Search

Cameras have been widely used in traffic operations. While many technologically smart camera solutions in the market can be integrated into Intelligent Transport Systems (ITS) for automated detection, monitoring and data generation, many Network Operations (a.k.a Traffic Control) Centres still use legacy camera systems as manual surveillance devices. In this paper, we demonstrate effective use of these older assets by applying computer vision techniques to extract traffic data from videos captured by legacy cameras. In our proposed vision-based pipeline, we adopt recent state-of-the-art object detectors and transfer-learning to detect vehicles, pedestrians, and cyclists from monocular videos. By weakly calibrating the camera, we demonstrate a novel application of the image-to-world homography which gives our monocular vision system the efficacy of counting vehicles by lane and estimating vehicle length and speed in real-world units. Our pipeline also includes a module which combines a convolutional neural network (CNN) classifier with projective geometry information to classify vehicles. We have tested it on videos captured at several sites with different traffic flow conditions and compared the results with the data collected by piezoelectric sensors. Our experimental results show that the proposed pipeline can process 60 frames per second for pre-recorded videos and yield high-quality metadata for further traffic analysis.

Porous cellulose films have been reported as sustainable and highly-efficient non-power radiative cooling (PRC) materials but are still challenged by their insufficient atmospheric window (AM) emissivity and complex preparation technology. Herein, bio-mimetic skin structure inspired 3D porous cellulose acetate hybrid hollow AlPO4 (3D PCA/h-AlPO4) nano-laminated PRC film is creatively fabricated by a proposed solvent polarity matching strategy. Owing to the ultra-rapid extraction of high-quality solvent NMP by water relying on their stronger matching polarity, the PRC film self-forms ultra-rapidly within 5 min and is composed of skin-like wrinkle surface, porous matrix, and dense nanolayer. The wrinkle surface and dense h-AlPO4 nanolayer has been demonstrated to have higher atmospheric window emissivity (∼94.7% and 96.6%) than the smooth surface (88.7%). The self-formed uniform cavities of about 2 μm can contribute to a high solar reflectivity (∼97%). As proof of concept, the bio-structural film is used for all-day cooling under different atmospheric radiations, and about 13.5 °C/5 °C (daytime high solar irradiance of 988 W m−2) and ∼1.5 °C/0.5 °C (nighttime) temperature reduction is realized at low/high (12 °C/35 °C) environment temperatures, respectively. We believe the bio-structure design and ultra-rapid film-formation strategy would inspire the development of polymer-based PRC materials for the highly-efficient energy saving applications.

Diffusion Tensor Imaging (DTI) is a new Magnetic Resonance Imaging (MRI) method that is newly used in clinical practice with no traumatic advantages, which can quantitatively analyze the microstructure of the brain, such as quantitative analysis of the dispersion characteristics of lesions and normal tissues, thus providing more information for the diagnosis and differential diagnosis of diseases. At the same time, DTI has a unique advantage in the imaging of white matter fibers in the brain, can visually show the relationship between intracranial lesions, especially tumors and autologous fibers, more conducive to guiding the operation, in order to maximize the removal of tumors and protect functional areas. In order to make the dispersal volume better applied to clinical, this topic mainly hopes to provide a new basis for the benign-malignant classification of brain tumors and the diagnosis of different brain tumors, FA (fractional anisotropy) value. It can provide a stronger basis for the diagnosis of brain tumors, which is rarely done in previous research institutes. Previous studies have rarely dealt with the signal strength of FA diagrams, and extensive research in this area is needed. It is hoped that the signal strength of FA graphs will help diagnose different tumors and grade the malignant degree of brain tumors. Through this experiment, it can be proved that this purpose can be achieved by magnetic resonance dispersion volume imaging. In this study, a comparative analysis method was used to determine the ADC and FA values on the magnetic resonance dispersion weighted image and dispersal volume image of glioma and benign malignant meningioma, thus revealing the important role of this method in the diagnosis and differential diagnosis of tumors in the brain and its application value.

Deformation-induced martensitic transformation (DIMT) has been used for designing high-performance alloys to prevent structural failure under static loads. Its effectiveness against fatigue, however, is unclear. This limits the application of DIMT for parts that are exposed to variable loads, although such scenarios are the rule and not the exception for structural failure. Here we reveal the dual role of DIMT in fatigue crack growth through in situ observations. Two antagonistic fatigue mechanisms mediated by DIMT are identified, namely, transformation-mediated crack arresting, which prevents crack growth, and transformation-mediated crack coalescence, which promotes crack growth. Both mechanisms are due to the hardness and brittleness of martensite as a transformation product, rather than to the actual transformation process itself. In fatigue crack growth, the prevalence of one mechanism over the other critically depends on the crack size and the mechanical stability of the parent austenite phase. Elucidating the two mechanisms and their interplay allows for the microstructure design and safe use of metastable alloys that experience fatigue loads. The findings also generally reveal how metastable alloy microstructures must be designed for materials to be fatigue-resistant.

Passive daytime radiative cooling materials significantly alleviate the current energy shortage and global warming without consuming any energy. However, conventional inorganic nanoparticles can hardly achieve excellent cooling effect during daytime due to their low reflectivity. Broadband super white nanoparticle-based radiators demonstrate excellent solar reflectivity under direct sunlight and low absorption. Accordingly, we introduced TiO2 nanoparticles with the advantages of high refractive index, high infrared emissivity and low solar absorption into the porous cellulose-based film via a facile phase separation strategy. The cavities were distributed around 5 μm in size due to the limited evaporation of solvents, while the introduction of TiO2 nanoparticles effectively enhanced the solar reflectance to 0.97. Moreover, the scattering model demonstrated that the high solar reflectivity was mainly produced by the randomly agglomerate effect of nanoparticles and these nano-clusters also contribute to a high mid-infrared emissivity. As a result, the photonic film presented about 10 °C cooling capacity even under the high-density solar radiation (~897 W/m2) according to the out-door measurement. The random nano-clusters filling strategy avoids the complicate process of specific size particles and has a significant potential for propel the polymer-based cooling technology to satisfy the need of large-scale thermal management.

Vehicle re-identification (Re-ID) plays a key role in many smart traffic management systems. Re-identifying a vehicle can be very challenging because the differences in visual appearances between pairs of vehicles are sometimes extremely subtle if they have the same colour and the same model. Given an image of a vehicle, most existing techniques adopt a global feature representation where details may be ignored. In this paper, we propose an Self-Attention Stair Feature Fusion model to learn the discriminative features for vehicle Re-ID. The model is designed to extract multi-level features in order to capture as much small details as possible. We also propose a Temporal Bayesian Re-Ranking method to exploit the spatial-temporal information in the vehicles’ travel patterns. Our algorithm has been tested against state-of-the-art techniques on popular benchmarks. The results show that our algorithm outperforms other state-of-the-art techniques by a large margin.

Visual tracking (VT) is the process of locating a moving object of interest in a video. It is a fundamental problem in computer vision, with various applications in human-computer interaction, security and surveillance, robot perception, traffic control, etc. In this paper, we address this problem for the first time in the quantum setting, and present a quantum algorithm for VT based on the framework proposed by Henriques et al. [IEEE Trans. Pattern Anal. Mach. Intell., 7, 583 (2015)]. Our algorithm comprises two phases: training and detection. In the training phase, in order to discriminate the object and background, the algorithm trains a ridge regression classifier in the quantum state form where the optimal fitting parameters of ridge regression are encoded in the amplitudes. In the detection phase, the classifier is then employed to generate a quantum state whose amplitudes encode the responses of all the candidate image patches. The algorithm is shown to be polylogarithmic in scaling, when the image data matrices have low condition numbers, and therefore may achieve exponential speedup over the best classical counterpart. However, only quadratic speedup can be achieved when the algorithm is applied to implement the ultimate task of Henriques's framework, i.e., detecting the object position. We also discuss two other important applications related to VT: (1) object disappearance detection and (2) motion behavior matching, where much more significant speedup over the classical methods can be achieved. This work demonstrates the power of quantum computing in solving computer vision problems., 10 pages, 9 figures

Visual tracking has been an active research area in computer vision for decades. However, the performance of existing techniques is still challenged by various factors, such as occlusion and change in appearance of the target. In this paper, we propose a novel framework based on correlation filtering and probabilistic finite state machines (FSMs) to handle occlusion. In our tracking framework, the target is partitioned into several parts whose occlusion states are automatically detected. A set of states for the target is defined in terms of the combination of the parts’ occlusion states. The probabilistic FSMs are then used to model the target’s state transitions so as to reduce the effect of noise in the output response maps of correlation filters. Our target model’s update strategy is adaptable online depending on the estimated state of the target. Extensive experiments have been performed on several public benchmarks and the proposed algorithm achieves competitive results against state-of-the-art techniques.

Visual tracking is a very challenging problem in computer vision as the performance of a tracking algorithm may be degraded due to many challenging issues in the scenes, such as illumination change, deformation, and background clutter. So far no algorithms can handle all these challenging issues. Recently, it has been shown that correlation filters can be implemented efficiently and, with suitable features and kernel functions incorporated, can give very promising tracking results. In this paper, we propose to learn discriminative correlation filters that incorporate information from the variances of the target's appearance features. We have evaluated our filters against several recent tracking methods on the OTB benchmark dataset. Our results show that the additional feature variances help to improve the robustness of the correlation filters in complex scenes.