We address the problem of multi-view multi-target tracking using multiple stationary cameras in the application of hockey tracking and test the approach with data from two cameras. The system is based on the previous work by Okuma et al. [50]. We replace AdaBoost detection with blob detection in both image coordinate systems after background subtraction. The sets of blob-detection results are then mapped to the rink coordinate system using a homography transformation. These observations are further merged into the final detection result which will be incorporated into the particle filter. In addition, we extend the particle filter to use multiple observation models, each corresponding to a view. An observation likelihood and a reference color model are also maintained for each player in each view and are updated only when the player is not occluded in that view. As a result of the expanded coverage range and multiple perspectives in the multi-view tracking, even when the target is occluded in one view, it still can be tracked as long as it is visible from another view. The multi-view tracking data are further processed by trajectory smoothing using the Maximum a posteriori smoother. Finally, automatic camera selection is performed using the Hidden Markov Model to create personalized video programs.

This thesis describes a method for the recognition of objects in an unconstrained environment with a widely ranging illumination, imaged from unknown view points and complicated background. The general problem is simplified by placing specially designed patterns on the object that allows us to solve the pose determination problem easily. There are several key components involved in the proposed recognition approach. They include pattern detection, pose estimation, model acquisition and matching, searching and indexing the model database. Other crucial issues pertaining to the individual components of the recognition system such as the choice of the pattern, the reliability and accuracy of the pattern detector, pose estimator and matching and the speed of the overall system are addressed. After establishing the methodological framework, experiments are carried out on a wide range of both synthetic and real data to illustrate the validity and usefulness of the proposed methods. The principal contribution of this research is a methodology for Tagged Object Recognition (TOR) in unconstrained conditions. A robust pattern (calibration chart) detector is developed for off-the-shelf use. To empirically assess the effectiveness of the pattern detector and the pose estimator under various scenarios, simulated data generated using a graphics rendering process is used. This simulated data provides ground truth which is difficult to obtain in projected images. Using the ground truth, the detection error, which is usually ignored, can be analysed. For model matching, the Chamfer matching algorithm is modified to get a more reliable matching score. The technique facilitates reliable Tagged Object Recognition (TOR). Finally, the results of extensive quantitative and qualitative tests are presented that show the plausibility of practical use of Tagged Object Recognition (TOR). The features characterising the enabling technology developed are the ability to a) recognise an object which is tagged with the calibration chart, b) establish camera position with respect to a landmark and c) test any camera calibration and 3D pose estimation routines, thus facilitating future research and applications in mobile robots navigations, 3D reconstruction and stereo vision.

No description available.

abstract: Camera calibration has applications in the fields of robotic motion, geographic mapping, semiconductor defect characterization, and many more. This thesis considers camera calibration for the purpose of high accuracy three-dimensional reconstruction when characterizing ball grid arrays within the semiconductor industry. Bouguet's calibration method is used following a set of criteria with the purpose of studying the method's performance according to newly proposed standards. The performance of the camera calibration method is currently measured using standards such as pixel error and computational time. This thesis proposes the use of standard deviation of the intrinsic parameter estimation within a Monte Carlo simulation as a new standard of performance measure. It specifically shows that the standard deviation decreases based on the increased number of images input into the calibration routine. It is also shown that the default thresholds of the non-linear maximum likelihood estimation problem of the calibration method require change in order to improve computational time performance; however, the accuracy lost is negligable even for high accuracy requirements such as ball grid array characterization. / Dissertation/Thesis / M.S. Electrical Engineering 2012

The purpose of this research was to develop, implement, and evaluate a citizen science program to survey and monitor for jaguars (Panthera onca) and ocelots (Leopardus pardalis) within key mountain ranges in southwest Arizona. I had three main objectives: 1.) develop and implement a citizen science program to train participants with little or no knowledge of wildlife conservation, 2.) identify program effectiveness in comparison with experts in terms of the quality of data gathered and the financial and administrative costs/benefits, 3.) understand what creates an effective citizen science program and how it can be replicated in the future. This research examined the efficacy of using citizen science as a tool to connect the public and scientific community through education and a research-based project. A network of citizen scientists were provided extensive education and training by field experts to participate in jaguar and ocelot monitoring in southern Arizona. The scale and scope of this project is unprecedented in the citizen science literature; participants drove and hiked off-trail monitoring wildlife cameras monthly for a minimum of one year in extremely rugged backcountry conditions in the mountains near the U.S./Mexico border. Citizen science participants were challenged by terrain, weather, and border issues during their time on this project. Despite these challenges, our group of nine citizen scientists successfully downloaded and sorted 28,637 photos from 22 cameras at 14 monitoring sites over 12 months. They logged a total of 327 hours of fieldwork including driving, hiking, and performing camera maintenance. After finishing a site visit, citizen scientists also sorted photos identifying wildlife, logging approximately 109 hours of data organization. In addition, 100% of our citizen scientists adhered to our protocols that we tested and implemented in the field over a one-year period. Citizen scientists continue to monitor remote wildlife cameras in the rugged mountainous regions of this southern Arizona study area. Citizen science is often criticized because of skepticism from the scientific community regarding data integrity, quality control, and potentially biased data. This research aimed to exam the data integrity of citizen scientists by comparing it to the data analyzed by experts. When comparing the ability of citizen scientists and field team experts in sorting and correctly identifying wildlife, I found strong positive correlations between the levels of data quality. These high positive correlations indicate that, with training, citizen scientists are capable of accurately identifying wildlife from camera data nearly as well as the field team experts, and can be an excellent surrogate for experts. Connecting volunteers to the natural landscape through hands-on science research has the potential to create many positive experiences. Citizen science can increase participant’s knowledge of science, build trust and foster understanding, and can create a more informed public. Despite these benefits, data integrity is the most important aspect of research and data collected by non-scientists remains heavily scrutinized. Collaboration amongst professionals, educators, program designers, and data managers is necessary to ensure that project goals are achieved while maintaining scientific integrity. Continuing to examine citizen science programs is important to advance the field of citizen science and foster meaningful relationships between the public and scientists. This study provided a unique opportunity to use non-scientists to augment data collection and assessment in the scientific workplace to advance jaguar and ocelot conservation. We strongly believe that citizen scientists remain an underutilized resource for helping scientists collect and analyze data in a climate of reduced funding and increasing need for long-term monitoring.

We address the problem of multi-view multi-target tracking using multiple stationary cameras in the application of hockey tracking and test the approach with data from two cameras. The system is based on the previous work by Okuma et al. [50]. We replace AdaBoost detection with blob detection in both image coordinate systems after background subtraction. The sets of blob-detection results are then mapped to the rink coordinate system using a homography transformation. These observations are further merged into the final detection result which will be incorporated into the particle filter. In addition, we extend the particle filter to use multiple observation models, each corresponding to a view. An observation likelihood and a reference color model are also maintained for each player in each view and are updated only when the player is not occluded in that view. As a result of the expanded coverage range and multiple perspectives in the multi-view tracking, even when the target is occluded in one view, it still can be tracked as long as it is visible from another view. The multi-view tracking data are further processed by trajectory smoothing using the Maximum a posteriori smoother. Finally, automatic camera selection is performed using the Hidden Markov Model to create personalized video programs. / Science, Faculty of / Computer Science, Department of / Graduate

No description available.

The topic of this master's thesis is devoted to design and realization of autonomous camera system of the selected object with possibility of remote access. The content of the theoretical part is description of components of the camera system and close analysis of the used software, including the selection of suitable components. The practical part is approaching to the initial installation of the system and verification of the functionality of individual components based on simple commands. The next step is developing an mobile application comunicattting with the camera system based on low data flow.

The aim of the work was to create an automatic positioning system, with optics for radiation concentration and a body for its collection, by monitoring the sun across the sky using a camera. At the beginning are introduced the concentrator systems and the movement of the sun from the perspective of the observer. Follows description of the various potential components which requires a functional system. The last part deal with the implementation of selected components for operation as a whole system, verification of functionality by accurate tracking of the sun across the sky and measuring the performance of the concentrator system with automatic positioning.

Following text focuses on use of wireless technologies in OrthoPilot navigation system developed by B.Braun company. Description of OrthoPilot software is followed by overview of available wireless technologies highlighting their both advantages and disadvantages. Practical part consists of two main parts, mostly dealing with electronic circuits. First part describes development process of camera-wireless printed circuit board which substitutes currently used RS-422 cable connection between PC and stereo camera. Part of this chapter covers programming in C++ in order to make interface compatible with the rest of current OrthoPilot software. Second bigger part deals with remote controller development using prototyping board mikroMedia for XMEGA. Besides electrical circuits design, chapter describes also software part - microcontroller programming in C language. Thesis is concluded by discussing system limitations and ideas for future development.