Aiming at the fact that the existing video person re-identification methods cannot effectively extract the spatiotemporal information between consecutive frames of the video, a person re-identification network based on non-local attention and multi-feature fusion was proposed to extract global and local representation features and time series information. Firstly, the non-local attention module was embedded to extract global features. Then, the multi-feature fusion was realized by extracting the low-level and middle-level features as well as the local features, so as to obtain the salient features of the person. Finally, the similarity measurement and sorting were performed to the person features in order to calculate the accuracy of video person re-identification. The proposed model has significantly improved performance compared to the existing Multi-scale 3D Convolution (M3D) and Learned Clip Similarity Aggregation (LCSA) models with the mean Average Precision (mAP) reached 81.4% and 93.4% respectively and the Rank-1 reached 88.7% and 95.3% respectively on the large datasets MARS and DukeMTMC-VideoReID. At the same time, the proposed model has the Rank-1 reached 94.8% on the small dataset PRID2011.

To solve the problems of low precision and poor generalization performance of traditional image segmentation algorithms on the plant images under bias light in plant factory, a method based on neural network and deep learning for accurately segmenting the plant images under artificial bias light in plant factory was proposed. By using this method, the segmentation accuracy on the original test set of bias light plant images is 91.89% and is far superior to that by other segmentation algorithms such as Fully Convolutional Network (FCN), clustering, threshold and region growth. In addition, this method has better segmentation effect and generalization performance than the above methods on plant images under different color lights. The experimental results show that the proposed method can significantly improve the accuracy of plant image segmentation under bias light, and can be applied to practical plant factory projects.

The current vibration detection system has the problems of large delay, poor controllability of sensor network and low detection precision. In order to solve the problems, a new vibration detection system based on Process field bus (Profibus) technology was proposed. Firstly, the complex calculations such as Kalman filtering and Fast Fourier Transform (FFT) were implemented at each node of detection device. The network load was reduced by about 95% compared with the traditional scheme of transmitting large amount of original vibration data, the network transmission time and the computing time of workstation were shortened, the real-time performance and computing power of system were improved. Secondly, the Profibus protocol was used to realize the management and data transmission of the vibration detection device network, which ensured the stability and controllability of the sensor network. In addition, the high-precision vibration sensor was used on the vibration detection node device, and the vibration data was filtered on the detection node. The detection precision was up to 0.0039 mg. What's more, the function customization and secondary development of detection device were convenient by the independent design and development of the Profibus protocol slave station. The RT-Thread embedded system kernel was used to achieve resource allocation and task scheduling on the detection node device, and the real-time performance and reliability were improved. The experimental results show that, the proposed system can process the original vibration data of the vibration field quickly, and transmit the processed data to the workstation computer with high real-time performance. Meanwhile, the Profibus network composed of vibration detection devices can display the status information of node in real-time and remind users promptly when there is a network fault. The network has good controllability.

The existing evaluation of dynamic traffic capacity does not consider the controller's cognitive load. In order to improve the accuracy of air traffic flow management, a new sector dynamic traffic capacity evaluation model based on controller's cognitive load and improved ant colony algorithm was constructed. Firstly, a dynamic flight constrained region model which could describe the dynamic influence factors of the sector was constructed, and the dynamic control guided path planning was realized by improving ant colony algorithm, which could meet the calculation speed requirement of air traffic flow management. Then, the concept of control and guidance load intensity was proposed and used for the construction of sector dynamic traffic capacity evaluation model, which expanded the concept of controller's total cognitive load. Finally, taking the Sanya regulatory sector as an example, the dynamic traffic capacity of the sector at 9 moments in the next 2 hours was evaluated by taking 15 min as an interval. The results of example verification show that the results of dynamic traffic capacity obtained by the proposed model are different from the actual operating results by an airplane, and the effects are ideal.

Focusing on the issue that the Correlation Filter (CF) has poor performance in tracking fast motion object, a Long-term Kernelized Correlation Filter (LKCF) tracking algorithm based on optical flow combining with Kernel Correlation Filter (KCF) was proposed. Firstly, while tracking with the tracker, a value of Peak-to-Sidelobe Ratio (PSR) was calculated. Secondly, the position was achieved in the last frame, optical flow was used to calculate coarse position when the value of PSR less than a threshold in the current frame, which means tracking failure. Finally, accurate position was calculated using the tracker again according to the coarse position. The results of experiment compared with four kinds of tracking algorithms such as Compressive Tracking (CT), Tracking-Learning-Detection (TLD), KCF and Spatio-Temporal Context (STC) show that the proposed algorithm is optimal in distance accuracy and success rate which are 6.2 percentage points and 5.1 percentage points higher than those of KCF. In other words, the proposed algorithm is robust to the tracking of fast motion object.

In order to improve the efficiency of power transmission line inspection by Unmanned Aerial Vehicle (UAV), a new method was proposed for detecting broken transmission lines and defects of foreign body based on the perception of line structure. The transmission line image acquired by UAV was easily influenced by the background texture and light, the gradient operators of horizontal and vertical direction which can be used to detect the line width were used to extract line objects in the inspection image. The study on calculation of gestalt perception of similarity, continuity and colinearity connected the intermittent wires into continuous wires. Then the parallel wire groups were further determined through the calculation of parallel relationship between wires. In order to reduce the detection error rate, spacers and stockbridge dampers of wires were recognized based on a local contour feature. Finally, the width change and gray similarity of segmented conductor wire were calculated to detect the broken part of wire and foreign object defect. The experimental results show that the proposed method can detect broken wire strand and foreign object defect efficiently under complicated backgrounds from the transmission line of UAV images.

Researches on the audio classification have involved various types of vector features. However, multi-semantics of audio information not only have their own properties, but also have some correlations among them. Whereas, to a certain extent, the simple vector representation cannot represent the multi-semantics and ignore their relations. Tensor Uniform Content Locator (TUCL) was brought forward to express the semantic information of audio, and a three-order Tensor Semantic Space (TSS) was constructed according to the semantic tensor. Tensor Semantic Dispersion (TSD) can aggregate some audio resources with the same semantics, and at the same time, the automatic audio classification can be accomplished by calculating their TSD. And Radical Basis Function Tensor Neural Network (RBFTNN) was constructed and used to train intelligent learning model. For the problem of multi-semantic audio classification, the experimental results show that our method can significantly improve the classification precision in comparison with the typical method of Gaussian Mixture Model (GMM), and the classification precision of RBFTNN model is obviously better than that of Support Vector Machine (SVM).

Because of the advantage of EJB in mantenance and extensibility,J2EE architecture was widely used in developing large E-commerce application.But the elusiveness of EJB means more spending and hard to use.A simplified.NET-based container managed persistence was introduced.It provided an excellent persistence container.

Besides the needed data,there are lots of navigation information and advertisements in the Web pages.A DOM tree comparison algorithm was proposed.It compared several pages within a class,and recognized the main contents in pages.Experiment results show that it is feasible and effective.