Dissertation > Excellent graduate degree dissertation topics show

Detection of Pine Forest Area Changes in Regions Damaged by Bursaphelenchus Xylophilus in Zhejiang Province

Author: JinWei
Tutor: GeHongLi
School: Zhejiang Forestry College
Course: Forest Management
Keywords: Remote sensing Pine wilt disease Pinewood Water Construction land Linear transformation method
CLC: S763
Type: Master's thesis
Year: 2009
Downloads: 48
Quote: 0
Read: Download Dissertation

Abstract

Zhejiang pine wilt disease infested areas. The purpose of this paper is the indirect assessment of pine wilt disease caused by remote sensing dynamic monitoring of the hazard area pine forest area change loss. In this paper, the remote sensing estimated area change process of the pine forest of pine wilt disease hazard area in Zhejiang Province since 1994. Use of a total of 3 King, 4 of 12 King remote sensing image and Zhejiang ground forest survey data. The results are consistent with the actual situation. Proposed linear transformation classification. Having a classification value according to the selected two or two or more bands, these bands are arranged according to certain rules (in descending or small to large), the luminance value as the dependent variable, the serial number as an independent variable to create a linear regression equation , then the slope and intercept of the linear equations as a basis for classification. This method in water and non-water separation, the application test land for construction and non-construction land, coniferous and non-coniferous forests separation, to good effect. Especially the separation of the first two types of ground, the accuracy is significantly higher than the maximum likelihood method. The advantage of this method when the number of bands is more than two more significantly, it can be compressed into 2-dimensional space of the high-dimensional data, in addition to the compressed data, and more on the plane can be intuitively data characteristic analyzed. High-dimensional nonparametric lookup table classification. Valuable first step to select the band; second step to establish a single class of high-dimensional histogram, and histogram smoothing; the third step, a single class of histogram density values ??converted into probability values; fourth step The probability histogram convert high-dimensional lookup table. Certain class classification characteristic bands. Such as the separation of water and non-water use TM2, 3, 4, 5 band, and compressed four-dimensional space to two-dimensional space using a linear transformation method, these bands and this approach can better distinguish between water and shadow. Separation of land for construction and non-construction land can take advantage of TM3, 4,7 band, coniferous and non-coniferous forest can take advantage of TM4, 5 band. Concluded, according to the analysis of the spectral curve of pine and fir on the data of 12 King TM data rely solely on spectral information is not sufficient to distinguish between these two species. According to this situation, this paper, Zhejiang Province, local the ground forest resources survey data pine forest area is estimated as a secondary data. Pine / fir forest remote sensing estimated by the ratio of the area of ??coniferous forest, you can get the King pine forest area estimates, coniferous forest remote sensing maps. According to the remote sensing estimates, 13 years of Zhejiang entire area of ??pine wilt disease hazards pine forest area decreased by 26.1%. The pine area change is not the only reason to pine wilt disease hazards, such as Zhejiang Province in recent years to promote the the needle change width \The decrease in the area of ??pine wilt disease hazards in the end how much, and also require a lot of research.

Related Dissertations

CLC: > Agricultural Sciences > Forestry > Forest Protection > Forest Pest and Disease Control
© 2012 www.DissertationTopic.Net Mobile