Search

The locator is the key component to control the track of contact wire in overhead catenary system (OCS) for high-speed railway. Once the locator slope is out of bound, it would pose a huge hazard to the safety of the high-speed trains and threat to the human life and property damage. In this work, a novel end-to-end locator slope calculation framework is presented for locator slope real-time inspection in high-speed railway system. The pipeline is composed of two stages: locator contour detection and slope calculation. In order to precisely detect the locator contours in OCS images captured from high-speed extreme environments, a novel detection mechanism including rough detection and fine fitting is proposed. For the fast slope calculation through only one camera, monocular vision model is modified by two novel assumptions to calculate the locator space coordinates. Rigorous experiments are performed across a number of standard locator slope calculation benchmarks, showing a large improvement in the precision and speed over all previous methods. Finally, the effectiveness of proposed framework is demonstrated through a real-world application of the high-speed rail OCS inspection system.

High-speed railway intelligent monitoring and management system is composed of schedule integration, geographic information, location services, and data mining technology for integration of time and space data. Assistant localization is a significant submodule of the intelligent monitoring system. In practical application, the general access is to capture the image sequences of the components by using a high-definition camera, digital image processing technique and target detection, tracking and even behavior analysis method. In this paper, we present an end-to-end character recognition method based on a deep CNN network called YOLO-toc for high-speed railway pillar plate number. Different from other deep CNNs, YOLO-toc is an end-to-end multi-target detection framework, furthermore, it exhibits a state-of-art performance on real-time detection with a nearly 50fps achieved on GPU (GTX960). Finally, we realize a real-time but high-accuracy pillar plate number recognition system and integrate natural scene OCR into a dedicated classification YOLO-toc model.

In recent years, due to the growing popularity of the human network behavior, we meet the new generation of big data era. This makes the emergence of characteristics that large data storage capacity and fast business growth. For large storage capacity, instability, low retrieval time caused by large data information, this paper demonstrate a distributed storage and server clustering approach to solve the problem of tradition storage capacity, high-availability of retrieval and servers. For the arrival of the era of big data, how to improve the retrieval time, the storage capacity and stability of system, these provide for the future construction of a guiding role.

Recent years, many vision-based vehicle detection methods have been proposed. But just a few of them paid attention to vehicle type classification in a real-time video, which is an important part in intelligent traffic system. The Haar features and Histograms of Oriented Gradients (HOG) features have been proposed as useful features for vehicle detection respectively. In this paper, we propose a method based on combined the Haar features and HOG features to detect vehicles in videos and classify them into two types. The experiment results show that this method can classify and detect the vehicles in multi-orientations with good classification results, and it verified the effectiveness of the propose method with real vehicle samples.

An approach of 3D face recognition by using of facial surface classification image and PCA is presented. In the step of pre-processing, the scattered 3D points of a facial surface are normalized by surface fitting algorithm using multilevel B-splines approximation. Then, partial-ICP method is utilized to adjust 3D face model to be in the right front pose for a better recognition performance. By using the normalized facial depth image been acquired through the two previous steps, and by calculating the Gaussian and mean curvatures at each point, the surface types are classified and the classification result is used to mark different kinds of area on the facial depth image by 8 gray-levels. This achieved gray image is named as Surface Classification Image (SCI) and the SCI now represents the 3D features of the face and then it is input to the process of PCA to obtain the SCI eigenfaces to recognize the face. In the experiments conducted on 3D Facial database ZJU-3DFED of Zhejiang University, we obtained the rank-1 identification score of 94.5%, which outperformed the result of using PCA method directly on the face depth image (instead of SCI) by 16.5%.

Minimum ignition energy is one of the important characteristics for electrostatic sensitivity of explosives and determination of the spark discharge energy is necessary for decision making of explosion prevention. In this paper, the real spark discharge energy was determined by the integration of the measured voltage and discharge current in a needle-to-plate configuration. The influence of the charging voltage and polarity, the charging capacitance and the electrode gap on the spark discharge energy was investigated. It is shown that the energy deposited into the discharge system is always smaller than that stored in the charging capacitor, especially at lower voltages that the discharge shows one current pulse and the energy efficiency is less than 30%. The discharge becomes complete at higher voltages when the discharge shows double pulses and the energy efficiency is close to 100%. For given charging capacitance and voltage, the real discharge energy is determined by the discharge mode rather than the electrode gap or the voltage polarity.