Lijuan Qin

A new closed-form method for pose estimation from three Z-like lines

In this paper, we present a new method for estimating object pose from 2-D to 3-D straight line correspondences. This method can get the closed-form solution of object pose with respect to the camera from three Z-like lines. The advantage of this method is that it can compute object pose more quickly than iterative method because it avoids the iterative process of iterative methods. Whatpsilas more, this method is simple to solve. Especially, we provide a detailed account of the computational aspects of this method. At last, experimental results which demonstrate our method works well and fast are presented.

Pose Determination of 3D Object Based on Four Straight Lines

Pose determination of 3D object with respect to a camera is an important problem in computer vision. Straight line is one feature which exists widely in natural environments. In this paper, a new pose determination method from four straight line correspondences is presented. This algorithm is simple to solve and has good real-time characteristic. Furthermore, unique solution for object pose can be obtained. At last, simulation results show this method is reasonable and has good real-time characteristic.

Algorithm for attitude determination from three door-like lines

In this paper, we propose a novel closed-form method to determine object attitude with respect to the camera for three door-like lines. The advantage of this closed-form method is that it is simple. Furthermore, it can compute object pose more quickly than iterative methods because it avoids the iterative process of iterative methods. This method can locate object uniquely at last, thus it solves the multi-solution problem and facilitate the application of pose estimation algorithm in practice. Especially, we provide a detailed account of the computational aspects of this method. At last, experimental results which demonstrate our method works well and fast are presented.

Study on single camera vision measurement using coplanar features

3D measurement in computer vision is to determine attitude of an object with respect to camera. Environments exist a lot of straight line features. New 3D measurement method of single camera from four coplanar line correspondences is presented in this paper. This algorithm is linear to solve. Simulation results show this method is feasible.

New Algorithm for Determining Object Attitude Based on Trapezoid Feature

Trapezoid is one feature which is widely presented in natural environments. In this paper, we propose a new pose estimation method based on trapezoid. The algorithm is simple to solve. It has good real-time characteristic. Furthermore, there exists unique solution for object pose from rectangle to locate. Thus, it has good practical value in application. At last, simulation results show this method is reasonable and has good real-time characteristic.

Research on Optimum Position for Straight Lines Model

Quantization errors are the primary source that affects the accuracy of pose estimation. For the model at different placement positions, quantization errors have different effects on the results of pose estimation. The analysis of optimum displacement for the model can increase the accuracy of pose estimation. In this paper, mathematical model of the propagation of quantization errors from a two-dimensional image plane to the 3D model is set up. Whats more, optimization function for the analysis of optimum placement position of model is set up. For given model, we obtain optimum placement position of model with respect to camera. At last, the simulation results show that it has better pose estimation accuracy at optimum place than at other places.

Research on monotony character of a new location method

We present a new pose estimation method from line correspondences and introduce a new iterative method for it. The iterative method works fast. In some cases, the process of search satisfies the monotony character which is the premise for convergence of iterative method. Especially, we introduce one monotony region and provide a detailed account of the proof of monotony at this region. The value of the proof is to provide theoretical basis for the iterative method and direct the application of the location method.

A new approach for location from rectangle correspondence

Location of an object with respect to a camera is an important problem in computer vision. Rectangle is one feature which exists widely in natural environments. In this paper, a new location method from rectangle correspondence is presented. This algorithm is simple to solve. It has good real-time characteristic. Furthermore, we obtain unique solution for object pose from rectangle to locate. Thus, it has good practical value in application. At last, simulation results show this method is reasonable and has good real-time characteristic.