Lane detection is one of the most critical tasks in autonomous driving. In the past few years, due to the development of deep neural networks, lane detection approaches using onboard sensors like cameras and LiDAR have been proven to be effective ways to improve performance. In contrast to the camera-based scheme, LiDAR-based lane detection exhibits remarkable robustness to varying lighting conditions. This paper proposes a novel cyclic-shifting group convolution (CSGConv) module. Compared with classical group convolution, the proposed CSGConv module can efficiently promote information exchange among different group feature channels and reduce the high computational burden in current LiDAR-based lane detection networks. Then, a cross stage partial CSGConv (CSPCSG) block is designed to enlarge the receptive field and improve detection performance, especially when the lanes are curved or occluded. Subsequently, a LiDAR-based lane detection network, CSGNet, is put forward by integrating the CSPCSG block. Experiments and ablation tests on the dataset illustrate that our strategy achieves the highest 83.9% overall F1-score. Compared with the current state-of-the-art LiDAR-based lane detection method LLDN-RW, our results exhibit 2.5 times faster and 39% reduction of FLOPs, which indicates less resources are required in the online process.

The Bell-Bloom magnetometer is an instrument for measuring weak magnetic fields that are widely used in geophysical exploration, earthquake monitoring, natural disaster monitoring, and other fields. In geophysical exploration, the magnetometer can detect changes in underground materials; and in magnetic field monitoring, it can accurately detect magnetic field anomalies caused by earthquakes. Relaxation is a crucial characteristic of the Bell-Bloom magnetometer, and unclear relaxation information can hinder the design and improvement of the Bell-Bloom magnetometer, thereby affecting its potential applications. This study proposes an intelligent algorithm called relaxation time of free precession signal (RT-FPS) for measuring the relaxation time of the Bell-Bloom magnetometer parameters to address these issues. Experimental results demonstrate the algorithm’s superior convergence efficiency, fitting accuracy, and noise robustness. Moreover, the algorithm exhibits rapid convergence and high computational accuracy with a minimum sum of squared residuals$1.8386 \times 10^{-11} \text { s}^{2}$. The algorithm is robust against different types of noise and is minimally affected by data quality, with minimum errors of 0.1 ms and$0.71~\mu \text{s}$for$T_{1}$and$T_{2}$, respectively. This study can enhance the performance of the Bell-Bloom magnetometer and its potential applications in magnetic anomaly detection and geomagnetic field monitoring. Our shareable code and data sources are available athttps://github.com/baicaidezhenshi/RT-FPS-DATA.git.

The free precession style Bell–Bloom atomic magnetometer is widely used in geophysical exploration, earthquake monitoring, and natural disaster monitoring to obtain magnetic field information by measuring the Larmor frequency of the free precession signal. However, the free precession signal complexity makes it challenging to accurately acquire the Larmor frequency using conventional frequency measurement methods, limiting its applicability. This study proposes a magnetic information of free precession signal data (MI-FPD) algorithm for measuring the Larmor frequency of the free precession signal in the Bell–Bloom atomic magnetometer. The MI-FPD algorithm accurately determines the Larmor precession frequency based on free precession signal data, providing precise magnetic field information. The algorithm outperforms established algorithms regarding convergence efficiency, accuracy, and noise resistance, with an optimal determination coefficient$R^{2}$of 0.99984, an optimal range of less than 1.04 nT, and an optimal root mean square error (RMSE) of 189 pT. These results demonstrate the effectiveness of the proposed algorithm in accurately obtaining magnetic field information in the free precession style Bell–Bloom atomic magnetometer. This capability enables widespread application of the free precession style Bell–Bloom atomic magnetometer in geomagnetic monitoring and disaster early warning within the geoscience domain. The shareable data are available athttps://github.com/baicaidezhenshi/MI-FPD-DATA.git.

The ground-airborne frequency-domain electromagnetic method (GAFDEM) is a geophysical technique designed for the efficient exploration of resistivity imaging in areas characterized by complex terrain. Conventional frequency-domain transformation methods are limited in accurately capturing the amplitude and phase of time-varying electromagnetic responses due to various noise and interference. As a result, the resistivity imaging results may be inaccurate and even exhibit false anomalies. To address this issue, we propose a novel orthogonal recursive least squares (ORLS) method. ORLS employs orthogonal signals as reference inputs and incorporates the adaptive filtering algorithm RLS, which enables real-time estimation of the amplitudes and phases of time-varying electromagnetic signals at multiple frequencies. The ORLS method overcomes the limitations of conventional frequency-domain transformation methods, such as the signal stationarity requirement and restricted frequency resolution. By ensuring real-time processing, ORLS enhances the accuracy of parameter estimation while maintaining efficiency. By simulating multifrequency signals and noisy data with different signal-to-noise ratios (SNRs), the effectiveness and stability of the ORLS algorithm are verified. Furthermore, the simulated and measured results are compared with frequency-domain analysis methods such as Fourier transform, which indicate that compared to frequency-domain analysis methods, the ORLS method reduces the average root mean square error (RMSE) of amplitude and phase by 57.79% and 85.97%, respectively, with similar estimation errors for each frequency component and no differences between frequencies. Moreover, the phase results of ORLS are easy to unwrap. Therefore, the GAFDEM data processed by the ORLS method hold significant importance in achieving high-precision underground resistivity imaging.

The transfer characteristic of the receiver sensor (TCRS) in a helicopter transient electromagnetic (HTEM) system can severely distort the measured data, especially early in the off-time. Removing TCRS is essential for an accurate and quantitative description of HTEM data. Precise Measurement and suppression of the TCRS are challenging due to measurement errors and noise in the HTEM system. We propose to use a calibration device to train the Wiener filter and then construct a TCRS deconvolution function to remove TCRS and enhance the HTEM system’s shallow detection capabilities. First, we design a calibration device and get its ideal and actual response in an existing HTEM system. Then the Wiener filter is trained using the measured and ideal response of the calibration device. Finally, we use a Wiener filter to obtain a recorded signal free of TCRS. Field tests show that the Wiener filter effectively suppresses the influence of TCRS on the recorded signal. After processing the observed signal with the Wiener filter, the blind time is reduced from 88.54 μs to 67.71 μs. The blind spot is reduced by 14.90 m for 100 Ω∙m homogeneous earth and 33.32 m for 500 Ω∙m homogeneous earth. The accuracy of the processed signals in channel two is improved by 16.93%. Accuracy for other channels is enhanced by 1.5%.

With the development of online real estate trading platforms, multi-modal housing trading data, including structural information, location, and interior image data, are being accumulated. The accurate appraisal of real estate makes sense for government officials, urban policymakers, real estate sellers, and personal purchasers. In this study, we propose an interpretable multi-modal stacking-based ensemble learning (IMSEL) method that deals with various modalities for real estate appraisals. We crawl the structural and image data of real estate in Chengdu city, China from the nation's largest real estate transaction platform with the location information, including public services, within 2 km of the real estate using Baidu map. We then compare the predictive results from IMSEL with those from previous state-of-art methods in the literature in terms of the root mean square error, mean absolute percentage error, mean absolute error, and coefficient of determination (R2). The comparison results show that IMSEL outperformed the other methods. We verified the improvement of introducing a data transformation strategy and deep visual features through a 10-fold cross-validation. We also discuss the managerial implications of our research findings.

Skin lesion diagnosis is a key step for skin cancer screening, which requires high accuracy and interpretability. Though many computer-aided methods, especially deep learning methods, have made remarkable achievements in skin lesion diagnosis, their generalization and interpretability are still a challenge. To solve this issue, we propose an interpretability-based multimodal convolutional neural network (IM-CNN), which is a multiclass classification model with skin lesion images and metadata of patients as input for skin lesion diagnosis. The structure of IM-CNN consists of three main paths to deal with metadata, features extracted from segmented skin lesion with domain knowledge, and skin lesion images, respectively. We add interpretable visual modules to provide explanations for both images and metadata. In addition to area under the ROC curve (AUC), sensitivity, and specificity, we introduce a new indicator, an AUC curve with a sensitivity larger than 80% (AUC_SEN_80) for performance evaluation. Extensive experimental studies are conducted on the popular HAM10000 dataset, and the results indicate that the proposed model has overwhelming advantages compared with popular deep learning models, such as DenseNet, ResNet, and other state-of-the-art models for melanoma diagnosis. The proposed multimodal model also achieves on average 72% and 21% improvement in terms of sensitivity and AUC_SEN_80, respectively, compared with the single-modal model. The visual explanations can also help gain trust from dermatologists and realize man-machine collaborations, effectively reducing the limitation of black-box models in supporting medical decision making.

Driving cycle is a crucial topic for the auto industry. It is developed to provide a quantitative measure on the fuel consumption and emission of a vehicle. In recent years, massive amount of driving data has been collected but has not yet been commonly used for the evaluation of driving cycle. We believe the collection of such data and the advancement in analytics models may provide a fresh perspective for the construction of driving cycle. Therefore, we propose a novel clustering-based optimization method for the construction of driving cycles. We employ the principal component analysis and spectral clustering algorithms to eliminate redundant features and analyze data structure. We further develop an adaptive optimization algorithm to select the appropriate kinematic segments to form a representative driving cycle. To demonstrate the effectiveness of our method, we compare our performance against the baselines including the New European Driving Cycle (NEDC), Federal Test Procedure (FTP), and Markov chain-based methods. The model performance is evaluated with real driving data from two cities in Fujian, China. Our proposed method is shown to be superior to all baselines. In addition, based on our optimized driving cycle, we can also estimate the fuel consumption to evaluate its energy economy. To sum up, this study offers a novel methodology to establish the driving cycle based on real and localized traffic data, where the constructed driving cycle can further be used for the development of energy economy and emission control.

ThispaperpresentsBigDL (adistributeddeeplearning framework for Apache Spark), which has been used by a variety of users in the industry for building deep learning applications on production big data platforms. It allows deep learning applications to run on the Apache Hadoop/Spark cluster so as to directly process the production data, and as a part of the end-to-end data analysis pipeline for deployment and management. Unlike existing deep learning frameworks, BigDL implements distributed, data parallel training directly on top of the functional compute model (with copy-on-write and coarse-grained operations) of Spark. We also share real-world experience and "war stories" of users that havead-optedBigDLtoaddresstheirchallenges(i.e., howtoeasilybuildend-to-enddataanalysisanddeep learning pipelines for their production data).

The growth of the number of people in the monitoring scene may increase the probability of security threat, which makes crowd counting more and more important. Most of the existing approaches estimate the number of pedestrians within one frame, which results in inconsistent predictions in terms of time. This study, for the first time, introduces a quadratic programming (QP) model with the network flow constraints to improve the accuracy of crowd counting. Firstly, the foreground of each frame is segmented into groups, each of which contains several pedestrians. Then, a regression‐based map is developed in accordance with the relationship between low‐level features of each group and the number of people in it. Secondly, a directed graph is constructed to simulate constraints on people's flow, whose vertices represent groups of each frame and arcs represent people moving from one group to another. Finally, by solving a QP problem with network flow constraints in the directed graph, the authors obtain consistency in people counting. The experimental results show that the proposed method can reduce the crowd counting errors and improve the accuracy. Moreover, this method can also be applied to any ultramodern group‐based regression counting approach to get improvements.

The small impact of the geomagnetic field enables a magnetic tensor gradiometer to be easily installed on the flight platform for airborne geophysical exploration, especially for the detection of shallow buried mines and magnetic moving targets. Using the fluxgates as the core components, the gradiometer has the advantages of wide temperature range, low cost, and high resolution, but has the disadvantage of relatively low accuracy. This is because of the scale factor error, the nonorthogonal error, the misalignment, the zero offset, the dynamic error in a fluxgate, and the inconsistency among the error of fluxgates. In this letter, a correction method for a towed airborne magnetic tensor gradiometer is proposed that is composed of four fluxgates arranged in a cross-shaped structure. The theoretical framework of the proposed method is based on the static error model and the dynamic characteristics of single fluxgate, the feature that the gradient tensor of the geomagnetic field at high altitude is approximately zero, and on the phenomenon that the unchanged tensor rotation as a result of nonuniform magnetic field on the ground can indicate the inconsistency of the scale factors among different tensor components. The actual flight results using a helicopter have demonstrated and validated the performance and effectiveness of the proposed method. The improvement ratios of the field tensor components are from 359.6 to 1765, and the RMS of each component has reached to the level of 1 nT/m.