Time filter

Source Type

Zhang L.,Langfang Teachers College | Wang D.,Langfang Vocational and Technical College
Metallurgical and Mining Industry | Year: 2015

In order to improve the precision of automatic calibration of computer vision, RAC calibration algorithm is used to realize automatic calibration. First, some briefly introduction of calibration method should be known, then the details of Mathematical model of the RAC calibration algorithm and the calibration process will be introduced. Finally, we use the simulation example to validate the performance of this method. The results of the experiments show that this method has realized the automatic calibration of computer vision. The fit of the measured value and the actual value is high, so this method has high application value. Source

Zhang L.,Langfang Teachers College | Li Q.,Langfang Vocational and Technical College
International Journal of Earth Sciences and Engineering | Year: 2015

According to the problems such as low accuracy and long time when standard C4.5 algorithm in data mining to the table tennis match, an evaluation and decision model of table tennis match based on improved C4.5 algorithm is proposed. First introduce the optimization theory, improve the formula of amount of information by using the properties of convex function, to improve the accuracy of the classification in decision tree classifier. Then draw lessons from the thought of information gain, and optimize rules of candidate attributes to C4.5 algorithm by using the parameters support and credibility of association rule mining, and then optimize the operation time by the simplified calculation method of the division of information. Finally, apply the improved C4.5 algorithm to evaluation and decision of table tennis match, and perform the video data mining of table tennis game. Seen from the simulation results, compared to standard C4.5 algorithm, the proposed evaluation and decision model for the of table tennis match based on improved C4.5 algorithm, has better performance in data mining. © 2014 CAFET-INNOVA TECHNICAL SOCIETY. All rights reserved. Source

Zhang L.,Langfang Teachers College | Wang D.,Langfang Vocational and Technical College
Metallurgical and Mining Industry | Year: 2015

As the standard particle filter algorithm has the problems that the tracking accuracy is not so good and the robustness is poor in the application of motion object detection, this paper proposes a basketball player tracking model based on hierarchical multiple feature fusion particle filter. First of all, use the color histogram distribution to model the target, and use a weighted function to take samples of target to establish the histogram, and then optimize the progressive nature of the histogram parameters to ensure the fidelity of complex function and data, then according to the characteristics of the shape, use the SIFT to execute shade optimization for the particle filter, finally construct the basketball player target tracking model incorporating the above methods. Simulation experiment shows that the improved algorithm shows excellent tracking performance whether there is a shelter or not. © 2015. Metallurgical and Mining Industry. Source

Zhang L.,Langfang Teachers College | Ren F.,Langfang Vocational and Technical College
Journal of Computational Information Systems | Year: 2015

Aiming at the poor accuracy and the short stability problems existed when the existing video retrieval algorithm is retrieving sports video, this paper proposed a sports video retrieval algorithm based on key frame extraction and image matching. Firstly the key frame extraction strategy based on comentropy and edge matching rate is adopted, the sport video is transformed into picture comentropy, then the improved Canny algorithm is adopted to extract its edge point. Furthermore the edge matching rate is used to extract the key frame of sport video, after that the LBP operator is introduced, and it is carried on the histogram transformation optimization and adaptive weighted optimization. At last the improved LBP operator is adopted to do picture matching in the extracted key frame. The simulation experiment shows that, the sports video retrieval algorithm based on key frame extraction and image matching which is proposed by this paper has good effects in the application of sport video retrieval, comparing with the SIFT algorithm, it has better stability and accuracy. Copyright © 2015 Binary Information Press. Source

Discover hidden collaborations