Study On Real-time Rendering And Rapid Collision Detection Of Rice Population Growth

Virtual crop visualization is a combination of agricultural science and computer science,As the virtual crop simulation can effectively shorten the crop growth cycle,so in agricultural research and agricultural teaching has a great role in promoting,and the real world crop to the size of the population management,so to achieve large-scale virtual crop visual simulation has Important practical significance.Virtual crop visualization has made many research results after years of development,But there are still two problems:1)Most of the studies have been targeted at crop organs or individuals,In the visual simulation of the virtual crop population growth,as the number of plants more,and each state of the crop grows differently,so the the amount of computer resources consumption increase,result in the rendering efficiency is low,can not satisfy the real-time drawing;2)In the study of collision detection of virtual crop organs,The research has focused on the collision detection of static crops.When the crop population is increased and the collision detection of the dynamic growth process is needed,the detection efficiency is low and the requirements are not satisfied.So to solve these problems can effectively promote the development of virtual crop simulation,and thus promote the development of virtual agriculture,In view of the above problems,the main contributions of this paper include:(1)Design a multi-real-time rendering technology integration of the drawing methodIn order to realize the real-time rendering of the large-scale rice population growth process,this paper presents a method based on the morphological structure of rice,which is based on a variety of real-time rendering.In order to realize the real-time rendering of large-scale rice population growth process,Technology integration of the drawing method,the method of integration of the model simplification technology,horizon cutting technology,display list technology,detailed level of technology and other optimization rendering technology,the main idea is as follows:Improved the way of rice leaf data acquisition,and reduced the calculation time consuming;On the basis of this,we use grid simplification technique to simplify the various organs of rice,and then form the individual model of rice with different resolution levels and encapsulate the drawing process into multi-level display list;At the time of drawing,the field cutting technique is used to determine the plants to be plotted and to draw the display list of the corresponding resolution hierarchy rice model according to the detail level technique.Experiments were carried out on the whole growth period of indica rice YD6 and japonica rice WXJ14.The results showed that:1)The system can only simulate the growth process of about 50 rice plants without using real-time rendering technology.Based on the improvement of leaf data acquisition mode,the display list technique and model simplification technique can simulate 1000 strains in 3 minutes The morphological changes of the whole growth period of rice,when the population size continues to increase,still can not meet the demand;2)The visualization simulation of rice population growth process can be carried out in 10 minutes to complete the visual simulation of 14000 rice whole growth period,which greatly improves the efficiency of simulation and satisfies the requirements of drawing.(2)Design a rapid collision detection method for rice leavesThe collision detection between the rice groups is to detect the penetrating phenomenon of the leaves in time,and to respond to the response strategy,so as to avoid the phenomenon and increase the realism of the virtual rice simulation.But in the rice population in the collision detection due to the larger number of leaves,resulting in lower detection efficiency,In order to solve the technical problem of low efficiency of inter-blade collision detection in the dynamic growth simulation of rice population,this paper proposes a method which has rapid speed of leaf construction(Mixed Level Tree Model)and collision detection(CGGPU Utilization).According to the morphological characteristics of rice leaves,The calculation method of the new OBB bounding box direction axis is proposed to replace the traditional calculation method of covariance matrix and mean value,which reduces the complexity of OBB bounding box construction;The CPGPU acceleration scheme was designed by using the dependence of the collision detection between the leaves of the individual plants and the leaves of the population,And improve the efficiency of collision detection between rice groups.Experiment on large-scale population leaves of indica rice YD6 rice at tillering stage,The results show that:1)the collision detection method proposed in this paper is more than 50%less time than the traditional AABB,OBB method,which effectively improves the collision detection speed between blades;2)When the rice population is large,the CPGPU parallel acceleration is 98%less than the running time on the CPU,and the collision detection efficiency is greatly improved.