Advanced

Developing a web-based system to visualize vegetation trends by a nonlinear regression algorithm

Wei, Yufei LU (2016) In Student thesis series INES NGEM01 20161
Dept of Physical Geography and Ecosystem Science
Abstract
Comparing with traditional linear regression methods that used to monitor vegetation trends, a nonlinear regression algorithm (PolyTrend) developed by Jamali et al. (2014) can provide more accurate information of vegetation trends by fitting a polynomial line with a degree of up to three for ASCII-formed time-series NDVI (Normalized Difference Vegetation Index) dataset of a single pixel. To extend the ability of the PolyTrend algorithm for processing time-series NDVI satellite imagery and to increase its accessibility, a web-based system for visualizing vegetation trends by the PolyTrend algorithm has been developed. The PolyTrend web-based system allows users to define the value of statistical significance of the PolyTrend algorithm, the... (More)
Comparing with traditional linear regression methods that used to monitor vegetation trends, a nonlinear regression algorithm (PolyTrend) developed by Jamali et al. (2014) can provide more accurate information of vegetation trends by fitting a polynomial line with a degree of up to three for ASCII-formed time-series NDVI (Normalized Difference Vegetation Index) dataset of a single pixel. To extend the ability of the PolyTrend algorithm for processing time-series NDVI satellite imagery and to increase its accessibility, a web-based system for visualizing vegetation trends by the PolyTrend algorithm has been developed. The PolyTrend web-based system allows users to define the value of statistical significance of the PolyTrend algorithm, the nominal range of the input data, and the range of desired NDVI input to be processed. It applies the PolyTrend algorithm to each pixel of the uploaded time-series NDVI satellite imagery dataset. It returns the types of vegetation changes, the slope of the changes of NDVI values in the whole time span, and whether the net change of NDVI increases or decreases during this period, in the forms of ASCII files (i.e. text files) and binary files (i.e. images). By refining the existing PolyTrend algorithm written in MATLAB and embedding it in a web environment, the PolyTrend web-based system has proved its ability in monitoring global vegetation trends using raw time-series NDVI satellite imagery. (Less)
Popular Abstract
An index called NDVI (Normalized Difference Vegetation Index) has been widely used to describe the reflectance characteristics of land features (Lillesand et al. 2008). The temporal dynamics of vegetation (i.e. vegetation trends) can be gained by monitoring the changes of NDVI. In all publications founded by the author, straight-line relationships have been used to describe vegetation trends. However, straight lines cannot fit to all real-world situations of vegetation growth. An algorithm (PolyTrend) developed by Jamali et al. (2014) solved this problem by assuming cubic-polynomial relationships exist in vegetation trends at the beginning. However, the original version of this algorithm could only accept time-series NDVI values of a... (More)
An index called NDVI (Normalized Difference Vegetation Index) has been widely used to describe the reflectance characteristics of land features (Lillesand et al. 2008). The temporal dynamics of vegetation (i.e. vegetation trends) can be gained by monitoring the changes of NDVI. In all publications founded by the author, straight-line relationships have been used to describe vegetation trends. However, straight lines cannot fit to all real-world situations of vegetation growth. An algorithm (PolyTrend) developed by Jamali et al. (2014) solved this problem by assuming cubic-polynomial relationships exist in vegetation trends at the beginning. However, the original version of this algorithm could only accept time-series NDVI values of a single pixel stored in a text file.

To enable the PolyTrend algorithm to process image-level information of NDVI and to disseminate this algorithm, an online system (i.e. the PolyTrend web-based system) that includes this algorithm was developed to allow users to upload raw time-series satellite imagery containing NDVI values that gained from the Internet. After the PolyTrend algorithm processes the imagery, the PolyTrend web-based system returns values of the range and the inclination of the net changes of NDVI and the types of vegetation trends classified by the algorithm. These results are downloadable in the forms of images and text files with explanation. The images map the temporal dynamics of vegetation directly while the text files can be imported to other software for generating other forms of data and revealing statistical results.

The PolyTrend web-based system provides a convenient way of monitoring global vegetation trends through the Internet with an innovative algorithm. (Less)
Please use this url to cite or link to this publication:
author
Wei, Yufei LU
supervisor
organization
course
NGEM01 20161
year
type
H2 - Master's Degree (Two Years)
subject
keywords
PolyTrend, nonlinear regression algorithm, vegetation trends, Physical Geography and Ecosystem Analysis, NDVI, Web Development, Django, MATLAB, Python
publication/series
Student thesis series INES
report number
387
language
English
id
8882588
date added to LUP
2016-06-20 15:22:33
date last changed
2016-06-20 15:22:33
@misc{8882588,
  abstract     = {Comparing with traditional linear regression methods that used to monitor vegetation trends, a nonlinear regression algorithm (PolyTrend) developed by Jamali et al. (2014) can provide more accurate information of vegetation trends by fitting a polynomial line with a degree of up to three for ASCII-formed time-series NDVI (Normalized Difference Vegetation Index) dataset of a single pixel. To extend the ability of the PolyTrend algorithm for processing time-series NDVI satellite imagery and to increase its accessibility, a web-based system for visualizing vegetation trends by the PolyTrend algorithm has been developed. The PolyTrend web-based system allows users to define the value of statistical significance of the PolyTrend algorithm, the nominal range of the input data, and the range of desired NDVI input to be processed. It applies the PolyTrend algorithm to each pixel of the uploaded time-series NDVI satellite imagery dataset. It returns the types of vegetation changes, the slope of the changes of NDVI values in the whole time span, and whether the net change of NDVI increases or decreases during this period, in the forms of ASCII files (i.e. text files) and binary files (i.e. images). By refining the existing PolyTrend algorithm written in MATLAB and embedding it in a web environment, the PolyTrend web-based system has proved its ability in monitoring global vegetation trends using raw time-series NDVI satellite imagery.},
  author       = {Wei, Yufei},
  keyword      = {PolyTrend,nonlinear regression algorithm,vegetation trends,Physical Geography and Ecosystem Analysis,NDVI,Web Development,Django,MATLAB,Python},
  language     = {eng},
  note         = {Student Paper},
  series       = {Student thesis series INES},
  title        = {Developing a web-based system to visualize vegetation trends by a nonlinear regression algorithm},
  year         = {2016},
}