Sparse Detector Imaging Sensor with Two-Class Silhouette Classification.

Area of Science:

• * Optoelectronics and Imaging Systems
• * Machine Learning for Pattern Recognition

Background:

• * Traditional imaging sensors can be costly and power-intensive.
• * There is a need for cost-effective, low-power surveillance solutions.

Purpose of the Study:

• * To design and test a novel active near-infrared sparse detector imaging sensor.
• * To evaluate algorithms for classifying object silhouettes.
• * To assess the sensor's potential for surveillance applications.

Main Methods:

• * Development of a prototype sensor with a 16-element sparse detector array.
• * Collection of object silhouette data (human, animal, vehicle).
• * Algorithm evaluation for binary classification (human vs. non-human).
• * Integration of a Webservice interface for network-centric operation.

Main Results:

• * Achieved >99% classification accuracy for human vs. non-human objects.
• * Demonstrated effective silhouette capture of various moving objects.
• * Prototype includes a Webservice interface for networked deployment.

Conclusions:

• * The sparse detector imaging sensor is a promising low-cost alternative for certain applications.
• * Potential for deployment in intelligent electronic fences and persistent surveillance.
• * Further optimization and testing are recommended for widespread adoption.