• Have a python3 environment
• Run pip install -r requirements.txt (it may be pip3)

• To start the GUI run ./start.sh

The vein-detection system utilizes a webcam, near-IR LEDs, and a Raspberry Pi. Cameras generally have an IR filter in front of the CCD to block IR light; however, in this project, the near-IR light is desired. A Raspberry Pi NoIR Camera will be used as it does not have an infrared filter. Based on the absorption curve for oxygenated hemoglobin shown in Figure 1, 850nm IR light provides the best balance between absorptivity and skin permeability, so 850nm IR LEDs will be used to illuminate the patient’s veins. Coupling the IR camera with an 850nm IR bandpass filter, the camera will be able to image the veins in the patient’s dorsum region.

The IR camera will send a continuous video stream to a Raspberry Pi for image processing. Using Python’s OpenCV library, the image processing algorithm will determine which vein has the largest diameter and locate the most parallel portion of the vein relative to the gantry. The coordinate of insertion will be displayed on a monitor and the technician will verify the pending vein. Upon verification, the Pi will send the coordinate to an Arduino Mega, which controls the IV gantry.