Image Analysis Approaches for Plant Phenotyping

In our post-genomic world, where ever increasing volumes of genetic information are obtained at great speed and little cost, the collection of phenotypic information is often a bottleneck to scientific progress. In stark contrast to genotyping, phenotyping is slow, and expensive in human time. Moreover, measurements are affected by the varying perception and interpretation of different observers. Image analysis has the potential to overcome these problems, but automatic interpretation of images of plants and animals remains very difficult.

My talk describes recent work on combining stereo and Time-of-Flight (ToF) images to estimate dense depth maps in order to automate plant phenotyping. I will also briefly describe another direction investigating statistical characteristics of the images for automatic plant phenotyping, and provides an overview of an EU-funded FP7 project, SPICY (Smart tools for Prediction and Improvement of Crop Yield).

Date and Time

2pm on May 10, 2011 in CS.101