Grid-EYE Dataset for Human Occupancy Detection and Counting in a Study Pod

This dataset contains thermal sensor data and corresponding ground truth annotations collected during controlled occupancy monitoring experiments in a library study pod environment. The data was gathered to evaluate a privacy-preserving human detection and tracking system using a low-resolution 8\u00d78 Panasonic Grid-EYE AMG8833 thermal sensor, Installed in the upper corner of an indoor 2-seater library study pod (dimensions: 2.20m \u00d7 1.05m \u00d7 2.10m). The dataset includes three 10-minute continuous recording sessions representing different occupancy scenarios: empty pod (0 persons), single occupant (1 person), and dual occupancy (2 persons). Each session contains synchronized thermal sensor readings processed through a deterministic algorithmic pipeline, blob detection visualization outputs, and ground truth video recordings for validation purposes. The CSV file contains timestamped records with a co-located DHT's temperature readings, the 8\u00d78 thermal grid arrays, and predictions of the algorithm - current people count and detected hot object count. This dataset supports research in thermal sensing, edge-computing applications for smart buildings, privacy-conscious sensing systems, and low-power IoT deployments where traditional camera-based solutions are unsuitable due to privacy constraints or resource limitations.