China
Africa
Explore chapters and articles related to this topic
Forest Ecosystem Monitoring Using Unmanned Aerial Systems
Published in David R. Green, Billy J. Gregory, Alex R. Karachok, Unmanned Aerial Remote Sensing, 2020
Cristina Gómez, Tristan R.H. Goodbody, Nicholas C. Coops, Flor Álvarez-Taboada, Enoc Sanz-Ablanedo
A small in-house-developed fixed-wing UAV plane of 2 kg made with EPO (expanded polyolefin) was flown over 1,000 ha in the area (Figure 11.10). The aircraft has an endurance of 40 minutes at 12 m × s−1 cruise speed, and the actual flights were <20 minutes at altitudes over 500 m. The positioning system of the plane used a coded-base single-frequency L1, GPS U-Blox receiver. This receiver was just used for aircraft navigation; therefore, its calculated position was not used during SfM (SFM) processing or georeferencing. The flight controller on the aircraft was an APM 2 autopilot, based on the open-source Ardupilot project. Ardupilot controller enables automatic mission flights based on waypoints of known latitude, longitude, and height. The firmware programmed into the flight navigates the UAV through all waypoints using information provided by the inertial system, precision barometer, electronic compass, and external GPS sensor. Two flights were planned with Mission Planner (open-source software) in east-west orientation (Figure 11.10), each one for carrying a single sensor (RGB or NIR). A Canon Powershot S100 12 MPix standard camera (4,000 × 3,000 pixels) modified with a low pass filter was employed to acquire infrared photography (ISO 160 F/2.8, T 1/640). The same camera with standard configuration acquired RGB photography (ISO 80 F/2.2 T 1/1000). The actual flights were carried out on 7 March 2014 at 12.30 (NIR) and on 8 April 2014 at 18.00 (RGB), over totally clear sky. Although the flights covered a larger area, the number of images used for this study was 89 RGB (GSD = 0.17 m) and 56 NIR images (GSD = 0.35 m) and had 80% forward- and 60% side-overlap rate. For georeferencing with Photoscan v.1, five GCPs measured on the ground with a Leica Viva GNSS receptor were employed, yielding residuals of 0.9 m (NIR imagery) and 0.3 m (RGB imagery). Mosaicking was based on the photo most perpendicular to the terrain. All images were acquired with the same exposition and were not radiometrically corrected.
Empirical Evaluation of a Complete Hardware and Software Solution for UAV Swarm Networks
Published in Fei Hu, Xin-Lin Huang, DongXiu Ou, UAV Swarm Networks, 2020
Carlos Felipe Emygdio De Melo, Maik Basso, Marcos Rodrigues Vizzotto, Matheus Schein Cavalheiro Correa, T'U Lio Dapper E Silva, Edison Pignaton De Freitas
The proposed architecture can be used on commercial drones that contain some variation of Ardupilot as firmware. Also, all the proposed software uses standard packages and libraries available in the ROS core, so there is no need for any other installation and complexities in the solution assembly.
An efficient in situ monitoring strategy for an active aquatic surface omni-directional sensing device
Published in Advanced Robotics, 2022
Yasuyuki Fujii, Dinh Tuan Tran, Joo-Ho Lee
Figure 7 illustrates the control architecture of the sensing device. The main computer has a flight controller that can estimate the position of the sensing device and control thrusters through PWM control. The flight controller is used as a controller of autonomous vehicles such as AUV and unmanned aerial vehicle (UAV). The position and pose of the sensing device are estimated using extended Kalman filter (EKF), which integrates the sensor data acquired through the GPS receiver, gyro sensor, accelerometer and magnetometer sensors. The EKF-based localization system is provided by Ardupilot. Ardupilot is an open source project for autonomous vehicle systems [27]. The sensing device determines its subsequent behavior based on the estimated position and pose. A 3-h yaw angle accuracy experiment conducted as a preliminary test showed a maximum error of 1.89 and an average error of 0.68. In preliminary experiments, the robot was held stationary at a point on land and its accuracy was checked against a reliable compass value.