OVERVIEW

We are exploring possibilities of using a network of low power wireless cameras for sensing applications. Our goal is creating a class of distributed lightweight image processing algorithms suitable for less capable computational platforms. These algorithms should enable the sensor nodes to acquire Meta-Data from the images, share that information across the network and enable network micro-server to convey coherent knowledge from this distributed information.

While such technology can be used for many applications we have focused on habitat ecology as our science motivation. There are many potential questions that cameraNet can explore. Estimation of color density of leaves conveys information that possibly in combination of other meteorological sensors can be used in prediction of health of the trees or for irrigation purposes in precision agriculture. Creating an occupancy profile of the birds nest using imagers help us, understand their life cycle while they prepare their nest, lay eggs and finally leave their habitat. Probabilistic estimation of motion contents of different cameras can be combined to form a belief about the amount of motion in the scene. Color cameras can be used to get the statistical contents of color to estimate visible portion of the incident light spectrum.

SYSTEMS / EXPERIMENTS

We are currently collaborating with Agilent Corp. in building the following hardware:

• A sensor board that can be attached to the Mote. It will carry the camera module, its own microcontroller, RAM and Flash memory. It is responsible for capturing images and operating low level commands.

We will also port TinyOS to the sensor board microcontroller and build following software modules:

• Set of modules on sensor board for capturing and controlling images,
• Set of modules for sending and receiving pictures to/from host.
• Image processing toolbox.
• Micro-Server tool box.

ACCOMPLISHMENTS

We have just started this project. Components have been selected. The first version of schematic has been prepared by Agilent and it is under discussion.

FUTURE DIRECTIONS

• Completed schematic and layout of the sensor board.
• Design and implementation of the software architecture and necessary components of host platform and sensor platform.
• Design and implementation of image processing toolbox.

PEOPLE

Faculty:

Prof. Deborah Estrin
Prof. Bill Kaiser
Prof. John Villasinor

Staff:

Mohammad Rahimi

Industry Partners:

Dr. Rick Baer (Agilent)
Dr. Jay Warrior