Dissertation > Agricultural Sciences > General theory > Agricultural scientific research principles, policies,and its elaborate

Approaches for Processing the Field Spatial Data and Practical Study for Precision Agriculture

Author LiuGang
Tutor WangZuoHua;ZuoPuSheng
School China Agricultural University
Course Agricultural Electrification and Automation
Keywords Precision Agriculture DGPS GIS Spatial Variability Geostatistics Neural Network
CLC S-01
Type PhD thesis
Year 2001
Downloads 556
Quotes 24
Download Dissertation

In these years, China keeps carrying out agricultural tech revolution, and focus closely on R&D of Precision Agriculture all over the world. Precision Agriculture抯 practice and rapid development could make us mostly free the latent capacity of farming land, increase efficiency of water and fertilizer using, save our environment from being polluted, and greatly improve crop yield and quality. Nowadays, Precision Agriculture is the main direction of continual agriculture development, and will be a very important research area that will accelerate agricultural tech revolution of China, at the beginning of the twenty first century.There are a series of science and technology in researching and practicing Precision Agriculture. The basic concept of Precision Agriculture is according to the spatial and temporal variability in field, treat the crop individually, generally optimize crop management and increase efficiency and benefit. Obviously, field information collection and process and field status description are the most important tasks. Especially, the advanced sensing technology for the real time fast field data collection, field spatial information distribution processing method, variable rate crop management technology and system analysis are the most important of all. And all of these need the support of agronomy, agriculture engineering, math, technical economics, and computer technology.At present, about processing field spatial distribute information, there are many research achievements published in journals around the world. But since the complexity of crop growing environment within the field, there is no such a method that could deal different conditions such as cultivation, and region, so new methods are still needed. In these two years, related research is carried out in China, especially about pattern of field soil fertility variability. But few publications are about finding the determined field spatial variable factor and quantificational describing the relationship between yield and field spatial information distribution.This paper based on Precision Agriculture development of China, and focused on several key problems for developing Precision Agriculture, especially field spatial information distribution processing method. The trail farm is located in Shunyi, Beijing. The area is 11 hectors. We sampled the soil, and got soil spatial distribute data (moisture, fertility), crop growing status distribute data (such as crop population, crop height, plant dry weight), yield distribute data (such as yield per mu, ears per mu, seeds number per ear, weight of thousand grain, plant yield). Then we use statistics and geostatistics analyzed field information spatial variability, and find out the relationship between yield and field spatial information distribution, and the neural network was used for the first time to deal with field spatial information distribution. A neural network model is built and tested for analyzing yield and field spatial information distribution. The main conclusion of the paper are listed as below:1)We should research, practice and demonstrate Precision Agriculture according the condition of China, the achievements of foreign countries, in different regions and different levels at the same time. And our research and practice should focus on Precision Irrigation and Precision Fertilizing.2)Reasonable field sampling method, suitable spatial interpolation method, is the guarantee of getting a field spatial variability information with higher accuracy and lower cost. Systematic grid sampling is the most popular method at present, but it could loss some important distribution information. Systematic unaligned sampling is the most ideal method. It is easy to find periodical distributing field information with systematic unaligned sampling method. Distance reverse interpolation is a relatively accurate interpolation method, but the area of neighborhood affects its precision. Kriging interpolation needs large amount of calculation, but it is a relative

Related Dissertations
More Dissertations