• ภาษาไทย
  • English

Geo-Informatics and Space Technology Development Agency (Public Organization):GISTDA

GISTDA To be Value based Organization in Geo-space Intelligence Deliveries by 2020

Field Server (Part 5): Interpolation and Smooth of Vegetation Phenology Graph

Field Server (Part 5): Interpolation and Smooth of Vegetation Phenology Graph

The previous part of Field Server described about segmenting AOI (Area of Interest), which in this place means “rice plot.” When obtain AOI, it will be built as Binary mask in order to find Average Excessive Green (ExG) [1] of all image pixel which is rice plot. Average ExG of each time series image will be used building Vegetation Phenology Graph as showing in picture 1.

05_fig_01

Picture 1 Using average ExG of each rice plot area image to create Vegetation Phenology Graph

Practically, daily image might be missing because of some obstacles, for example, problems from electricity or communication system as showing example of graph in picture 1 (Below). Therefore it is necessary to evaluate by Interpolation for the acquired Vegetation Phenology Graph. At the first stage, there are many format of Interpolation. Herein, the example is about fundamental format such as Linear interpolation, which is Curve fitting on Linear polynomials or Equation of the Straight Line. The “(x,y) point that needs in interpolation could be calculated from the equation as showing in picture 2. Linear Interpolation takes less time in calculation and effective to Interpolation. The missing data from the graph has high characteristic of Linear Interpolation. Other Interpolation format such as Polynomial or Splineinterpolation will have more complex calculation and the effectiveness is good to the graph format in various cases.

05_fig_02

รูปที่ 2 การประมาณค่าในช่วงแบบเชิงเส้น (Linear interpolation)

05_fig_03a

ก) หลังจากทำInterpolation

05_fig_03b

b) Smooth by using Savitzky-Golay filter

Picture 3 Vegetation Phenology Graph a) After Interpolation b) Smooth by using Savitzky-Golay filter

Vegetation Phenology Graph after Interpolation as showing in picture 3 a) the missing data will be estimated and filled in graph. Normally in the next process, graph will be Smooth. The popular method being used is Savitzky-Golay filter [2] showing in picture 3 b) Vegetation Phenology Graph will be Smooth to reduce the up/down swing of signal (Frequency), making it easier for analyzing graph and useful for furthermore evaluation.
As for the next part, we will talk about how Vegetation Phenology Graph in Smooth format could be specified the starting of growing season and ending of growing season date of each automatic cultivation cycle…Until next time…

Note
• Interpolation is the method of Curve fitting of Vegetation Phenology Graph in order to estimate data value on the location where the image is missing.
• The Interpolation method being used is Linear Interpolation which is the easy fundamental method. It is applied in a variety of signal evaluations, for example, graph (1D), picture (2D) or 3D data (3D).
• Vegetation Phenology Graph will be Smooth by using Savitzky-Golay filter.

 

Tags: