Applied Mechanics and Materials Vols. 631-632

Abstract: In this paper, we propose a novel, online, long-term tracking algorithm to track objects that possess fairly regular motion. Our tracker contains two terms: stabilizer and attractor. The stabilizer narrows the candidate location of an object at the next frame in a sequence by employing the Kalman filter, which enhances the speed of our tracker and brings stability. The attractor is an inner template of an object consisting of Harris corner pairs. By excluding distractors with a different inner template, our tracker is discriminative and accurate. Experiments on several benchmark sequences show our competitive performance.

Abstract: Vehicle detection based on video plays a more and more important role in the intelligent transportation system (ITSs), since it presents siginificant advantages such as simple structure, less equipment, no radiation and easier to secondary development and so on over other methods. Based on those the paper summarizes the principle of video velocity measurement and object detection, furthermore, focuses on the analysis of the main content and the latest progress of background difference, frame difference and optical flow. Finally, presents the strengths and weaknesses and the technical difficulties faced, and points out directions of future research in this area.

Abstract: Absolute orientation is a classic photogrammetric task for recovering the transformation relationship between two coordinate systems. In low-altitude photogrammetric field, arbitrary coordinate system in the network of aerial triangulation is spatial auxiliary coordinate system, which is quite different from ground photogrammetric coordinate. So the commonly used solution to absolute orientation with simplified parameters is unsuitable for low-altitude photogrammetry. In this paper, a novel solution of absolute orientation is proposed, the initial parameters of absolute orientation can be solved by the formula using centroids of coordinates instead of simplified parameters, then the parameters of absolute orientation are iteratively refined by Levenberg-Marquardt algorithm. Experimental results show that the solution is effective and practical. It may be suitable to implement absolute orientation in low-altitude photogrammetry.

Abstract: A common need in photogrammetry, robotics and computer vision is performing camera pose estimation. A comparative analysis is presented here for three classical and representative algorithms, including direct linear transform (DLT), EPNP and Cayley method, each of which computes the translation and rotation matrix using non-iterative method with six or more point correspondences. The comparison shows qualitative and quantitative experimental results to determine (1) the accuracy and robustness under the influence of different levels of noise, (2) the accuracy, robustness and efficiency for different sizes of point correspondences.

Abstract: The gray image can be viewed as a 3D terrain consisting of intensity waves. The Wave Equalization introduced in the paper is to weaken the influence of uneven illumination on the gray image binarization. The 2D wave is decomposed into several 1D waves corresponding to different directions for 1D equalization. Then PCA algorithm is used to compress all 1D results to the final image with uniform illumination. The extensive experiments in the paper had proved that our method has excellent performance and adaptability for various uneven illumination environments.

Abstract: Image recognition has been a research hotspot in the field of machine learning; this paper puts forward a kind of cascade algorithm based on SVM and AdaBoost. The algorithm to select the sample pretreatment, fixed size of window image segmentation into different areas, then using Haar - like rectangular figure characteristics of integral method for feature extraction, finally using AdaBoost cascade classifier to classify the SVM training. Through the face recognition experiments show AdaBoost cascade of SVM algorithm improve the classification accuracy, error rate get reduced obviously.

Abstract: The three components of the color images enhance the images directly in the RGB color space, will cause distortion of the image. So our paper will convert true color image from RGB space to YIQ space, the holomorphic filtering and histogram equalization are executed on Y component to make the image enhancement, the Y component contains a lot of image information; finally, the image is converted from YIQ color space to RGB color space once again. Experimental results show that the approach which were proposed in the paper, Combined with the method of holomorphic filtering and histogram equalization to overcome the uneven illumination, the image is dark and other shortcomings to achieve satisfactory enhancement.

Abstract: With the development of computer, multimedia and integrated circuit technology, more and more information are digital, so that the storage and processing including images, sounds and words, etc. After processing the information, of course, it eventually to be represented in the form of people can accept. FPGA has the characteristics of the device, occupies an important position in the information processing system, this paper expounds how to complete image acquisition processing system functions in the FPGA.

Abstract: The Many stereo matching problems can be converted to energy minimization problem, by establishment of special network graph to obtain the minimum graph cut, and then obtaining the optimal solution. For graph cuts algorithm, complete network graph include all vertices and disparity edges, the computation of time and space is huge. In this paper, we put forward a method by combing local and global stereo matching, set up a reduced network graph by the possible disparities values for each pixel, and then global optimization, to slove the maximum flow polynomial through CUDA parallel computing, greatly reduced the consumption of time and space.

Abstract: The ultrasonic echo signal of wear debris is influenced by many matters. It causes so much more noise. Therefore, it puts forward an improved online de-noising method for ultrasonic echo signal of wear debris in oil. In the dual-tree complex wavelet transform (DTCWT) field, a method, which combines the new non-linear threshold function with adaptive threshold, utilizes particle swarm optimization (PSO) for optimizing the parameter of non-linear threshold function to get the optimal solution. The result of de-noising method can be evaluated. Experimental results show that the proposed method has obvious effect on signal de-noising.