage warehouses are widely distributed,each warehouse can store larger scale grain,and most of the grain storage period is long term.Those characteristic make the workload of supervision and audit of grain reserves very huge.In addition,due to the huge amount of grain reserves,which involves huge economic interests,if individual grain depots have illegal behaviors such as empty warehouse instead of full warehouse,false interest rate discount,and exchanging bad grain for good grain,it will cause great economic losses to the country and affect the national grain security.Therefore,it is of great social value and economic significance to improve the methods and modes of supervision of grain reserves in different regions to ensure the quantity and quality safety of grain reserves.In order to solve the problems of time-consuming and difficult to find problems in the supervision of grain reserves,this paper proposed the technical route and basic method of physical supervision of grain reserves based on the characteristics of temperature field of grain bulk,based on the three characteristics of grain bulk,i.e.spatial-temporal continuity,periodicity and coordination of multi field coupling.By analyzing the three characteristics of historical grain temperature data,this paper proposed the inventory modal detection and classification method based on the correlation and continuity of grain temperature digital features,as well as the correlation of cloud maps features.Aiming at the characteristics of wide distribution of grain warehouses,this paper studied the grain warehouses group supervision mode,and finally built a supervision system of grain reserves.The main contents of this paper are as follows:(1)Basic principle to supervise grain reserves-Study on field characteristics of grain bulkThree characteristics of biological material bulk during normal storage(without external interference)were analyzed.The three characteristics include spatial-temporal continuity,periodicity and coordination of multi field coupling.Taking the grain bulk as the object,this paper analyzed the spatiotemporal continuity,periodicity and coordination of multi field coupling of grain storage bulk during storage,and analyzed the derived characteristics of temperature field which is spatiotemporal correlation.The fact was verified that the reasonable use of three characteristics can detect the historical inventory mode,including empty warehouse,new grain,aeration,condensation,mildew and other states.(2)Research on supervision method of grain reserve based on correlation and continuity of grain temperature digital characteristicsThe inventory modal detection method based on grain temperature digital feature correlation was improved.Firstly,the self-correlation and cross-correlation of temperature measurement plane and the self-correlation of temperature measurement line and the temperature measurement point were analyzed.According to the results,the self-correlation coefficient threshold and the cross-correlation threshold of planes,lines and point were set respectively.The modal detection experiment of a real grain warehouse was carried out by using the threshold.Then,based on the above detection results,the inventory mode classification method based on the continuity of grain temperature digital features(mainly including empty state,new grain and aeration)was proposed.Firstly,it was verified that the temperature difference between adjacent layers of grain bulk and the new variation ratio of grain temperature could be used to detect the empty mode,the temperature differencebetween adjacent layers and the standard deviation of grain temperature can be used to detect the new grain addition,and the grain temperature change rate and standard deviation change rate could be used to detect the aeration mode,and the threshold ranges of the above parameters were initially set.Then,grain temperature data from seven different provinces were selected to detect the three modes.The results showed that the average precision rate,average recall rate and F value of the three modes were 81%,80% and 87% respectively,which indicates that the method can basically meet the requirements of grain reserves supervision.Secondly,the characteristic parameters of grain temperature data from 68 grain warehouses in the second to seventh grain storage ecological areas were calculated.The threshold ranges of characteristic parameters were optimized by K-Means++,K-Mediods and DBSCAN clustering methods.The Rand Index(RI)was used to evaluate the clustering results.Results showed that the clustering effect of DBSCAN method was better(RI =0.9703).(3)Research on supervision method of grain reserve based on correlation of temperature field cloud map featuresThe paper proposes a method of grain reserve supervision based on the correlation of temperature field cloud map.First,the historical grain temperature data was preprocessed to generate the cloud map;then,the similarity of cloud map was calculated by using RGB color distribution of temperature field cloud map,and the abnormal judgment threshold was set accordingly;then,the similarity of each plane cloud map of adjacent time was calculated,and the time point of mode change was detected according to the threshold value set;the situation of 5 kinds of grain replacement was simulated and carried out.The results showed that the average recall rate and the average precision rate of the method based on the RGB color feature of cloud map were 98.6%,97.3%,and the operation rate was about 320 ms/time,which realized the detection of inventory mode change.According to the above methods,a new inventory mode classification method based on multi-feature fusion of temperature field cloud map was proposed.Firstly,BP neural network with double hidden layer was constructed.By comparing the classification effect of BP neural network when different feature combinations of temperature field cloud map were used as input matrix,it was found that when the combination of color coherence vector(CCV),texture feature(TFV)and smooth feature(SFV)was used as input matrix,the accuracy of classification was better(about 93.9%).The classification experiments of temperature field cloud map of different modes were carried out by using the established network,which include empty state,new grain state,aeration state,self-heating state and normal state.The results showed that the accuracy rate of empty state was higher than 98%,the accuracy rate of aeration state was between 82%and 89%,while the classification accuracy of new grain state,self-heating state and normal state was between 89%and 98%.(4)Research on grain warehouses group supervision mode of grain reserves in different regionsFrom the perspective of different managers,this paper analyzed the implementation mode of grain warehouses group supervision in different regions.From the perspective of National Food and Strategic Reserve Administration and the managers of China National Grain Reserve,this paper studied the implementation mode of grain warehouses group supervision in different regions.It is found that the implementation of the mode needs to establish a cloud platform for Grain Reserve supervision and a grain situation data server.The management organization uses the cloud platform to supervise the grain depots in different regions,so as to realize the supervision of grain warehouses group in different regions;This paper analyzes the necessary conditions for the implementation of grain warehouses group supervision in different regions,that is,the standardization and large-scale storage of grain condition data;This paper analyzes the common software architecture C/S and B/S architecture,and studies the application mode of grain warehouses group supervision in different regions using these two architectures.The research results can provide direction for the implementation of grain warehouses group supervision in different regions.(5)Digital and cloud map monitoring system for grain reserves and its application testThis paper analyzed the main function modules of the digital and cloud map monitoring system of grain reserves,and used Lab VIEW programming tool to design and complete the construction of the monitoring system of grain reserves.The system mainly includes six modules: data reading module,granary selection module,grain situation scanning module,cloud map analysis module,result display module and three temperature analysis module,which can realize the scanning and analysis of historical grain temperature,and the results can be output,save and other functions.More than 230000 groups of grain situation data from 592 grain warehouses in recent one year were detected by the system.The results showed that the accuracy rate of detecting empty warehouse mode was 94%,the accuracy rate of detecting new grain addition was 93%,the accuracy rate of detecting aeration mode was 95%,the accuracy rate of detecting condensation and mildew was 96% and 87%respectively.The results showed that the digital cloud map monitoring system of grain reserves could basically meet the requirements of grain reserves monitoring.The research and application of the system support the implementation of grain warehouses group supervision in different regions.