Cao Liying

Application of a spatial fuzzy clustering algorithm in precision fertilisation

Abstract In precision agriculture there exist two sorts of data—high‐dimensional space data and property data. We propose a precision fertilisation method based on a spatial fuzzy clustering algorithm and apply it to the National 863 Project “Corn Precision Task Systems Research and Application”. Our research methods include: (1) use of DGPS and GIS to obtain and to process soil spatial information and property information; (2) use of fuzzy clustering analysis to carry out systematic soil nutrient research and to develop classification models; (3) use of the 8‐adjacent connection traverse algorithm to carry out spatial clustering analysis and to apply the results to spatial clustering. This two‐stage clustering method is superior to the more traditional single‐stage clustering methods. Based on clustering results, we can make corn precision fertilisation decisions and can also make decisions relating to other precision agricultural field management issues and operations. It is shown by experiment that this method has good prospects of practical application.

A network intrusion detection method based on combined model

In order to make the detecting rate faster and improve the accuracy of network intrusion detection, this paper ameliorated a network intrusion detection method which was based on combining support vector machines and LVQ (Learning vector quantization) neural network algorithm The method combines the popularizing capability of SVM and the learning capability of LVQ neural network. It overcame the shortcomings of traditional neural network algorithm, such as the slower learning speed and the larger possibility of falling into local minimum. Examples proved that this combined model had faster speed and higher rate of accuracy. What is more, it better resolved a series of detecting problems, such as nonlinearity, small-sample, high-dimension and local minimum.

Study and application of grain yield forecasting model

There are some poor accuracy problems of grain yield prediction. GM (1, 1) prediction model and ARIMA (1,1,1) prediction model were established according to Jilin Province 1998-2011 grain yield data. In the same training sample, 2 kinds of methods were used to forecast grain yield. The average error is 7.88% and 12.32%, the average precision accuracy is 92.12% and 87.68% respectively. The test results show that the average prediction accuracy of grey system is higher than that of the time sequence model, and it can be applied to the prediction of grain yield.

Research on image vectorization based on mathematical morphology

Image vectorization plays an important role in the digital image processing. Because the traditional linear vectorization methods have some shortcomings including processing data slowerly, being sensitive to noises and being easy to be distorted, this paper proposes an image vectorization method based on mathematical morphology. This method consists of the image edge detection method based on the four structural elements template, the morphological sequential homotypic skeleton abstraction method based on the eight structural elements template and the vectorization method based on the dynamic change of pace about Freemans chain code. Examples have proved that this method can abstract high precision skeletons rapidly and attain high accuracy vector data measured by the arc segment. This method has many advantages, such as the faster processing rate, the higher accuracy and the less storage space. So it has certain feasibility and practicability in digital image vectorization.

A study on evaluation of farmland fertility levels based on optimization of the decision tree algorithm

Base on the organic matter, total nitrogen, available phosphorus and potassium content in the cultivated land of The 20 District in Dehui 20, it comes a study on using the decision tree ID3 algorithm for the evaluation of farmland fertility levels. Through experimental analysis, it gets the number of grade 6, grade sub-region soil nutrient content is similar or smaller differences, and significant differences among grades. Fertility Evaluation based on the ID3 algorithm farmland can be used for variable rate fertilization in the guidance of precision agriculture, to provide an effective division for farmland fertility level partition.

Research on evaluation of college teachers based on support vector machine algorithm

In order to reduce man-made mistakes and errors in the process of the evaluation of college teachers, this article applies the support vector machine theory to the evaluation system and sets the college teacher evaluation indicators according to some requirements by combining current college teacher evaluation indicators in colleges in accordance with related theories such as pedagogy. We divide the sample data into two parts with SVM principle and attain the training model through training the sample data in the evaluation system and take the intelligent evaluation and analysis on the prediction data with the training model. The experimentation and trial run showed that we can make the evaluation more accurate, reasonable and objective by taking the method based on SUM in evaluating the teacher evaluation data. So the prospect of the method is better.

Delineating soil nutrient management zones based on ID3 algorithm

Focused on the contents of organic matter, total N, available P and available K in patch data of Dehui city, Jilin Province, this paper studies on soil nutrient management zones using decision tree ID3 algorithm. By experimental analysis, we obtain six partitions and the result shows that sub-area differences in soil nutrient content are similar or less significant but differences between partitions. The method of soil nutrient management zones based on decision tree ID3 algorithm can be used to guide the variable-rate fertilization in precision agriculture, and it also provides an effective classification method for soil nutrient management zones.