|
|
Tracking People in Video Captured from a Drone
Lukáč, Jakub ; Orság, Filip (oponent) ; Goldmann, Tomáš (vedoucí práce)
This thesis deals with the problem of determining the location of a person through their distance from camera approximation. The location is derived from video which is captured using a drone. The goal here is to propose and test existing solutions, and state-of-the-art algorithms for each encountered subproblem of the tracking. This means overcoming challenges such as object detection, re-identification of persons in time, estimating object distance from the camera and processing data from various sensors. Then, I am using these methods to design the final solution which can operate in nearly real-time. Implementation is based on the use of Intel NCS accelerator unit with the cooperation of small computer Raspberry Pi. Therefore, the setup may be easily mounted directly to a drone. The resulting application can generate tracking metadata for detected individuals in the recording. Afterwards, the positions are visualised as paths for better end-user presentation.
|
|
|
Tracking People in Video Captured from a Drone
Lukáč, Jakub ; Orság, Filip (oponent) ; Goldmann, Tomáš (vedoucí práce)
This thesis deals with the problem of determining the location of a person and its approximation. The location is derived from video which is captured using a drone. The goal here is to propose and test existing solutions and state-of-the-art algorithms for each encountered subproblem. This means overcoming challenges such as object detection, re-identification of persons in time, estimating object distance from camera and processing data from various sensors. Then, I am using the methods to design the final solution which can operate in nearly real-time. Implementation is based on the use of Intel NCS accelerator unit with the cooperation of small computer Raspberry Pi. Therefore, the setup may be easily mounted directly to a drone. The resulting application can generate tracking metadata for detected individuals in the recording. Afterwards, the positions are visualised as paths for better end-user presentation.
|
|
|
Localization in Wireless Energy-Constrained Networks
Morávek, Patrik ; Vozňák, Miroslav (oponent) ; Křepelka, Václav (oponent) ; Komosný, Dan (vedoucí práce)
The doctoral thesis is devoted to localization in wireless networks, and particularly, to distance estimation. Localization is an important process in many wireless networks with both static and mobile nodes since it provides a position knowledge which can be further exploited during the application lifetime. The thesis presents a novel method for distance estimation based on received signal strength measurement. The method respects both the application accuracy requirements and dynamic ambient radio conditions while performing with minimal energy costs. Before the design of the novel method an experimental analysis of signal propagation for localization and energy consumption was performed. Based on the results of the analysis the novel Adaptive Energy-aware Distance Estimation method was proposed and subsequently evaluated in both simulator and experimental testbed under real conditions.
|
|
|
Reference Nodes Selection for Anchor-Free Localization in Wireless Sensor Networks
Šimek, Milan ; Makáň, Florian (oponent) ; Diviš, Zdeněk (oponent) ; Komosný, Dan (vedoucí práce)
The doctoral thesis is focused on a design of a novel anchor free localization algorithm for wireless sensor networks. As introduction, the incremental and concurrent anchor free localization algorithms are presented and their performance is compared. It was found that contemporary anchor free localization algorithms working in the concurrent manner achieve a low localization error, but dissipate signicant energy reserves. A new Boundary Recognition Aided Localization algorithm presented in this thesis is based on an idea to recognize the nodes placed on the boundary of network and thus reduce the number of transmission realized during the reference nodes selection phase of the algorithm. For the position estimation, the algorithm employs the multilateration technique that work eectively with the low number of the reference nodes. Proposed algorithms are tested through the simulations and validated by the real experiment with the wireless sensor network. The novel Boundary Recognition Aided Localization algorithm is compared with the known algorithms in terms of localization error and the communication cost. The results show that the novel algorithm presents powerful solution for the anchor free localization.
|
|
|
Computer Vision for Autonomous Vehicles
Lečbych, Michal ; Škrabánek, Pavel (oponent) ; Shehadeh, Mhd Ali (vedoucí práce)
Perceptive systems in autonomous cars are a heavily researched topic these days and an essential part of making fully autonomous vehicles possible. First, we make a short summary of the development of such a system, then we explain different approaches to make these systems possible, and we focus on object detection, as this will be the main part of our own created perceptive system. A new model for object detection is implemented, and some additional parts like distance estimation and lane detection are added.
|
|
|
Tracking People in Video Captured from a Drone
Lukáč, Jakub ; Orság, Filip (oponent) ; Goldmann, Tomáš (vedoucí práce)
This thesis deals with the problem of determining the location of a person through their distance from camera approximation. The location is derived from video which is captured using a drone. The goal here is to propose and test existing solutions, and state-of-the-art algorithms for each encountered subproblem of the tracking. This means overcoming challenges such as object detection, re-identification of persons in time, estimating object distance from the camera and processing data from various sensors. Then, I am using these methods to design the final solution which can operate in nearly real-time. Implementation is based on the use of Intel NCS accelerator unit with the cooperation of small computer Raspberry Pi. Therefore, the setup may be easily mounted directly to a drone. The resulting application can generate tracking metadata for detected individuals in the recording. Afterwards, the positions are visualised as paths for better end-user presentation.
|
|
|
Tracking People in Video Captured from a Drone
Lukáč, Jakub ; Orság, Filip (oponent) ; Goldmann, Tomáš (vedoucí práce)
This thesis deals with the problem of determining the location of a person and its approximation. The location is derived from video which is captured using a drone. The goal here is to propose and test existing solutions and state-of-the-art algorithms for each encountered subproblem. This means overcoming challenges such as object detection, re-identification of persons in time, estimating object distance from camera and processing data from various sensors. Then, I am using the methods to design the final solution which can operate in nearly real-time. Implementation is based on the use of Intel NCS accelerator unit with the cooperation of small computer Raspberry Pi. Therefore, the setup may be easily mounted directly to a drone. The resulting application can generate tracking metadata for detected individuals in the recording. Afterwards, the positions are visualised as paths for better end-user presentation.
|
|
|
Perception of space in virtual reality environments
Fajnerová, Iveta ; Vlček, Kamil (vedoucí práce) ; Vavrečka, Michal (oponent)
Cílem této diplomové práce bylo objasnit problematiku prostorového vidění s ohledem na navigaci ve virtuálním prostředí a neuronální základy odhadu vzdáleností. Pro tento účel byla vytvořena virtuální verze Úlohy hledání skrytého cíle, která je analogií Morrisova vodního bludiště pro lidi. Práce prezentuje výsledky experimentu s odstraňováním orientačních značek v kruhové aréně. Cílem experimentu bylo objasnit, jestli pro naši arénu platí předpoklad teorie kognitivního mapování o rovnocennosti orientačních značek při hledání skryté pozice cíle. Výsledky experimentu naznačují, že přesnost odhadu pozice cíle je dána nejen počtem viditelných značek, ale i jejich individuální hierarchií. Ta může být odvozena na základě jejich vzdálenosti od pozice cíle, ale v některých případech je ovlivněna výraznou identitou značky. Na tyto výsledky navazuje experiment využívající principy funkční magnetické rezonance s cílem objasnit neuronální základy odhadu vzdáleností ve virtuální aréně v egocentrickém a allocentrickém referenčním rámci. Výsledky potvrzují nálezy citovaných studií o účasti oblastí okcipitálního a parietálních laloku při odhadování vzdáleností v prostoru. Porovnání obou typů referenčních rámců ukázalo, že zatímco pro egocentrický odhad je specifická i aktivita premotorických korových oblastí, v...
|
|
|
Localization in Wireless Energy-Constrained Networks
Morávek, Patrik ; Vozňák, Miroslav (oponent) ; Křepelka, Václav (oponent) ; Komosný, Dan (vedoucí práce)
The doctoral thesis is devoted to localization in wireless networks, and particularly, to distance estimation. Localization is an important process in many wireless networks with both static and mobile nodes since it provides a position knowledge which can be further exploited during the application lifetime. The thesis presents a novel method for distance estimation based on received signal strength measurement. The method respects both the application accuracy requirements and dynamic ambient radio conditions while performing with minimal energy costs. Before the design of the novel method an experimental analysis of signal propagation for localization and energy consumption was performed. Based on the results of the analysis the novel Adaptive Energy-aware Distance Estimation method was proposed and subsequently evaluated in both simulator and experimental testbed under real conditions.
|
|
|
Reference Nodes Selection for Anchor-Free Localization in Wireless Sensor Networks
Šimek, Milan ; Makáň, Florian (oponent) ; Diviš, Zdeněk (oponent) ; Komosný, Dan (vedoucí práce)
The doctoral thesis is focused on a design of a novel anchor free localization algorithm for wireless sensor networks. As introduction, the incremental and concurrent anchor free localization algorithms are presented and their performance is compared. It was found that contemporary anchor free localization algorithms working in the concurrent manner achieve a low localization error, but dissipate signicant energy reserves. A new Boundary Recognition Aided Localization algorithm presented in this thesis is based on an idea to recognize the nodes placed on the boundary of network and thus reduce the number of transmission realized during the reference nodes selection phase of the algorithm. For the position estimation, the algorithm employs the multilateration technique that work eectively with the low number of the reference nodes. Proposed algorithms are tested through the simulations and validated by the real experiment with the wireless sensor network. The novel Boundary Recognition Aided Localization algorithm is compared with the known algorithms in terms of localization error and the communication cost. The results show that the novel algorithm presents powerful solution for the anchor free localization.
|
host ::
RSS.