Intelligent detection of road cracks is crucial for road maintenance and safety. Due to the interference of illumination and different background factors, the road crack extraction results of existing deep learning methods are incomplete, and the extraction accuracy is low. We designed a new network model, called AR-UNet, which introduces a convolutional block attention module (CBAM) in the encoder and decoder of U-Net to effectively extract global and local detail information. The input and output CBAM features of the model are connected to increase the transmission path of features.The BasicBlock is adopted to replace the convolutional layer of the original network to avoid network degradation caused by gradient disappearance and network layer growth. We tested our method on DeepCrack, Crack Forest Dataset, and our own labeled road image dataset (RID). The experimental results show that our method focuses more on crack feature information and extracts cracks with higher integrity. The comparison with existing deep learning methods also demonstrates the effectiveness of our proposed method. The code is available at: https://github.com/18435398440/ARUnet.

When the speed of permanent magnet motor is relatively high, the fundamental frequency of stator voltage and current will rise. However, due to the limitation of switching frequency of power electronic devices, the switching times of power devices will be greatly reduced in a fundamental cycle of stator voltage. For the traditional control method with constant switching frequency mode, the control algorithm is executed only once in a carrier wave cycle, which will result in a low control frequency operating condition for the PMSM control system. At this time, the dynamic and steady-state control performance of stator current will be decreased. In response to above problem, this paper proposed a deadbeat predictive current control method based on oversampling scheme, which the core algorithm of current loop operated four times and updated the PWM duty twice in one switching cycle. The number of control times of the proposed method is greater than that of the conventional method when the conventional deadbeat method and the proposed method adopt the same switching cycle. Therefore, the control method does not increase the switching frequency of power devices, but improves the dynamic and stable performance of the system, especially under the highspeed condition.This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.

The study of high-precision building change detection is essential for the sustainable development of land resources. However, remote sensing imaging illumination variation and alignment errors have a large impact on the accuracy of building change detection. A novel lightweight Siamese neural network building change detection model is proposed for the error detection problem caused by non-real changes in high-resolution remote sensing images. The lightweight feature extraction module in the model acquires local contextual information at different scales, allowing it to fully learn local and global features. The hybrid attention module consisting of the channel and spatial attention can make full use of the rich spatiotemporal semantic information around the building to achieve accurate extraction of changing buildings. For the problems of large span of changing building scales, which easily lead to rough extraction of building edge details and missed detection of small-scale buildings, the multi-scale concept is introduced to divide the extracted feature maps into multiple sub-regions and introduce the hybrid attention module separately, and finally, the output features of different scales are weighted and fused to enhance the edge detail extraction capability. The model was experimented on the WHU-CD and LEVIR-CD public data sets and achieved F1 scores of 87.8% and 88.1%, respectively, which have higher change detection accuracy than the six comparison models, and only cost 9.15 G MACs and 3.20 M parameters. The results show that our model can achieve higher accuracy while significantly reducing the number of model parameters.

With the increasing economic growth in developing nations, soil heavy metal pollution has become a growing concern. Monitoring the heavy metal concentration in soil through remote sensing is crucial for safeguarding the ecological environment. However, the current indoor spectral measurement method has limitations, such as the discrete soil sampling space and weak spectral characteristics of soil heavy metals, leading to a poor robustness of remote sensing inversion models. This study presents a novel approach to address these challenges by incorporating a spatial feature of pollution sources and sinks to evaluate the spatial factors affecting pollutant diffusion and concentration. An integrated learning model, combining spatial and spectral information, is developed to estimate heavy metal content in soil using Sentinel-2A satellite data. A total of 235 soil samples were collected in Jiyuan, China, and the effective spectral transformation characteristics of Sentinel-2A data were screened. The impact of spectral characteristics, topographic characteristics, and spatial characteristics on retrieving soil heavy metal lead (Pb) and cadmium (Cd) content were analyzed. The optimal inversion method was determined through various integrated learning models, and the spatial distribution of heavy metals Pb and Cd was mapped. The results indicate that the accuracy of the inversion model was significantly improved by incorporating terrain features and spatial features of pollution sources. The Blending integrated learning method showed a 65.9% and 73.2% reduction in the RMSE of Pb and Cd, respectively, compared to other regression models. With R2 values of 0.9486 and 0.9489 for Pb and Cd, respectively, and a MAPE less than 0.2, the Blending model demonstrated high prediction accuracy.