Rancang Bangun Robot Penghindar Halangan Berbasis Kamera Menggunakan Deteksi Kontur

Abstrak

So far, most obstacle avoidance robots are built using reference from some distance sensors placed on the robot body, the weakness of this system is that robots can only avoid obstacles without knowing where the target is located. In this research built a camera-based obstacle avoidance robot with 0.67x wide lens placed 2 meters above the arena, so that the position of robots, obstacles and targets can be known in real-time. The image of the camera will be processed using a laptop that has installed the application CodeBlocks with OpenCV library, the output is a character command movement of a DC motor that is sent on the robot through bluetooth communication. Contour detection is used to distinguish selected objects with surrounding disruptions. From the test results, the HSV range calibration system developed has been able to produce accurate color ranges for color filtering, using auto-click and tuning-manual methods. The actual distance testing calibration based on pixel distance produces a linear graph trend, so the linear equations can be used to convert pixel spacing to actual distance in centimeters. Robot rate test results, built systems can make the robot walk closer to the target and avoid obstacles in front of the robot, although sometimes experienced oscillation but the robot is able to return stable, this is because the built system does not use oscillation damper method such as PID. Keywords - Robot Avoidance, Camera, Contour Detection.