Research On The Quantitative Identification Of Green Plant Water And Fertilizer Coupling Based On The Hmm Algorithm

Chlorophyll fluorescence is a non-destructive probe for green plants that can reflect the effects of water and fertilizer on plants. Plants in response to changes in environmental conditions, the amplitude information of chlorophyll fluorescence will change accordingly.Based on the relationship, the relationship between chlorophyll fluorescence intensity information and water fertilizer coupling is studied by the way of fluorescence detection. The basis for optimal irrigation with water and fertilizer is provided. The rational irrigation and fertilization of modern agriculture and forestry in our country have been realized.Based on the principle of chlorophyll fluorescence generation, the fluorescence data were collected by fluorescence kinetic imaging using LED inducement. A hardware detection system was established to obtain fast and slow phase fluorescence intensity profiles. Single variable moisture or nitrogen fertilizer experiments are designed under experimental ideal conditions.And coupling experiment of water and fertilizer is implemented under actual growth conditions.The fluorescence intensity profiles are analyzed and compared when both water and nitrogen are applied to the plant at the same time under the actual growth conditions, which confirm the effectiveness and limitations of current methods. Traditional methods can not effectively detect the fluorescence situation under the simultaneous action of water and fertilizer.Therefore, the processing algorithrm of HMM was proposed to model the complex information of fluorescence kinetic curves by comparing and analyzing the characteristics of fluorescence intensity profile.The identification function of different kinds of water and fertilizer concentration was realized.The reliability of water and fertilizer coupling on plant theory was verified.In the HMM training recognition, algorithm is applied to the research of fluorescence detection. The Baum-Welch algorithm is used to study the classification of dynamic fluorescence information from the aspects of iteration times, state value selection, and selection of training identification models. The initial matrix is ascertained as the fixed matrix. The optimal number of iterations, the number of states and the number of samples are analyzed. The combination of original chlorophyll profile and rate curve is selected as the characteristic information, which selects half of the sample points in this combination information. The best state of modeling and recognition can be achieved.In order to analyze the effectiveness of the method proposed in this thesis, taking the lucky tree as the research object, a part of the data sample is used for modeling, and the other is used to predict and statistically identify rates. The experimental scheme was proposed under different conditions, which analyse the correct rate of identification by quantitative and qualitative methods under different water and fertilizer concentration conditions. Study shows that the HMM algorithm can identify the redundant information under different growth conditions. The recognition rate can also be obtained. The increase of training sample is beneficial to the increase of recognition accuracy. The difference of high and low leaves in distinct plants at different locations is very obvious under the same stress. The effect of water on plant is slightly stronger than that of nitrogen fertilizer. The characteristics of stress datas are more obvious and easy to be identified in the process of modeling, which was consistent with the hydrophilicity of lucky tree. In the actual planting, certain concentrations can reflect the plant sensitivity,abnormal information and high stress characteristics under the conditions of the seedling period.These results are consistent with the law of plant growth and illustrate the complexity of the fluorescence phenomenon in the plant. The method proposed in this thesis can be used to guide the plant water and fertilizer.The fluorescence data analysis shows that the coupling phenomenon of water and fertilizer is determined by using the mathematical model of HMM, which can identify the interaction of water and fertilizer coupling. All of these indicate that the analytical method presented in this thesis which can be used to describe the change of the growth environment on the analysis of plant growth finely.