detectBayts: Calculate change probability and detect changes

In jreiche/bayts: Probabalistic approach to combine multiple time series and Bayesian probability updating for near real-time forest change detection

Description

Iterative Bayesian updating of the conditional probability of change (PChange) based on PNF and detection of change

Usage

detectBayts(bayts, chi = 0.5, PNFmin = 0.5, start = NULL, end = NULL)

Arguments

bayts	"bayts" time series data frame created with createBayts
chi	Threshold of Pchange at which the change is confirmed; Default=0.5
PNFmin	threshold of pNF above which the first observation is flagged; Default=0.5
start	Start date of monitoring period. Default=NULL (start of "bayts" time series data frame).
end	End date of monitoring period. Default=NULL (end of "bayts" time series data frame)

Details

Short method description: Observations at time t (current observation) are flagged to be potentially changed in the case that the conditional NF probability (PNF) is larger than 0.5. For a flagged observation, the conditional probability of change (PChange) is computed by iterative Bayesian updating (calcPosterior), using the previous observation (t − 1), the current observation (t), as well as i upcoming observations (t + i) to confirm or reject a change event at t. A change is confirmed in case PChange exceeds a given threshold "chi". A detailed description is provided in Reiche et al. 2015 (Chapter 2.1.4).

Value

Updated "bayts" time series data frame with changes if detected.

Flag: "0" = no change flagged; "Flag" = change flagged (iterative Bayesian updating is ongoing); "oldFlag" = old flagged change that was rejected; "Change" = confirmed change and observations that were initially flagged as change

Author(s)

Johannes Reiche (Wageningen University)

References

Reiche et al. (2015): A Bayesian Approach to Combine Landsat and ALOS PALSAR Time Series for Near Real-Time Deforestation Detection. Remote Sensing. 7(5), 4973-4996; doi:10.3390/rs70504973

jreiche/bayts documentation built on Feb. 3, 2021, 1:12 a.m.