Research On The Technology Of Rice Ear Harvesting Task Planning Based On Image Density

As an agricultural field crop,the rice ear is related to the food supply of more than one billion people.The traditional rice ear harvesting operation has low precision and slow operation speed.There is an urgent need for technological innovation to plan rice ear harvesting tasks to achieve precise harvesting and efficient operation.Improve the harvesting level of rice ears.Therefore,how to reasonably plan the task of harvesting rice ears has become an urgent problem to be solved,which is an NP-complete problem.This thesis studies the image segmentation algorithm and task planning algorithms to accurately segment the rice ear image,refine the harvesting tasks,rationally allocate the harvesting tasks,plan the harvesting job scheduling,and design and implement the rice ear harvesting management system on this basis.The main work is as follows:(1)In order to solve the problem that the existing segmentation algorithm does not perform well in the segmentation of the rice ear image,a rice ear segmentation algorithm RSAap based on atrous pyramid structure is proposed.The algorithm first proposes the redistributed atrous convolution RAC,obtains the weight of each feature,calculates and converts the weights in the positive end of the different atrous rate,redistributes the weights of the features to roughly detect segmentation targets of different scales.In order to improve the accuracy of rice ear segmentation,a reverse recursive pyramid RRP is proposed.The results obtained by the pyramid are recursively back to the backbone layer and the input information is transmitted to the pyramid network structure again as input at the same time.Iterate repeatedly to extract the precise segmentation target,then add the result to the residual module of the backbone network with feedback features.Finally,a more targeted loss function is proposed to optimize the segmentation algorithm.The experimental results show that this algorithm is significantly better than other algorithms in the accuracy and efficiency of segmentation of rice ear image.(2)In order to solve the problems of uneven load distribution of harvesting tasks and unreasonable path planning,a multi-region harvesting task planning algorithm MA-ACO based on ant colony algorithm is proposed.The algorithm first abstracts the rice ear region into a two-dimensional grid map,combines the results of the rice ear segmentation to generate a rice ear density matrix,divides the relative operation speed according to the rice ear density matrix,takes the operation time as the path cost,quantifies the operation tasks,and makes uniform planning.The job scheduling path of each harvester achieves the goal of maximizing overall revenue.Experimental results show that the algorithm is closer to the global optimal solution and has better stability.(3)In order to practically apply the RSAap of the rice ear segmentation algorithm based on the hollow pyramid structure and the ACO-based harvesting task planning algorithm to help the rice ears to be harvested accurately and efficiently,the rice ear harvesting management system was designed and implemented using Vue,Spring Boot and other technologies.The system uses the algorithm module’s rice ear image segmentation and harvesting task planning to generate harvester task allocation scheduling routes,displaying the user’s current harvesting position,rice ear image information,job travel speed,driving direction,etc.to achieve precise harvesting and efficient operations.