|
|
Ball Trajectory Estimation from Images
Heřman, Zdeněk ; Materna, Zdeněk (referee) ; Španěl, Michal (advisor)
Main goal of this thesis is to design a method for ball trajectory estimation from images, and to evaluate the proposed method using an experimental setup. Both the method and the experimental setup were designed to be suitable for a golf simulator. The proposed solution uses a stereo camera for flying ball tracking. Accurracy of the ball detection in images is further improved by using a pulsed light source. A solution reaches an accuracy that can be compared with cheaper simulators that already exist. Nevertheless prize is lower.
|
|
|
Ensuring Precision of Stereo Image Processing
Kuník, Oliver ; Španěl, Michal (referee) ; Zemčík, Pavel (advisor)
This thesis deals with the possibilities of refining the results of stereo image processing and examines the methods of stereo image stabilization. Proposes a method of stabilizing the stereo recording with respect to the designated reference stereo image to keep the images in the calibration position of the camera. An affine transformation is used to eliminate the movement of stereo images. The positions of the corner points found with subpixel precision are used for calculating the transformation. This thesis includes a comparison of the stabilization efficiency of a stereo record intended for vehicle speed measurements.
|
|
|
Fusion of drone sensors for motion planning
Semerád, Jakub ; Novotný, Josef (referee) ; Janoušek, Jiří (advisor)
This bachelor thesis deals with drone flight planning in an unknown environment. In order for a drone to be able to perceive its environment, it must be equipped with the necessary sensory equipment. This thesis discusses possible sensor equipment suitable for flight coordination and collision avoidance. Also, software for sensor data processing and drone motion planning are described and used here. LiDAR and depth camera were used for the experiment in this thesis. In the experiment the data from both sensors are fused together through ROS software and displayed in RViz simulation environment. This data are suitable for subsequent use in drone motion planning and collision avoidance.
|
|
|
Obstacle Avoidance System
Dražil, Jan ; Raichl, Petr (referee) ; Novotný, Josef (advisor)
This thesis deals with methods for the problem of navigation and movement of autonomous vehicles between obstacles. The thesis first describes in general terms algorithms for motion planning and obstacle avoidance. Then, the thesis discusses the PX4 firmware for UAV control. The thesis proposes a custom obstacle avoidance algorithm designed for UAVs. The sensor used for obstacle detection is a stereo camera. This algorithm was implemented in Python using the ROS framework and tested in the Gazebo simulation environment. The results are discussed.
|
|
|
Obstacle Avoidance System
Dražil, Jan ; Raichl, Petr (referee) ; Novotný, Josef (advisor)
This thesis deals with methods for the problem of navigation and movement of autonomous vehicles between obstacles. The thesis first describes in general terms algorithms for motion planning and obstacle avoidance. Then, the thesis discusses the PX4 firmware for UAV control. The thesis proposes a custom obstacle avoidance algorithm designed for UAVs. The sensor used for obstacle detection is a stereo camera. This algorithm was implemented in Python using the ROS framework and tested in the Gazebo simulation environment. The results are discussed.
|
|
|
Fusion of drone sensors for motion planning
Semerád, Jakub ; Novotný, Josef (referee) ; Janoušek, Jiří (advisor)
This bachelor thesis deals with drone flight planning in an unknown environment. In order for a drone to be able to perceive its environment, it must be equipped with the necessary sensory equipment. This thesis discusses possible sensor equipment suitable for flight coordination and collision avoidance. Also, software for sensor data processing and drone motion planning are described and used here. LiDAR and depth camera were used for the experiment in this thesis. In the experiment the data from both sensors are fused together through ROS software and displayed in RViz simulation environment. This data are suitable for subsequent use in drone motion planning and collision avoidance.
|
|
|
Ensuring Precision of Stereo Image Processing
Kuník, Oliver ; Španěl, Michal (referee) ; Zemčík, Pavel (advisor)
This thesis deals with the possibilities of refining the results of stereo image processing and examines the methods of stereo image stabilization. Proposes a method of stabilizing the stereo recording with respect to the designated reference stereo image to keep the images in the calibration position of the camera. An affine transformation is used to eliminate the movement of stereo images. The positions of the corner points found with subpixel precision are used for calculating the transformation. This thesis includes a comparison of the stabilization efficiency of a stereo record intended for vehicle speed measurements.
|
|
|
Ball Trajectory Estimation from Images
Heřman, Zdeněk ; Materna, Zdeněk (referee) ; Španěl, Michal (advisor)
Main goal of this thesis is to design a method for ball trajectory estimation from images, and to evaluate the proposed method using an experimental setup. Both the method and the experimental setup were designed to be suitable for a golf simulator. The proposed solution uses a stereo camera for flying ball tracking. Accurracy of the ball detection in images is further improved by using a pulsed light source. A solution reaches an accuracy that can be compared with cheaper simulators that already exist. Nevertheless prize is lower.
|
guest ::
