There are many factors leading to construction safety accident. The rule presented under the influence of these factors should be a statistical random rule. To reveal those random rules and study the ...probability prediction method of construction safety accident, according to stochastic process theory, general stochastic process, Markov process and normal process are respectively used to simulate the risk-accident process in this paper. First, in the general-random-process-based analysis the probability of accidents in a period of time is calculated. Then, the Markov property of the construction safety risk evolution process is illustrated, and the analytical expression of probability density function of first-passage time of Markov-based risk-accident process is derived to calculate the construction safety probability. In the normal-process-based analysis, the construction safety probability formulas in cases of stationary normal risk process and non-stationary normal risk process with zero mean value are derived respectively. Finally, the number of accidents that may occur on construction site in a period is studied macroscopically based on Poisson process, and the probability distribution of time interval between adjacent accidents and the time of the nth accident are calculated respectively. The results provide useful reference for the prediction and management of construction accidents.
Three-dimensional human pose estimation from depth maps is a fast-growing research area in computer vision. The distal joints of the human body are more flexible than the proximal joints, making it ...more difficult to estimate the distal joints. However, most existing methods ignore the difference between the distal joints and proximal joints. Moreover, the distal joint can be constrained by the proximal joint on the same kinematic chain. In our work, we model the human skeleton as the tree structure called the human-tree. Then, motivated by the WPL (weighted path length) in the data structure, we propose a WPL-based loss function to constrain the distal joints with the proximal joints in a global-to-local manner. Extensive experiments on benchmarks demonstrate that our method can effectively improve the performance of the distal joints.
In recent years, with the development of deep learning methods, hand pose estimation based on monocular RGB images has made great progress. However, insufficient labeled training datasets remain an ...important bottleneck for hand pose estimation. Because synthetic datasets can acquire a large number of images with precise annotations, existing methods address this problem by using data from easily accessible synthetic datasets. Domain adaptation is a method for transferring knowledge from a labeled source domain to an unlabeled target domain. However, many domain adaptation methods fail to achieve good results in realistic datasets due to the domain gap. In this paper, we design a self-looping adversarial training strategy to reduce the domain gap between synthetic and realistic domains. Specifically, we use a multi-branch structure. Then, a new adversarial training strategy we designed for the regression task is introduced to reduce the size of the output space. As such, our model can reduce the domain gap and thus improve the prediction performance of the model. The experiments using H3D and STB datasets show that our method significantly outperforms state-of-the-art domain adaptive methods.
High-resolution scanning radar mapping of the surface is an effective tool for addressing concerns in local environmental and social investigation fields. Regrettably, the azimuth resolution of a ...scanning radar is constrained by the antenna beamwidth. Multiple super-resolution approaches have been applied to the scanning radar to enhance the azimuth resolution, but they suffer from limited resolution improvement. In this paper, a methodology to derive surface estimates from the scanning radar at an improved azimuth resolution is proposed. We first consider the truncated spectrum by discarding the unreliable frequencies to suppress the noise amplification. Then, based on the iterative adaptive approach (IAA), a novel inverse filtering method is formulated to obtain lower sidelobes and a higher resolution. Finally, by taking advantage of the Fourier property of the steering matrix and the Toeplitz structure of the covariance matrix, we exploit the Gohberg-Semencul representation and the data-dependent trigonometric polynomials to derive a fast IAA (FIAA)-based inverse filtering to mitigate the computational burden. Simulation results and real data processing demonstrate that the proposed FIAA-based inverse filtering outperforms the existing super-resolution approaches in resolution improvement and results in a higher computational efficiency.
Firstly, this paper analyzes the basic principles and processes of the spatial pattern changes of land use in towns and villages, and the result shows that the land resource demands of urban ...development and population growth lead to the spatial pattern changes. Secondly, in order to grip land use changes better, the paper proposes a method for the simulation of spatial patterns. The simulating method can be divided into two parts: one is a quantitative forecast by using the Markov model, and the other is simulating the spatial pattern changes by using the CA model. The above two models construct the simulative model of the spatial pattern of land use in towns and villages. Finally, selecting Fangshan which is a district of Beijing as the experimental area, both the quantity and spatial pattern changing characteristics are investigated through building a changing dataset of land use by using spatial analysis methods based on the land use data in 2001, 2006 and 2008; CA–Markov is used to simulate the spatial pattern of land use in Fangshan for 2015.
The growing population in China has led to an increasing importance of crop area (CA) protection. A powerful tool for acquiring accurate and up-to-date CA maps is automatic mapping using information ...extracted from high spatial resolution remote sensing (RS) images. RS image information extraction includes feature classification, which is a long-standing research issue in the RS community. Emerging deep learning techniques, such as the deep semantic segmentation network technique, are effective methods to automatically discover relevant contextual features and get better image classification results. In this study, we exploited deep semantic segmentation networks to classify and extract CA from high-resolution RS images. WorldView-2 (WV-2) images with only Red-Green-Blue (RGB) bands were used to confirm the effectiveness of the proposed semantic classification framework for information extraction and the CA mapping task. Specifically, we used the deep learning framework TensorFlow to construct a platform for sampling, training, testing, and classifying to extract and map CA on the basis of DeepLabv3+. By leveraging per-pixel and random sample point accuracy evaluation methods, we conclude that the proposed approach can efficiently obtain acceptable accuracy (Overall Accuracy = 95%, Kappa = 0.90) of CA classification in the study area, and the approach performs better than other deep semantic segmentation networks (U-Net/PspNet/SegNet/DeepLabv2) and traditional machine learning methods, such as Maximum Likelihood (ML), Support Vector Machine (SVM), and RF (Random Forest). Furthermore, the proposed approach is highly scalable for the variety of crop types in a crop area. Overall, the proposed approach can train a precise and effective model that is capable of adequately describing the small, irregular fields of smallholder agriculture and handling the great level of details in RGB high spatial resolution images.
Multi-scale feature fusion techniques and covariance pooling have been shown to have positive implications for completing computer vision tasks, including fine-grained image classification. However, ...existing algorithms that use multi-scale feature fusion techniques for fine-grained classification tend to consider only the first-order information of the features, failing to capture more discriminative features. Likewise, existing fine-grained classification algorithms using covariance pooling tend to focus only on the correlation between feature channels without considering how to better capture the global and local features of the image. Therefore, this paper proposes a multi-scale covariance pooling network (MSCPN) that can capture and better fuse features at different scales to generate more representative features. Experimental results on the CUB200 and MIT indoor67 datasets achieve state-of-the-art performance (CUB200: 94.31% and MIT indoor67: 92.11%).
Accurate urban land-use mapping is a challenging task in the remote-sensing field. With the availability of diverse remote sensors, synthetic use and integration of multisource data provides an ...opportunity for improving urban land-use classification accuracy. Neural networks for Deep Learning have achieved very promising results in computer-vision tasks, such as image classification and object detection. However, the problem of designing an effective deep-learning model for the fusion of multisource remote-sensing data still remains. To tackle this issue, this paper proposes a modified two-branch convolutional neural network for the adaptive fusion of hyperspectral imagery (HSI) and Light Detection and Ranging (LiDAR) data. Specifically, the proposed model consists of a HSI branch and a LiDAR branch, sharing the same network structure to reduce the time cost of network design. A residual block is utilized in each branch to extract hierarchical, parallel, and multiscale features. An adaptive-feature fusion module is proposed to integrate HSI and LiDAR features in a more reasonable and natural way (based on “Squeeze-and-Excitation Networks”). Experiments indicate that the proposed two-branch network shows good performance, with an overall accuracy of almost 92%. Compared with single-source data, the introduction of multisource data improves accuracy by at least 8%. The adaptive fusion model can also increase classification accuracy by more than 3% when compared with the feature-stacking method (simple concatenation). The results demonstrate that the proposed network can effectively extract and fuse features for a better urban land-use mapping accuracy.
As the main problem of the durability deterioration of concrete structures, the corrosion of steel bars is usually made by the method of electrified corrosion with a short cycle and low cost. ...However, there is a big difference between the actual corrosion depth and the theoretical corrosion depth after the reinforcement is electrified. In this paper, through the accelerated corrosion test of steel bars, the change law and influence factors of corrosion efficiency of steel bars in concrete simulated pore solution and NaCl solution are studied. The test results show that the corrosion efficiency of reinforcement in the NaCl solution is higher than that in the concrete simulated pore solution, and the corrosion efficiency in the NaCl solution changes in two stages with the corrosion degree of reinforcement. The corrosion efficiency of concrete in the simulated pore solution decreases with the increase of corrosion degree of reinforcement, which is more significant than that in the NaCl solution. Under the same conditions, the corrosion efficiency is higher in the chloride ion solution with high concentration, and the influence of chloride ion concentration change in the simulated pore solution of concrete on the corrosion efficiency is more significant. The corrosion efficiency of reinforcement under low current density is higher than that under high current density.
Because a scanning radar system works as a noncoherent sensor, it is suitable for any geometry situation, and it has significant and extensive applications, such as surveillance, autonomous landing ...of aircraft, navigation, and guidance. After the pulse compression technique of improving range resolution was presented, angular resolution became crucial for a scanning radar system. In this paper, a scheme of angular superresolution based on maximum a posteriori (MAP) framework is proposed. First, the received signal in azimuth is modeled as a mathematical convolution of the antenna pattern and the targets' scattering. Then, the principle of the angular superresolution algorithm, superresolution performance analysis, and computational implementation are presented. The algorithm can endure more significant disturbance of noise than conventional approaches. Simulation validates that the method can improve the radar angular resolution at least four times, even when the signal-to-noise ratio (SNR) is 10 dB. Furthermore, real data processing has proved the effectiveness of the proposed method.
PDF
