Computer Vision in Robotic Zen Garden Systems: An Embedded Approach for Contour Detection and Pattern Generation

This paper presents an integrated system that combines computer vision and robotics to automate the creation of traditional Zen garden patterns. The system uses a single Raspberry Pi to detect contours in real time and generate aesthetically pleasing patterns around them, mimicking natural elements like water ripples and waves. By leveraging OpenCV for contour detection and custom pattern generation algorithms, the system produces dynamic, high-quality designs based on the placement and different sizes and shapes of objects within the garden. The experimental results of this system include high detection accuracy, with precision and recall rates exceeding 90% for object classification, and the ability to generate smooth, fluid patterns that closely align with traditional Zen garden aesthetics. Furthermore, the system demonstrates robust real-time performance and ensures seamless motor control. This paper showcases a cost-effective and efficient solution for blending robotics with creative automation in Zen garden design.