Sparse sensor based silhouette detector

This is my undergraduate thesis project. It is an intrusion detection system that uses an array of infrared LEDs on a pole, and an array of corresponding sensors on another pole placed across from the path to be monitored. If an object goes through the path, the sensors detects and records a silhouette of the object which can be analyzed to classify the object. The system is awaiting field deployment.

elephant-sensor
Early prototype of the detector, sans the casing.

The project was motivated by the need to collect elephant demography -population age structure- in the wild.  From the data collected, the height of the passing object can be known to an accuracy of less than one inch. Since height is highly correlated to age in elephants, the system can be used to access elephant demography.

The system is designed to be cheap and low power while being robust to the elements. The system works in any lighting condition, is much more immune to fog than a visual system, works even if partially covered in mud, and requires little post processing of data. For more information, read my undergraduate thesis.

humble-beginnings
Humble beginnings of the sensor in my dorm room.
case_3d
A quick mockup of proposed casing.

I would like to thank Dr. Karpagam Chelliah for inspiring this project as well as for helpful discussions and continued moral support.