Other resources:
: PDF (6.3MB)
: PPT (2.3MB)
Abstract: Cameras are a natural choice of sensors for monitoring natural habitats. They are cheap, both in terms of cost and energy, and at the same time data-rich. They are remote, not requiring contact, and passive, not requiring to broadcast signals into the environment, and at the same time they can be tuned to be sensitive to different bands, most commonly the visible and near-infrared spectra.

However, their use in natural habitats presents significant challenges, due to the variability that objects or events of interest can manifest depending on "nuisance factors" such as illumination, vantage point and occlusion. For instance, the image of a moss is a function of its CO2 uptake, which is of interest, but also of the illumination, depending on the time of the day, the day of the season, the weather, and the presence of cast shadows. It also depends on the vantage point and the geometric layout in three dimensions, and the presence of occlusions of line-of-sight. Similarly, events of interest exhibit complex dynamics, for instance the motion of birds or the configuration of a swarm of pollinators, but so do nuisances such as the complex background foliage moving under the elements. (September 2009)

Other resources:
: PDF (13MB)
: PPT (4.3MB)
Abstract: There exist many natural phenomena where direct measurement is either extremely invasive or impossible. For example, measuring a plant's CO2 flux requires enclosing a plant in a chamber and measuring pollination events for a given flower requires a human. Instead of using traditional sensors (like temperature, light, etc), we propose using imagers as sensors, since they are the most natural means of observing these phenomena. Like all other sensors, image sensors require calibration; however, our calibration is much more complex since we aren't measuring the phenomena directly. Our specialized calibration is composed of state-of-the-art computer vision, image processing, and statistical learning algorithms. In this talk, we will present a few applications, associated calibration procedures, and preliminary results. (July 2009)

Abstract: Relative spectral reflectance is an illumination invarient image feature that is related to many ecological phenomena that are difficult to measure, such as plant CO2 uptake. We describe a procedure to estimate the relative spectral reflectance of known subject using color image features. Through application, we show that this procedure produces accurate estimates in the presence of changing field conditions. Using this procedure, we can use imagers as sensors to measure natural phenomena that cannot be easily measured using any other available sensing modality. (November 2008)

Abstract: There exist many natural phenomena where direct measurement is either impossible or extremely invasive. We propose using imagers as sensors by constructing a procedure that uses images to obtain approximate measurements of these phenomena. This procedure, composed of state-of-the-art computer vision, image processing, and statistical learning algorithms, will be evaluated in the context of a specific application and shown to be general through multiple instantiations. We show through application, that many of these algorithms make unacceptable assuptions about their input. We will describe a methodology that can be used to augment existing algorithms, making them robust to field conditions present in ecological applications. In this paper, we rigorously define the proposed procedure and begin to evaluate its accuracy in the context of an example application. (September 2008)

Abstract: There are questions which can't easily be answered by traditional ecological sensors. Indirect sampling of phenomena through image analysis can provide a non-invasive and inexpensive means of answering these questions. Domain specific knowledge will serve to make this image analysis more tractable than general vision. This presentation serves as an introduction to our vision and preliminary results. (October 2007)

Abstract: There exist many natural phenomena where direct measurement is either impossible or extremely invasive. To obtain approximate measurements of these phenomena we can build prediction models based on other sensing modalities such as features extracted from data collected by an imager. These models are derived from controlled experiments performed under laboratory conditions, and can then be applied to the associated event in nature. In this paper we explore various different methods for generating such models and discuss their accuracy, robustness, and computational complexity. Given sufficiently computationally simple models, we can eventually push their computation down towards the sensor nodes themselves to reduce the amount of data required to both flow through the network and be stored in a database. The addition of these models turn in-situ imagers into powerful biological sensors, and image databases into useful records of biological activity. (September 2007)

Abstract: In order to increase the quality and quantity of Computer Scientists graduating for universities, we need to spark their interest early. In this talk, I talk about a few differnt project ideas tailored to students of various ages to peak their interest. In practice, I have found that these particular projects originally caught the attention of many computer scientists I know and work with. (July 2007)