VISION-BASED TRACKING TURRET USING ESP32-CAM AND YOLO WITH A DECOUPLED HOST-PC ARCHITECTURE

This paper presented a Vision-Based Tracking Turret in which a host PC runs YOLO detection and an ESP32 board drives the pan-tilt servos and a laser indicator, with an ESP32-CAM as the wireless camera. Splitting detection from actuation avoids the processor contention found in single-processor trackers. The prototype tracked a person indoors with an average confidence of 0.92 at close range and a delay of about 80 to 150 ms, at a hardware cost of about INR 1,940. Future work will add a thermal sensor for dark scenes, Kalman filter prediction to bridge missed detections, GPU or edge-AI inference to remove the PC, multi-target tracking and a web dashboard for remote monitoring.

1)    P. Viola and M. J. Jones, "Robust Real-Time Face Detection," Int. J. Computer Vision, vol. 57, no. 2, pp. 137-154, May 2004. 2)    N. Dalal and B. Triggs, "Histograms of Oriented Gradients for Human Detection," in Proc. IEEE CVPR, 2005, vol. 1, pp. 886-893. 3)    W. Liu et al., "SSD: Single Shot MultiBox Detector," in Eur. Conf. Computer Vision (ECCV), Springer, 2016, pp. 21-37. 4)    J. Redmon, S. Divvala, R. Girshick and A. Farhadi, "You Only Look Once: Unified, Real-Time Object Detection," in Proc. IEEE CVPR, 2016, pp. 779-788. 5)    Raza et al., Raspberry Pi 4 face-tracking turret using an OpenCV Haar cascade, 2019. [Full citation to be added.] 6)    Jain and Kumar, Raspberry Pi 4 tracking turret using MobileNet-SSD, 2021. [Full citation to be added.] 7)    Ultralytics, "YOLOv8 Documentation." [Online]. Available: https://docs.ultralytics.com 8)    PySerial Development Team, "PySerial Documentation." [Online]. Available: https://pyserial.readthedocs.io 9)    Arduino, "ESP32Servo Library Documentation." [Online]. Available: https://madhephaestus.github.io/ESP32Servo/annotated.html