Research On Agricultural Machinery Equipment Based Intelligent Simulation Behavior

Agricultural mechanization and agricultural machinery equipment are important foundations for transforming agricultural development methods and improving rural productivity,and are an important support for the implementation of rural revitalization strategies.The realization of informationization and intelligence of agricultural machinery equipment is an important way to improve the level of agricultural mechanization.In order to realize the intelligent design of agricultural machinery equipment and the simulation of the operation behavior and performance of the whole machine,this paper researched the assembly behavior of key components of agricultural machinery equipment that incorporates machine vision,developed a simulation platform for harvesting operations,and performed harvest simulation test research under different conditions.The intelligent simulation behavior of the equipment had been studied,and the following researches had been carried on:1)Object-oriented modeling methods of agricultural machinery equipment was researched.By analyzing the implementation method of the object-oriented modeling method,an object-oriented simulation model construction method suitable for Unity3 D was proposed;The construction process of the simulation model was expounded,including geometric shape model,collision body and virtual environment construction.The perspective roaming model was analyzed,and the artificial roaming control method for the behavior simulation of agricultural machinery equipment was proposed.The agricultural equipment operation status and behavior modeling were analyzed,and the behavior control model based on stress behavior and feedback behavior was studied.2)Assembly behavior based on color segmentation and SVM classifier was researched.The machine vision perception was integrated to research the interactive virtual assembly technology.Firstly,a virtual assembly platform was established,what contains a simulation model of parts and virtual hands,and implemented interaction based on behavioral inference rules.Then,the hand image was segmented based on the color space,and the recognition results was obtained through the SVM classifier,and the position information was obtained according to the principle of triangulation.Finally,The system was applied in visual system verification and interactive virtual assembly tests.The constructed SVM classifier had a success rate of 92.45%,an accuracy rate of 98.65%,a recall rate of 86.90%,and an F1 value of 92.41%.The time required for hand recognition and positioning of the vision system was65 ～ 100 ms/frame;and interactive virtual assembly of parts 90% success rate.The results proved that the platform meets the requirements of virtual assembly.3)Assembly behavior based on deep learning framework was researched.The basic composition of the convolutional neural network and the principle of the target detection network were analyzed.With reference to the Multiview Bootstrapping algorithm,a method for detecting and extracting human hand skeleton information based on 21 key points of the hand in the deep learning framework was proposed.scene simulation tests were performed.The results showed that this method can obtain more behavioral information,has good detection results in different perspectives,has better robustness and works well on virtual assembly.4)Agricultural machinery operation behavior was researched.Taking corn harvesting equipment as an example,a 3D simulation model of the harvesting equipment was constructed by using the Unity3 D virtual platform.According to the rheological characteristics of the material,a plant model was constructed with a virtual spring,and the interaction between the harvesting equipment and the plant was realized by establishing a behavior control model.The development of the harvesting simulation platform was completed,and the verification experiments and the harvesting operation simulation tests under different conditions were performed using the platform.The results showed that the virtual environment modeling was reliable,the agricultural equipment model was operating normally,and the interaction with the plants was real.The simulation platform was effective,and the fruit leakage rate decreases rapidly with the increase of the traveling speed.The higher the plant density,the slower the fruit leakage rate declines,which was in line with the actual harvest law.The harvesting simulation platform can evaluate the performance of harvesting equipment.