A Vision System to Detect 3D Wound Surfaces for Plasma Medical Applications

Dielectric Barrier Discharge (DBD) plasma jets are increasingly being used in medical applications. Due to their non-equilibrium and non-thermal properties, DBD plasma is an effective method for enhancing wound healing and sterilization. When the wound surface is exposed to the plasma jet, the jet scans the area, initiating crucial chemical reactions between reactive nitrogen and oxygen species with the liquid biofilm. This generates secondary reactive species that play a significant role in the wound healing process.

One challenge in utilizing DBD plasma is achieving precise and controllable scanning of the plasma jet, which requires process automation. To address this, a triple-joint robotic arm has been designed at SJSU to carry the plasma jet and follow a predetermined tool path to scan the flat wound surface in a 2D configuration. However, since wounds on the human body are often irregular and not flat, 3D plasma scanning is necessary.

This project describes a vision system designed to detect the wound and calculate the position of key points on the wound surface in 3D. Several steps were involved in developing the vision system. First, computer vision methods were implemented using OpenCV in Python to detect the wound in real-time. Validation checks ensure that the detected contours correspond to the wound’s actual shape. Next, the coordinates of key points on the wound were calculated in 3D space, relative to the origin of the robotic arm. These coordinates are then sent to the processor controlling the robotic arm, enabling it to follow a path between the key points.

The developed vision system demonstrated high accuracy in detecting the wound and calculating the positions of key points. This poster will present the wound detection process and the calculations used to determine the positions of the key points. The mechanical design of the system and its recorded performance will also be discussed.