3D Mapping of Quadcopters Using Stereo Cameras for Indoor Environment

This research involves creating a 3D mapping using technology of stereo camera, utilizing a total of 3 quadcopters equipped with three Raspberry Pi boards. Optical flow sensor or PWM3901 sensor exhibits precise and accurate movement for the 3 quadcopters with an average error of 20 - 30 centimeter. Communication among three quadcopters are Raspberry Pi microcontroller for sending data, receiving data, and streaming video that are occurred over the internet using the HTTP Protocol with Flask API Server. Each quadcopter has a maximum lifting force averaging 5.59 kilogram, 5.49 kilogram, and 5.63 kilogram, respectively. For creating 3D mapping, the three Raspberry Pi devices send images to the computer in form of point cloud creation. The technique that is used to merge the point cloud are disparity calculation, WLS filter, and Key point detection. Each set of point cloud from quadcopter are combined for the point cloud of 3 quadcopters with a slight error from overall dimension, but 3D mapping can indicate the detail in the map correctly.