Comparison of 3D Image Acquisition Devices for the Detection, Recognition and Reconstruction of Wooden Logs
Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
One of the important research topics in the field of robotic vision is the detection andrecognition of different real life objects. With the development of new three dimensional (3D)image acquisition systems, robots can detect the world in its full three dimensional reality. Thisincreases the utility of using robots in more real world applications like in factories,automobiles, forestry etc. In forestry log recognition and pose estimation in the forestenvironment are essential for automating the machines used for harvesting and logging thetrees. The detection algorithms will provide the precise location of the log to the automatedcrane mounted on the forest machines for its collection.In this thesis work we have used and compared three different types of 3D image acquisitionsystems for the detection, recognition and reconstruction of a wooden log present on theground using the algorithm proposed by Yeon Choo Park. The algorithm basically considerswooden logs as a cylindrical surface and tries to detect and segment all such surfaces from therest of the scene. Then the pose of each segmented surface is calculated and its missing datapoints are reconstructed by using the information available. Finally the orientation and positionof the log is transferred to the crane for its successful collection.The image acquisition devices are 1) Stereo camera 2) Structure Light camera and 3) Time ofFlight camera system. These 3D imaging systems are tested in the indoor crane laboratoryfacility available at Umeå University. The individual performances of these imaging systems areevaluated depending on the quality of information provided by them and the results of therecognition software. Finally some future research topics are suggested for the improvement ofthe current algorithm.
Place, publisher, year, edition, pages
Electrical Engineering, Electronic Engineering, Information Engineering
IdentifiersURN: urn:nbn:se:umu:diva-58419OAI: oai:DiVA.org:umu-58419DiVA: diva2:548371
Master's Programme in Robotics and Control
Westerberg, SimonFreidovich, Leonid