The Study Of Soil Heavy Metals Concentration Prediction Based On Neural Network

With the rapid development of agriculture and urbanization, part of the soil has been polluted by heavy metals. Plants would absorb heavy metal from soil due to plants have absorption character. Thus, this will cause food pollution, and it will pose significant health impact to humans by biologic chain. Therefore, the contents of heavy metals in the soil have important implication for soil quality evaluation, protecting human health and maintaining social economy sustainable development.The characteristics of heavy metal concentration changing in soil are nonlinearity,large time delay and so on. It is difficult to set up a precise model to predict the concentration of heavy metal in soil. Based on Neural network has the advantages of self adjustment, self adaptability, self-learning and fault tolerance, these virtues is very suitable for dealing with complex problems such as soil prediction. Using neural network model and MATLAB neural network toolbox, and through adaptive learning method and dynamic optimization model,the prediction model of heavy metal content in soil was established. The main contents are as follows:(1) The introduction of neural network. The key factors are the working principle of neural network and analyzing of building prediction model. At the same time this easy will calculation neural network and explanation hidden layer.(2) This article is introducing how to build prediction system, which is working principle and basic methods (input layer, hidden layer, output layer and the parameters set).(3) Using high technology (Italy HANNAHI83200) to make a soil testing. The prediction model was established by the correlation between heavy metals, analyzing the cause of errors that may exist through horizontal and vertical contrast.(4) Finally, making a summary of the article and the bug of detection system. And make a direction of the following work.