We present a solution to animating interactions between characters and thin, non-rigid bodies using a passive optical motion capture system. Prior work in human body motion capture can accurately recreate human motion but this work is not adequate because it does not allow for interactions with a non-rigid body. Prior work in face and cloth motion capture handles non-rigid planes but rope is better handled with a curved spline rather than a curved plane. The segmented motion is in the form of un-indexed motion capture data. After segmenting the motion of the thin, non-rigid body and the human character the separated motion capture data can be recreated individually. The recreated motion streams are then recombined using 3D modeling and animation software. The presented solution also improves techniques for recreating thin, non-rigid body motion from un-indexed motion capture data. Using the linear combination of two predicted marker positions our solution can accurately track motion capture markers through each frame of the motion capture data. This also allows our solution to invent marker positions when gaps are present in the motion capture data. Our improvements allow users to reconstruct the motion of both a human character and a thin, non-rigid body simultaneously from motion capture data gathered in a mixed motion capture session.

This research presents a procedure for interactive segmentation and automatic tracking of moving objects in a video sequence. The user outlines the region of interest (ROI) in the initial frame; the procedure builds a refined mask of the dominant object within the ROI. The refined mask is used to model a spline template of the object to be tracked. The tracking algorithm then employs a motion model to track the template through a sequence of frames and gathers the 3D affine motion parameters of the object from each frame. The extracted template is compared with a previously stored library of 3D shapes to determine the closest 3D object. If the extracted template is completely new, it is used to model a new 3D object which is added to the library. To recreate the motion, the motion parameters are applied to the 3D object in a virtual environment. The procedure described here can be applied to industrial problems such as traffic management and material flow congestion analysis.

Many bridges are damaged by collisions from over-height vehicles resulting in significant impact to the transportation network. To reduce the likelihood of impact from over-height vehicles, steel pedestals have been used as a cost-effective, efficient means to increase bridge clearance heights. However, these steel pedestals installed on more than 50 bridges in Georgia have been designed with no consideration of seismic loads and may behave in a similar fashion to high-type steel bearings. Past earthquakes have revealed the susceptibility of high-type bearings to damage, resulting in the collapse of several bridges. Although Georgia is located in a low-to-moderate region of seismicity, earthquake design loads for steel pedestals should not be ignored. In this study, the potential vulnerabilities of steel pedestals having limited strength and deformation capacity and lack of adequate connection details for anchor bolts is assessed experimentally and analytically. Full-scale reversed cyclic quasi-static experimental tests are conducted on a 40' bridge specimen rehabilitated with 19" and 33" steel pedestals to determine the modes of deformation and mechanisms that can lead to modes of failure. The inelastic force-deformation hysteretic behavior of the steel pedestals obtained from experimental test results is used to calibrate an analytical bridge model developed in OpenSees. The analytical bridge model is idealized based on a multi-span continuous bridge in Georgia that has been rehabilitated with steel pedestals. The analytical bridge model is subjected to a suite of ground motions to evaluate the performance of the steel pedestals and the overall bridge system. Recommendations are made to the Georgia Department of Transportation (GDOT) for the design and construction of steel pedestals. The results of this research are useful for Georgia and other states in low-to-moderate seismic zones considering the use of steel pedestals to elevate bridges and therefore reduce the likelihood of over-height vehicle collisions.

This research presents a procedure for interactive segmentation and automatic tracking of moving objects in a video sequence. The user outlines the region of interest (ROI) in the initial frame; the procedure builds a refined mask of the dominant object within the ROI. The refined mask is used to model a spline template of the object to be tracked. The tracking algorithm then employs a motion model to track the template through a sequence of frames and gathers the 3D affine motion parameters of the object from each frame. The extracted template is compared with a previously stored library of 3D shapes to determine the closest 3D object. If the extracted template is completely new, it is used to model a new 3D object which is added to the library. To recreate the motion, the motion parameters are applied to the 3D object in a virtual environment. The procedure described here can be applied to industrial problems such as traffic management and material flow congestion analysis.

No description available.

The nonlinear interactions between aerodynamic forces and wing structures are numerically investigated as integrated dynamic systems, including structural models, aerodynamics, and control systems, in the time domain. An elastic beam model coupled with rigid-body rotation is developed for the wing structure, and the natural frequencies and mode shapes are found by the finite-element method. A general unsteady vortex-lattice method is used to provide aerodynamic forces. This method is verified by comparing the numerical solutions with the experimental results for several cases; and thereafter applied to several applications such as the inboard-wing/twin-fuselage configuration, and formation flights. The original thought that the twin fuselage could achieve two-dimensional flow on the wing by eliminating free wing tips appears to be incorrect. The numerical results show that there can be a lift increase when two or more wings fly together, compared to when they fly alone. Flutter analysis is carried out for a High-Altitude-Long-Endurance aircraft wing cantilevered from the wall of the wind tunnel, a full-span wing mounted on a free-to-roll sting at its mid-span without and with a center mass (fuselage). Numerical solutions show that the rigidity added by the wall results in a higher flutter speed for the wall-mounted semi-model than that for the full-span model. In addition, a predictive control technique based on neural networks is investigated to suppress flutter oscillations. The controller uses a neural network model to predict future plant responses to potential control signals. A search algorithm is used to select the best control input that optimizes future plant performance. The control force is assumed to be given by an actuator that can apply a distributed torque along the spanwise direction of the wing. The solutions with the wing-tip twist or the wing-tip deflection as the plant output show that the flutter oscillations are successfully suppressed with the neural network predictive control scheme. / Ph. D.

視覺追蹤(visual tracking)一直是傳統電腦視覺研究中相當重要的議題，許多電腦視覺的應用都需要結合視覺追蹤的幫助才能實現。近年來擴增實境(augmented reality)能快速成功的發展，均有賴於視覺追蹤技術上之精進。擴增實境採用視覺追蹤的技術，可將虛擬的物件呈現在被追蹤的物體(真實場景)上，進而達成所需之應用。 由於在視覺追蹤上，被追蹤之物體易受外在環境因素影響，例如位移、旋轉、縮放、光照改變等，影響追蹤結果之精確度。本研究中，我們設計了一套全新的圖形標記方法作為視覺追蹤之參考點，能降低位移、旋轉與光照改變所造成追蹤結果的誤差，也能在複雜的背景中定位出標記圖形的正確位置，提高視覺追蹤的精確度。同時我們使用立體視覺追蹤物體，將過去只使用單一攝影機於二維影像資訊的追蹤問題，提升至使用三維空間的幾何資訊來做追蹤。然後透過剛體(rigid)特性找出旋轉量、位移量相同的物件，並且結合一致性隨機取樣(random sample consensus)之技巧以估測最佳的剛體物件運動模型，達到強健性追蹤的目的。 另外，我們可由使用者提供之影片資訊中擷取特定資料，透過建模技術將所產生之虛擬物件呈現於使用者介面(或被追蹤之物體)上，並藉由這些虛擬物件，提供真實世界外之資訊，達成導覽指引(或擴增實境)的效果。 實驗結果顯示，我們的方法具有辨識時間快、抗光照變化強、定位準確度高的特性，適合於擴增實境應用，同時我們設計的標記圖形尺寸小，方便適用於導覽指引等應用。 / Visual tracking is one of the most important research topics in traditional computer vision. Many computer vision applications can not be realized without the integration of visual tracking techniques. The fast growing of augmented reality in recent years relied on the improvement of visual tracking technologies. External environment such as object displacement, rotation, and scaling as well as illumination conditions will always influence the accuracy of visual tracking. In this thesis, we designed a set of markers that can reduce the errors induced by the illumination condition changes as well as that by the object displacement, rotation, and scaling. It can also correctly position the markers in complicated background to increase the tracking accuracy. Instead of using single camera tracking in 2D spaces, we used stereo vision techniques to track the objects in 3D spaces. We also used the properties of rigid objects and search for the objects with the same amount of rotation and displacement. Together with the techniques of random sample consensus, we can estimate the best rigid object motion model and achieve tracking robustness. Moreover, from the user supplied video, we can capture particular information and then generate the virtual objects that can be displaced on the user’s device (or on the tracked objects). Using these techniques we can either achieve navigation or guidance in real world or achieve augmented reality as we expected. The experimental results show that our mechanism has the characteristics of fast recognition, accurate positioning, and resisting to illumination changes that are suitable for augmented reality. Also, the size of the markers we designed is very small and good for augmented reality application.

No description available.