Scribd
Upload
38 views12 pages

Time Series Analysis: Johannes Reiche, Johannes Balling and Eliakim Hamunyela

Time series analysis of satellite imagery allows for accurate monitoring of forest change dynamics over long periods of time. Analyzing full time series provides information about the timing, magnitude and speed of forest changes like logging or recovery that bi-temporal analysis cannot. Different time series analysis methods exist for break detection, segmentation and probabilistic analysis of forest disturbance that can help near real-time monitoring using dense satellite collections like Sentinel-1 and -2.

Uploaded by

Zamazam Hasen
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
38 views12 pages

Time Series Analysis: Johannes Reiche, Johannes Balling and Eliakim Hamunyela

Time series analysis of satellite imagery allows for accurate monitoring of forest change dynamics over long periods of time. Analyzing full time series provides information about the timing, magnitude and speed of forest changes like logging or recovery that bi-temporal analysis cannot. Different time series analysis methods exist for break detection, segmentation and probabilistic analysis of forest disturbance that can help near real-time monitoring using dense satellite collections like Sentinel-1 and -2.

Uploaded by

Zamazam Hasen
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 12

Time series analysis

Johannes Reiche1, Johannes Balling1 and Eliakim


Hamunyela2

1Wageningen University
2University of Namibia
Why time series analysis?

[Figure from Vos et al 2019]


Why time series analysis?

[Photo by Brent Stirton/Getty Images]

[Credit: Illustration by Nathan Johnson/PNNL]

[Photo by Mongabay.com]

[Credit: Verbesselt et al., 2010]

• Understand long-term forest changes and dynamics


• Near real-time change detection and alerting 3
[Photo by Daniel Beltra /Greenpeace]
Satellite images for monitoring tropical forest
Bi-temporal versus time series
Gold mining

Forest logging

[Photo by Johannes Reiche/Wageningen


University]
Landsat image (Guyana), 2015 [Credit: Johannes Reiche/Wageningen University]

Landsat image (Guyana), 2019


Satellite images for monitoring tropical forest
Bi-temporal versus time-series
Bi-temporal image
data

[Graphic by Johannes Reiche/Wageningen University]

Bi-temporal change detection (e.g. image differencing; delta classification)


+ conceptually simple and quick
+ easy to interpret
- no accurate timing of changes or understanding of dynamics
- inaccurate estimate of change magnitude
Satellite images for monitoring tropical forest
Bi-temporal versus time-series
Time-series data

forest

recovery
logging

[Graphic by Johannes Reiche/Wageningen University]

Time series analysis


+ accurate monitoring of forest change dynamics and timing
Satellite images for monitoring tropical forest
Bi-temporal versus time-series
Time-series data

change
information
post-change
information

[Graphic by Johannes Reiche/Wageningen University]*


Time series analysis
+ accurate monitoring of forest change dynamics and timing
+ analysing change (e.g. change magnitude) and post-change parameters (e.g. recovery speed,
follow up land use)
- conceptually more advanced
- mainly pixel-based
Satellite images for monitoring tropical forest
Time series analysis methods (examples)
• Break-detection
• BFAST (Verbesselt et al., 2010)
• CCDC (Zhu & Woodcock 2014)
• ....

[Source: Verbesselt et al., 2010]


Satellite images for monitoring tropical forest
Time series analysis methods (examples)
• Break-detection
• BFAST (Verbesselt et al., 2010)
• CCDC (Zhu & Woodcock 2014)
• ....

[Source: Zhu & Woodcock 2014]


Satellite images for monitoring tropical forest
Time series analysis methods (examples)
• Break-detection
• Segmentation
• LandTrendr (Kennedy et al., 2010)
• ....

[Source: Kennedy et al., 2010]


Satellite images for monitoring tropical forest
Time series analysis methods (examples)
• Break-detection
• Segmentation
• Probabilistic approaches
• Bayts (Reiche et al., 2015)
• ...

Landsat NDVI
Senintel-1 VV-backscatter
Near real-time forest disturbance
detection (Sentinel-1 + Landsat)
[Source: Johannes Reiche/Wageningen University]
Satellite images for monitoring tropical forest
Conclusions
• Time series analysis
• Understand long-term changes and dynamics
• Near real-time change detection
• Selection of time series methods depends on application
• Dense & high resolution Sentinel-1 (radar) and -2 (optical) time
series available
Deforestation detected in a
dense Sentinel-1 time series
(Indonesia)

[Source: Johannes Reiche/Wageningen University]

You might also like