Mobile Phone Holder Gimbal

A face recognition tracking system made entirely from recycled old parts (one old computer power supply providing 12V; a discarded GT2560 REV_A control board from a 3D printer; two stepper motors salvaged from disassembled machines; one personal mobile phone used as a camera).

The provided models are SOLIDWORKS 2021 parts and assemblies. You can adjust the dimensions yourself based on the stepper motors, 3D printer accuracy, and material sinkage rate.

Face recognition software: Python (opencv, YOLOV8N-FACE); control board firmware: Marlin-2.1.x (Configuration.h needs modification); driver: FTDI FT232RQ driver-CDM2123620_Setup; burning tool: AVRDUDESS-2.20-setup; firmware compilation using Python commands: pio run -t clean -e mega2560
pio run -e mega2560.

To connect your phone to the computer as a webcam, you need Iriun Webcam (PC and Android versions), download it yourself (do not use Droidcam, as USB connection is difficult and Wi-Fi connection cannot read the screen).

The recognition model is too large to upload: yolov8n-face-lindevs.pt

Disadvantages: The phone is too heavy, the stepper motor's torque is insufficient, and it sags when tilted 45 degrees. The code is poorly written; it works but is extremely laggy. The basic functions are normal. Those interested can modify the code themselves, and please share it with me if you do, thank you.