| Insects are one of the important factors leading to post-harvest food storage losses.During the storage of raw grain,timely and accurate monitoring of the occurrence of early insects in grain silos is the main basis for comprehensive management of stored-grain insects.The early monitoring of the occurrence of insects in grain storage in grain silos is mostly based on the sampling method and probe-and-trap method.The detection process is time-consuming and labor-intensive,and the level of information is low.In recent years,with the continuous introduction of electronic monitoring methods,electronic probes will gradually become the main means of monitoring insects in the grain silos.This paper designs and implements a new type of electronic tube detection equipment and its classification and recognition algorithm based on the tube detection trap and infrared photoelectric sensor.The main work is as follows:1.Designed and developed an Online Insect Trap Device based on Infrared Photoelectric Sensor(OITD-PI),including OITD-PI structure,circuit and low-level driver software.Designed and developed a data transmission device as a control hub for on-line monitoring and storage automation of insects information,mainly including control management software running on the data transmission device and a bus communication and data uploading scheme.The data transmission device and any OITD-PI constitutes an on-line monitoring device for stored-grain insects.The data transmission device requests,summarizes,classifies,and uploads the infrared photoelectric sequence data of the insects entering the OITDs-PI,and realizes the monitoring of the occurrence of insects in the grain pile.2.An infrared photoelectric sequence data set of nine main stored-grain insects adults was established.Through the analysis of variance and data visualization methods,it is verified that its data distribution characteristics are independent of the position where the insect enters the OITD-PI trapping section,so the stability of its data distribution is determined,and the entire infrared photoelectric sequence can be used to study the classification and recognition algorithm.3.Through in-depth analysis of the state of stored-grain insects when passing through the sensor beam range,an infrared photoelectric sequence recovery algorithm is proposed to solve the problem of infrared photoelectric sequence distortion and ensure the accuracy of feature extraction.Based on the geometric appearance of the insects,three methods of extracting the size,irregularity and shape of the insects were proposed.A Gaussian kernel support vector machine model was established for these three characteristics.Through laboratory tests,the discrimination of two major insect categories,internal-feeding insects and external-feeding insects,reached weighted average classification accuracy of 88.6%,recall rate of 88.4%,and F1 value of 88.5%.4.Research and analyze the impact of OITD-PI sampling frequency,sensor tolerance,component aging,environmental changes and other factors on the effect of equipment monitoring and identification,and propose solutions for standardization of standard parts and calibration of voltage change factors for specific effects.5.Through the simulation of the deployment of the grain depot,the equipment performance test under laboratory conditions was completed.The test results show that the maximum communication distance of the equipment can reach 3800m,the maximum load limit is 15,and the data communication time under full load is in 15 seconds,the accuracy of insect counting reached 99.7%.The on-line monitoring equipment for grain storage insects proposed in this article can accurately monitor the number and type of insects in the grain stack,so that the grain storage management staff can timely understand the occurrence of insects in the grain stack and take early control measures.To reduce stored grain losses due to insects and ensure stored grain security.The research work in this paper is of great significance and application prospects for the realization of green grain storage and the intelligentization of grain storage in China. |