The Study Of Picking Decision Plan And Fruits Maturity Distribution Prediction Based On Dynamic Bayesian Network

Because of lacking of information and cooperation with each other, the capacity of resisting natural disaster and market risk is low in our agriculture production. Now information technology can solve this problem effectively. The application of information technology transformed agriculture production from extensive to precision and realized agricultural sustainable development.The development of agriculture is related to good time and good land. The time here means meteorological conditions which is one of the most important factors in agriculture production. The agricultural production is the process of fully using, utilizing, adjusting the climate, terrain, natural resource, which is a complete system. In order to achieve a high yield of agricultural products, it has to take the economic factors, social factors and technical factors into consideration in this system. So it is important to make proper agricultural activities responding to these factors in this complex system.Based on the research about agricultural informatization, this paper provided a method for picking fruits in order to get the maximum profit. As we all known, abnormal weather influenced seriously on fruit in maturity stage. For example, heavy rainfall will damage large area fruits. So we want to avoid this by picking fruits in advance. But earlier picking will influence its economic value because of the price is related to its maturity stage. And when to pick will also be an important problem which would also relate to its profit. This paper provided an intelligent decision algorithm which can provide the optimal picking scheme for peasant when facing abnormal weather. The picking scheme can provided the maximum profit for peasant according to analyze complex factors.There are two primary works in this paper. Firstly we provide an algorithm for fruits’maturity distribution prediction which can caculate probability of every maturity stage. Then we provide a decision algorithm to give the optimal picking plan for peasant under the condition of abnormal weather. The detailed works are descried in the following:(1) Determine the current state of fruits’maturity. Because different mature period of fruit has different skin color, so we can determine the state of the maturity according to color histogram.(2) Provide a prediction algorithm of fruit maturity distribution. According to the process and influenced factors of fruit ripening, this paper provides a prediction algorithm based on Dynamic Bayesian Network. Though the prediction algorithm, we can calculate the maturity distribution ratio, which provides the basis for the picking plan.(3) Build a fruit income effect model. The factors affecting fruit yield can be divided into four categories, such as the cost, the sudden factors, fruit growth and economic value. Each of these factors is affected by lots of other factors. This paper detailed analyzes the impact of each kind of factor, and then abstracts the elements into mathematical model.(4) Establish and optimize the fruit picking decision model. This model mainly solves the problem of picking fruits in different maturity. And the decision is based on the income. Therefore, on one hand, this paper establishes the objective function based on the cost and benefit; on the other hand, it also establishes the constraints by harvesting ability and fruit yield rule. According to the constraint condition, this paper proposes recommendation algorithm the optimal picking scheme, optimizes objective function, then get a picking scheme.(5) Build the experiment platform. According to the forecast algorithm and picking model, this paper builds the simulation platform, which verifies the effectiveness of the algorithm.