In this work, we present an augmented reality (AR) approach for position based service using a smartphone in an indoor environment. The AR method, combined with position estimation, provides a user with a smartphone with a service that is specific to a particular position without using a marker or any other hardware device. The position in an indoor environment is estimated using an IMU sensor only in the smartphone. The accuracy of the position and heading direction of the user is improved by integrating the values from the accelerometer and the gyro using Principal Component Analysis(PCA) and Extended Kalman Filter(EKF). Then, a drift noise of the estimated position is reduced by a registration step performed at a specific position. The estimated position is given to the position based service, which is provided to the user on the smartphone screen through AR. The concept of the proposed method is demonstrated with some examples.

This paper proposes a method for augmenting a 3D CAD model onto an image using a mobile device. An image and its depth map of a target are obtained using a Phab 2 mobile device. The image is processed to extract line segments of the target. Next a rectangular planar region of the target is selected, which is then refined using the depth data. The chosen region is then compared with the CAD model, and a planar face in the CAD model that matches the selected region is obtained using various geometric properties. Using the matching planar faces, a pose of the camera is computed, which is then used for augmenting the CAD model onto the image correctly. The test results demonstrate that the method can be used for real fabrication in a complex environment.

This paper introduces a multiple landmine extraction algorithm from the data which are obtained from the ground penetrating radar (GPR). Conventional algorithms focus on extracting a single landmine. The landmine, however, is not buried under the ground alone in the real cases; it is buried with stones or other landmines. Therefore, detection of multiple landmines is an important problem. The proposed algorithm is composed of two steps. The first step finds the number of buried landmines using the energy projection method. The second step extracts individual buried objects from the data that have multiple landmines using the symmetric filtering method. The experiments show that the proposed algorithm demonstrates the potential of detecting multiple landmines.

We present methods for the computation of roots of univariate and bivariate nonlinear polynomial systems as well as the identification of their multiplicity. We first present an algorithm, called the TDB algorithm, which computes the values and the multiplicities of roots of a univariate polynomial. The procedure is based on the concept of the degree of a certain Gauss map, which is deduced from the polynomial itself. In the bivariate case, we use a combination of resultants and our procedure for the univariate case, as the basis for developing an algorithm for locating the roots and computing their multiplicities. Our methods are robust and global in nature. Complexity analysis of the proposed methods is included together with comparison with standard subdivision methods. Examples illustrate our techniques.