As part of my PhD I spent 6 months as a visiting researcher in the CENS project at UCLA. These very productive 6 months resulted in the creation of a toolkit and the analysis of image series that were collected during 6 years. This page describes and links to the toolkit.
EcoIP is a toolkit that facilitates the automation of phenological data analysis derived from digital image series. It uses a Naive Bayesian Model to generate a probabilistic representation that it then applies to new images in the hope of getting responses for phenological phases. It is implemented as an open platform that encourages everyone to run, copy, distribute, study, change and improve the application. If contributions do occur, we hope to funnel them into the existing repository. All the code is made available for browsing or download and examples are provided for every functionality to aid the user. EcoIP is designed as an R (R version 2.15.0, R Team, 2012) package and it can easily be installed in all platforms where R runs. Help is given to the user by providing documentation, sample data and executable examples. Two data sets are provided: One of wallflowers in a meadow (Erysimum capitatum) and another of an Oak tree (Quercus sp.) which the user can manipulate before tackling more complicated tasks. EcoIP has other functionalities relevant to its main goal such as color channel comparison, allowing small tweaks in signal generation and features to facilitate plotting. Here is the paper that was published for this project.
The following prezi presentation outlines what we did