Phenology Detection At IBIMET

I was a visiting researcher at the Institute of BIoMETeorology (IBIMET) in Sassari Italy in the month of July 2013. My work there was directed at applying known research from the Richardson lab for using color responses to characterize phenological behavior. You can find the relevant research in this paper.

Research at IBIMET centers on monitoring phenological behavior of the different plant species in the north-west of Sardinia in the Porto Conte – Capo Caccia Nature reserve. The people at IBIMET-Sassari have installed a pan-tilt-zoom camera in a location of interest and have collected images during several months of the species growing in the natural reserve. I have compiled these images in a this movie which shows, among other things, the flowering of a species (around the 8th second). This is picked up by the excess green signal calculated with the Richardson paper concepts.

The gist of my work in Sassari can be summarized in the following figure:

The blue line represents the "raw" signal from the Excess Green Color space on the image series. The red represents the fitted Sigmoid signal. The date is calculated using the inflection point of the sigmoid and is when the flowering occurred.

The blue line represents the “raw” signal from the Excess Green Color space on the image series. The red represents the fitted Sigmoid signal. The date is calculated using the inflection point of the Sigmoid and is when the flowering occurred.

We see how the color signal can pick up relevant phenological variations. The idea is to automate the whole process and have an automata go through large amounts of image databases trying to detect these types of behavior.

Posted in Uncategorized | Tagged , , | Leave a comment

Leaf Counting and Identification.

Automating plant phenotyping. We have added an initial approximation to leaf counting and identification. This will detect the leaf centers and give the number of leafs of a plant. Inspiration came from a paper presented at CVPPP20143-D Histogram-Based Segmentation and Leaf Detection for Rosette Plants. Jean-Michel Pape et. al. 2014. And while there is still some false negatives that need to be addressed, I think the use of a simple distance map, together with local maximum is a pretty good starting point. Here is an example image from our data set:

The circles mark the places where we detect a leaf.

The circles mark the places where we detect a leaf. Notice how the plants in the right did not get detected, this is due to segmentation errors.

The circles mark the detected leaf centers. Some of the leafs of these plants are not detected because they don't translate into a detectable maximum in the distance transformation.

The circles mark the detected leaf centers. Some of the leafs of these plants are not detected because they don’t translate into a detectable maximum in the distance transformation.

Image | Posted on by | Tagged , , , , , , , | Leave a comment

Automatic Season Detection

It’s not very difficult for a human to know when a season starts or ends from seeing an image series. It becomes very tedious to do it when you have millions of images to go through. This is where automatic season detection can aid efforts in academic (e.g. research on global warming) and industrial (e.g. Agriculture) contexts.

Posted in Uncategorized | Tagged , , , | Leave a comment

Image Recognition and Machine Learning are Cool/Scary :)

Check this link out. Cool because emotions (smile == happy), age, identity and lots more can now all be gathered automatically but its scary because emotions, age, identity and lots more can all be gathered automatically.

Check out the Smart Me Up company web page. It shows a really cool way of presenting all these concepts as well as how far these fields have come.

Posted in Uncategorized | 1 Comment

Pan, Zoom with Qt

This is a followup post for my previous pan/zoom post. It is so difficult to find a proper widget that implements Zoom with mouse wheel and pan with CTRL+LeftClick. IMO, it is something that is necessary when viewing an image. I finally found a QT implementation that was fast and simple to understand. Here is the link to the Stack Overflow discussion. Hopefully this post will increase the likelihood of ppl finding a good solution in QT for this issue.

Posted in Uncategorized | Leave a comment

Sound from image

What will they think of next? This awesome MIT project recovers sound from a high-speed video feed. It’s awesome, scary and interesting all at the same time :)

 

Posted in Uncategorized | Leave a comment

Awesome Segmentation for Arabidopsis Thaliana

In the past months I have been working with a project which has the purpose of addressing high throughput phenotyping. We recently had a very successful test that I want to share as a video. What we are trying to do is extract plant data automatically from individual images. Here is the video :)

 

Posted in Uncategorized | Tagged , , , , , | Leave a comment