Research On The Detection Method Of Grain Weevil In Paddy Grain Pile And Its Breed Pattern

China is a large grain producing country,and paddy,as the staple food for more than half of our population,can provide the human body with carbohydrates,proteins,fats,vitamins and minerals and other nutrients after its processing.As a seasonal food crop,good storage conditions can extend the shelf life of paddy and avoid food losses.It also provides a strategic reserve to secure food supply in case of emergency.China mainly uses warehouse to store paddy,and the environment inside the warehouse needs to be strictly controlled to guarantee the storage quality.In the process of paddy storage,grain pests and mildew have been important factors affecting the quality and safety of paddy.Under unsuitable storage conditions,sitophilus are prone to breed in paddy grain.Due to the small size of sitophilus,similar body color and environment,and complex hiding environment,the mechanism of temperature and humidity changes in paddy grain pile on their reproduction pattern is unclear,which makes it difficult to provide an effective prevention strategy for paddy grain storage pests.In this paper,an improved YOLOv5 algorithm based on deep learning is proposed to solve the above practical problems,which can accurately identify sitophilus in paddy grain pile.By building a grain weevil detection test platform,combined with COMSOL temperature and humidity simulation analysis,the sitophilus breed pattern within the paddy grain pile is explored,and the specific research contents are as follows:(1)Influencing factors of grain weevil breed pattern and characteristics of paddy grain pile.Preliminary analysis of the influencing factors of sitophilus breed pattern in the grain pile to provide research directions for subsequent studies.The physical characteristics of sitophilus and paddy grain pile were tested and analyzed.Sitophilus characteristics include:size and color.The characteristics of paddy grain pile include: water content,bulk weight,porosity,permeability,specific heat capacity,thermal conductivity,etc.To provide theoretical data support for the subsequent research of this paper.(2)Deep learning-based grain weevil detection method.Aiming at the practical problems of low detection accuracy in grain weevil target detection due to the small size of the worm,similar color and environment,and complex hiding environment,an improved Yolov5 algorithm based on deep learning is proposed.The recognition accuracy of sitophilus in paddy grain pile is improved by modifying the feature scale,adding a convolutional attention module,improving the Mosaic image enhancement technique,applying depthseparable convolution,and improving the loss function.The feasibility of this improved algorithm is verified by conducting comparison tests with traditional image recognition methods.(3)Grain weevil detection platform design and experiments.In order to investigate the breed pattern of grain weevils in different temperature and humidity environments in grain silos,a grain weevil detection test platform consisting of a simulated grain silo,an environmental condition monitoring module and a visual recognition detection module was designed and built.By conducting the bench test,we investigated the effects of temperature and humidity changes in the paddy grain pile on the breed of grain insects.(4)COMSOL-based research on grain weevil breed pattern.A COMSOL simulation model of temperature and humidity variation in paddy grain pile was established based on the characteristic parameters of paddy grain pile,and the accuracy of the simulation model was verified by combining the results of the aforementioned bench test.By analyzing the changes of temperature and humidity in the paddy grain pile under different conditions and combining the basic data of grain weevil changes with temperature and humidity obtained in the previous paper,the linear regression analysis method was used to obtain the distribution number of grain weevils as a function of temperature and humidity changes,and the sitophilus breed pattern in the paddy grain pile was analyzed.