Image Pairs Research Articles

Infrared and Visual Fusion Detection Framework for Oil and Gas Pipeline Environmental Inspection

The inspection of oil and gas pipeline retains important. These newly constructed pipelines may have defects near the circumferential welds and potential issues with water accumulation. Improper management of these issues can lead to severe consequences. Thus, the inner pipe structure detection prior to their operation is critically essential. Visual camera provides a direct view of the pipeline environment such as circumferential weld seams and water accumulation. In contrast, infrared thermal camera capture infrared radiation distribution from varied shapes and materials, offering an indirect insight into the pipeline environment. This paper designs an integrated infrared and visual NDT system for pipeline detection, capable of capturing time-synchronized infrared and visible image pairs along with corresponding odometer and gyroscope data for further processing. Moreover, an infrared and visual fusion detection algorithm is proposed, comprising coarse registration, base detectors, and probability-based late fusion strategy. The late fusion strategy enhances detection accuracy by leveraging the complementary information from both infrared and visual modalities. Furthermore, as the focus is primarily on the defects near the circumferential welds, the size of the circumferential welds in forward images can be used to predict their position in the circumferential visual detection system, achieving a 96.7% reduction in computational complexity compared to processing all images. The proposed system and algorithm were actually tested in a 26.8 km 1219 pipeline. The results verified the reliability of the system and the robustness and accuracy of the algorithm. The demo code is available at: [https://github.com/ljcuestc/Probability-based-late-fusion.git].

Refining disparity maps using deep learning and edge-aware smoothing filter

Stereo matching algorithm is crucial for applications that rely on three-dimensional (3D) surface reconstruction, producing a disparity map that contains depth information by computing the disparity values between corresponding points from a stereo image pair. In order to yield desirable results, the proposed stereo matching algorithm must possess a high degree of resilience against radiometric variation and edge inconsistencies. In this article convolutional neural network (CNN) is employed in the first stage to generate the raw matching cost, which is subsequently filtered with a bilateral filter (BF) and applied with cross-based cost aggregation (CBCA) during the cost aggregation stage to enhance precision. Winner-take-all (WTA) strategy is implemented to normalise the disparity map values. Finally, the resulting output is subjected to an edge-aware smoothing filter (EASF) to reduce the noise. Due to its resistance to high contrast and brightness, the filter is found to be effective in refining and eliminating noise from the output image. Despite discontinuities like adiron's lost cup handle or artl's shattered rods, this approach, based on experimental research utilizing a Middlebury standard validation benchmark, yields a high level of accuracy, with an average non-occluded error of 6.79%, comparable to other published methods.

Frequency-oriented hierarchical fusion network for single image raindrop removal.

Single image raindrop removal aims at recovering high-resolution images from degraded ones. However, existing methods primarily employ pixel-level supervision between image pairs to learn spatial features, thus ignoring the more discriminative frequency information. This drawback results in the loss of high-frequency structures and the generation of diverse artifacts in the restored image. To ameliorate this deficiency, we propose a novel frequency-oriented Hierarchical Fusion Network (HFNet) for raindrop image restoration. Specifically, to compensate for spatial representation deficiencies, we design a dynamic adaptive frequency loss (DAFL), which allows the model to adaptively handle the high-frequency components that are difficult to recover. To handle spatially diverse raindrops, we propose a hierarchical fusion network to efficiently learn both contextual information and spatial features. Meanwhile, a calibrated attention mechanism is proposed to facilitate the transfer of valuable information. Comparative experiments with existing methods indicate the advantages of the proposed algorithm.

Deep-learning model for background parenchymal enhancement classification in contrast-enhanced mammography

Background. Breast background parenchymal enhancement (BPE) is correlated with the risk of breast cancer. BPE level is currently assessed by radiologists in contrast-enhanced mammography (CEM) using 4 classes: minimal, mild, moderate and marked, as described in breast imaging reporting and data system (BI-RADS). However, BPE classification remains subject to intra- and inter-reader variability. Fully automated methods to assess BPE level have already been developed in breast contrast-enhanced MRI (CE-MRI) and have been shown to provide accurate and repeatable BPE level classification. However, to our knowledge, no BPE level classification tool is available in the literature for CEM. Materials and methods. A BPE level classification tool based on deep learning has been trained and optimized on 7012 CEM image pairs (low-energy and recombined images) and evaluated on a dataset of 1013 image pairs. The impact of image resolution, backbone architecture and loss function were analyzed, as well as the influence of lesion presence and type on BPE assessment. The evaluation of the model performance was conducted using different metrics including 4-class balanced accuracy and mean absolute error. The results of the optimized model for a binary classification: minimal/mild versus moderate/marked, were also investigated. Results. The optimized model achieved a 4-class balanced accuracy of 71.5% (95% CI: 71.2–71.9) with 98.8% of classification errors between adjacent classes. For binary classification, the accuracy reached 93.0%. A slight decrease in model accuracy is observed in the presence of lesions, but it is not statistically significant, suggesting that our model is robust to the presence of lesions in the image for a classification task. Visual assessment also confirms that the model is more affected by non-mass enhancements than by mass-like enhancements. Conclusion. The proposed BPE classification tool for CEM achieves similar results than what is published in the literature for CE-MRI.

Learning User Embeddings from Human Gaze for Personalised Saliency Prediction

Reusable embeddings of user behaviour have shown significant performance improvements for the personalised saliency prediction task. However, prior works require explicit user characteristics and preferences as input, which are often difficult to obtain. We present a novel method to extract user embeddings from pairs of natural images and corresponding saliency maps generated from a small amount of user-specific eye tracking data. At the core of our method is a Siamese convolutional neural encoder that learns the user embeddings by contrasting the image and personal saliency map pairs of different users. Evaluations on two public saliency datasets show that the generated embeddings have high discriminative power, are effective at refining universal saliency maps to the individual users, and generalise well across users and images. Finally, based on our model's ability to encode individual user characteristics, our work points towards other applications that can benefit from reusable embeddings of gaze behaviour.

ICNoduleNet: Enhancing Pulmonary Nodule Detection Performance on Sharp Kernel CT Imaging.

Thoracic computed tomography (CT) currently plays the primary role in pulmonary nodule detection, where the reconstruction kernel significantly impacts performance in computer-aided pulmonary nodule detectors. The issue of kernel selection affecting performance has been overlooked in pulmonary nodule detection. This paper first introduces a novel pulmonary nodule detection dataset named Reconstruction Kernel Imaging for Pulmonary Nodule Detection (RKPN) for quantifying algorithm differences between the two imaging types. The dataset contains pairs of images taken from the same patient on the same date, featuring both smooth (B31f) and sharp kernel (B60f) reconstructions. All other imaging parameters and pulmonary nodule labels remain entirely consistent across these pairs. Extensive quantification reveals mainstream detectors perform better on smooth kernel imaging than on sharp kernel imaging. To address suboptimal detection on the sharp kernel imaging, we further propose an image conversion-based pulmonary nodule detector called ICNoduleNet. A lightweight 3D slice-channel converter (LSCC) module is introduced to convert sharp kernel images into smooth kernel images, which can sufficiently learn inter-slice and inter-channel feature information while avoiding introducing excessive parameters. We conduct thorough experiments that validate the effectiveness of ICNoduleNet, it takes sharp kernel images as input and can achieve comparable or even superior detection performance to the baseline that uses the smooth kernel images. The evaluation shows promising results and proves the effectiveness of ICNoduleNet.

Deep learning-based super-resolution of structural brain MRI at 1.5T: application to quantitative volume measurement.

This study investigated the feasibility of using deep learning-based super-resolution (DL-SR) technique on low-resolution (LR) images to generate high-resolution (HR) MR images with the aim of scan time reduction. The efficacy of DL-SR was also assessed through the application of brain volume measurement (BVM). In vivo brain images acquired with 3D-T1W from various MRI scanners were utilized. For model training, LR images were generated by downsampling the original 1mm-2mm isotropic resolution images. Pairs of LR and HR images were used for training 3D residual dense net (RDN). For model testing, actual scanned 2mm isotropic resolution 3D-T1W images with one-minute scan time were used. Normalized root-mean-square error (NRMSE), peak signal-to-noise ratio (PSNR), and structural similarity (SSIM) were used for model evaluation. The evaluation also included brain volume measurement, with assessments of subcortical brain regions. The results showed that DL-SR model improved the quality of LR images compared with cubic interpolation, as indicated by NRMSE (24.22% vs 30.13%), PSNR (26.19 vs 24.65), and SSIM (0.96 vs 0.95). For volumetric assessments, there were no significant differences between DL-SR and actual HR images (p > 0.05, Pearson's correlation > 0.90) at seven subcortical regions. The combination of LR MRI and DL-SR enables addressing prolonged scan time in 3D MRI scans while providing sufficient image quality without affecting brain volume measurement.

New Method of Paired Comparison for Improved Observer Shortage Using Deep Learning Models

The aim of this study was to validate the potential of substituting an observer in a paired comparison with a deep-learning observer. Phantom images were obtained using computed tomography. Imaging conditions included a standard setting of 120 kVp and 200 mA, with tube current variations ranging from 160 mA, 120 mA, 80 mA, 40 mA, and 20 mA, resulting in six different imaging conditions. Fourteen radiologic technologists with >10 years of experience conducted pairwise comparisons using Ura's method. After training, VGG16 and VGG19 models were combined to form deep learning models, which were then evaluated for accuracy, recall, precision, specificity, and F1value. The validation results were used as the standard, and the results of the average degree of preference and significance tests between images were compared to the standard if the results of deep learning were incorporated. The average accuracy of the deep learning model was 82%, with a maximum difference of 0.13 from the standard regarding the average degree of preference, a minimum difference of 0, and an average difference of 0.05. Significant differences were observed in the test results when replacing human observers with AI counterparts for image pairs with tube currents of 160 mA vs. 120 mA and 200 mA vs. 160 mA. In paired comparisons with a limited phantom (7-point noise evaluation), the potential use of deep learning was suggested as one of the observers.

Validation of computer-based analysis of static ultrasound images of patellar and Achilles tendon enthesis territories

Aim: The aim of the present study is to conduct interobserver and intra-observer validation of computer analysis of static ultrasound images of entheseal territories of the Achilles and distal patellar tendons. Methods: Three rheumatologists with varying levels of experience underwent training in the use of ImageJ software for the analysis of 384 pairs of ultrasound images (long and short axis) from recorded studies of the Achilles and patellar tendons of both spondyloarthritis (SpA) patients and controls. Intra-observer and interobserver tests were conducted by calculating the differences in measurements of the same image at two different times by the same observer and by two different observers assessing the same image. The measurements included the area of analysis, the mean grayscale intensity, and the dispersion of grayscale intensity. Results: In the intra-observer test, no measurement showed a difference greater than 15%, ranging from 4.10% to 14.14%. In the interobserver test, no measurement exhibited a difference greater than 16%, ranging from 7.96% to 15.87%. The differences detected were evenly distributed among observers in both the intra-observer and inter-observer tests. Higher differences were detected in the analysis of images obtained from patient studies compared to control studies in almost all measurements. Conclusions: Whether analyzing control or patient ultrasound images of Achilles and patellar tendons, the intra-observer and interobserver agreement of computer-based analysis of static ultrasound images is more than acceptable and predominantly excellent.

Multi-task contrastive learning for change detection in remote sensing images

ABSTRACT This letter proposes a novel multi-task contrastive learning (MTCL) approach for change detection of high-resolution remote sensing images. Current self-supervised learning approaches have limitations in utilizing multiview information of remote sensing images, and they cannot directly train the change detection backbone network for feature learning. The investigated method exploits multiview information contained in remote sensing images to directly train the change detection backbone network. Specifically, multiview views are constructed from high-resolution remote sensing images by handcrafted feature extraction approach, which contain complementary information from different domains. Then each constructed view is utilized for contrastive learning, and the multi-task learning strategy of parameter sharing is used to learn more rich and robust information. Furthermore, instead of learning feature representations from single remote sensing image, the proposed approach directly utilizes image pairs to train the change detection backbone, further improving the performance of change detection network. Two benchmark datasets are employed for comparative experiments, and experimental results certify the effectiveness and superiority of the investigated approach.

True lateral imaging during lumbar medial branch radiofrequency neurotomy: Interobserver reliability

BackgroundTrue lateral imaging (TLI), obtained by superimposing bilateral lumbar spine structures and aligning superior endplate cortical bone, requires deliberate rotational adjustments of the laterally positioned fluoroscope in both the axial and longitudinal planes. True lateral segmental imaging is necessary to depict true and accurate radiofrequency (RF) cannula positioning relative to bony anatomy during lumbar medial branch radiofrequency neurotomy (LMBRFN). ObjectiveTo determine the interobserver reliability of TLI during LMBRFN. MethodsThis was a retrospective review of a prospectively generated collection of lateral fluoroscopic images to determine the interobserver reliability of TLI during LMBRFN. Lateral fluoroscopic images were prospectively collected from 34 consecutive L4-5 and L5-S1 LMBRFN procedures during routine clinical practice. Employing International Pain and Spine Intervention Society (IPSIS) LMBRFN and TLI techniques, an RF cannula was positioned parallel to the L3 and L4 medial branches and the L5 dorsal rami. During the normal course of TLI, untrue and final true lateral segmental images were obtained and saved. An original data set of 100 pairs of true and untrue lateral images was reviewed to verify true laterality using established criteria; disagreement was resolved by consensus or discarding ambiguous cases. To measure interobserver reliability (Cohen's Kappa), two blinded expert reviewers independently reviewed the image set, identifying the true lateral image and the plane requiring correction. ResultsThe observers agreed upon 98/98 true lateral RF-segment images (Kappa score 1.0 [1.00,1.00]). The observers agreed upon 86/98 maneuvers to correct the untrue RF-segment image. The Kappa score for determining the most appropriate corrective maneuver was 0.76 (0.63,0.89), showing substantial interobserver agreement. ConclusionsThe true lateral image of the targeted RF segment during LMBRFN was reliably determined with perfect interobserver agreement. Interobserver agreement was substantial regarding the maneuver to achieve TLI.

EMC+GD_C: circle-based enhanced motion consistency and guided diffusion feature matching for 3D reconstruction

AbstractRobust matching, especially the number, precision and distribution of feature point matching, directly affects the effect of 3D reconstruction. However, the existing methods rarely consider these three aspects comprehensively to improve the quality of feature matching, which in turn affects the effect of 3D reconstruction. Therefore, to effectively improve the quality of 3D reconstruction, we propose a circle-based enhanced motion consistency and guided diffusion feature matching algorithm for 3D reconstruction named EMC+GD_C. Firstly, a circle-based neighborhood division method is proposed, which increases the number of initial matching points. Secondly, to improve the precision of feature point matching, on the one hand, we put forward the idea of enhancing motion consistency, reducing the mismatch of high similarity feature points by enhancing the judgment conditions of true and false matching points; on the other hand, we combine the RANSAC optimization method to filter out the outliers and further improve the precision of feature point matching. Finally, a novel guided diffusion idea combining guided matching and motion consistency is proposed, which expands the distribution range of feature point matching and improves the stability of 3D models. Experiments on 8 sets of 908 pairs of images in the public 3D reconstruction datasets demonstrate that our method can achieve better matching performance and show stronger stability in 3D reconstruction. Specifically, EMC+GD_C achieves an average improvement of 24.07% compared to SIFT-based ratio test, 9.18% to GMS and 1.94% to EMC+GD_G in feature matching precision.

Triplet Epipolar Images for Spaceborne Three Linear Cameras

Abstract. Epipolar images can reduce conjugate points searching from a two-dimensional to a one-dimensional space, which significantly improves the robustness and efficiency of dense matching. However, two images matched the lack of redundancy and suffered from occlusions. The third image could introduce two more stereos and improve the accuracy and completeness of the three-dimension reconstruction, which were adopted by spaceborne three linear cameras (TLCs), the ALOS PRSIM and ZY-3. In this study, we propose a new triplet epipolar images generation method for spaceborne TLCs. Triplet epipolar images were defined as any two of the three epipolar image pairs. The triplet epipolar geometry requires the imaging rays of any conjugate point to be coplanar. For high-resolution satellite images, parallel projection can be used to approximate the imaging geometry of satellite images. Therefore, a coplanar principal optical axis is a fundamental requirement for spaceborne TLCs. We propose a general workflow to generate triplet epipolar images (TEIs), which includes free-net bundle adjustment of TLCs, building a triplet epipolar geometry, correcting the y-parallaxes, and generating a rational function model (RFM) for TEIs. The ZY3-02 satellite images were used to validate the proposed method. The root means square error (RMSEs) of the free-net bundle adjustment in the image space was 0.185 pixels, which proved the fine intrinsic accuracy. After compensation, the RMSEs of the y parallaxes of the three epipolar image pairs were 0.295, 0.310, and 0.370 pixels. Owing to the simple geometry of the TEIs, the RMSEs of the RFM replacements were within 0.001 pixels.

Learning content-aware feature fusion for guided depth map super-resolution

RGB-D data including paired RGB color images and depth maps is widely used in downstream computer vision tasks. However, compared with the acquisition of high-resolution color images, the depth maps captured by consumer-level sensors are always in low resolution. Within decades of research, the most state-of-the-art (SOTA) methods of depth map super-resolution cannot adaptively tune the guidance fusion for all feature positions by channel-wise feature concatenation with spatially sharing convolutional kernels. This paper proposes JTFNet to resolve this issue, which simulates the traditional Joint Trilateral Filter (JTF). Specifically, a novel JTF block is introduced to adaptively tune the fusion pattern between the color features and the depth features for all feature positions. Moreover, based on the variant of JTF block whose target features and guidance features are in the cross-scale shape, the fusion for depth features is performed in a bi-directional way. Therefore, the error accumulation along scales can be effectively mitigated by iteratively HR feature guidance. Compared with the SOTA methods, the sufficient experiment is conducted on the mainstream synthetic datasets and real datasets, i.e., Middlebury, NYU and ToF-Mark, which shows remarkable improvement of our JTFNet.

InSAR Digital Surface Model Refinement by Block Adjustment with Horizontal Constraints

Abstract. Interferometric Synthetic Aperture Radar (InSAR) technology is an important method to generate digital surface model (DSM). The past studies on space-borne derived DSM most focused on the elevation correction, due to the relative low resolution of DSM product. As a large number of high-resolution satellite data emerge, the horizontal discrepancies are needed to be considered. This paper proposes a DSM block adjustment method with horizontal constraints, aimed at eliminating the horizontal errors that exist between multiple DSM scenes with overlaps, achieving high precision and consistency in both the horizontal and vertical dimensions. Using ICESat-2 ATL08 point clouds as absolute elevation control and a reference DSM for horizontal control, the adjustment equations are constructed based on the constraint of tie points and controls. The experiment selects 7 image pairs of China TH2-01 SAR satellite, corresponding ICESat-2 ATL08 point and AW3D30 as reference DEM. The block adjustment results show that the proposed method improves the absolute vertical accuracy from 3.78 m to 2.56 m and reduces the average horizontal standard deviation between the InSAR derived DSMs and the reference AW3D30 from 15.31 m to 9.08 m.

Preliminary Derived DInSAR Coseismic Displacements of the 2022 Mw 5.7 Stolac Earthquake

On 22 April 2022, a Mw 5.7 earthquake was generated near Stolac (Bosnia and Herzegovina). The mainshock was succeeded by several aftershocks, three of which were significant. Two Mb 4.3 earthquakes occurred on 23 April 2022, and a Mw 4.8 earthquake was generated on 24 April 2022. Available data from fault mechanism solutions revealed that the mainshock activated a reverse fault, while the aftershock generated a normal fault with a right-lateral component. The Balkan Peninsula stands as one of the most active geodynamic areas in Central and Eastern Europe due to its location within the collision zone between Eurasian and African tectonic plates and the Anatolian microplate. Recorded earthquakes in Bosnia and Herzegovina are related to the energy generated by the subduction of the African tectonic plate under Eurasia. Furthermore, the seismicity of Bosnia and Herzegovina, particularly its southern part, is profoundly influenced by the subduction of the Adriatic microplate under the Dinarides. The Dinarides are a mainly fold and thrust belt that extends from the Southern Alps in the northwest to the Hellenides in the southeast and make dominant the tectonic system of Bosnia and Herzegovina. In this study, two pairs of SAR images obtained from the Sentinel-1 satellite mission were utilized to generate satellite LOS surface displacements using the DInSAR method. Moreover, LOS displacements were decomposed into vertical and east–west horizontal components by combining ascending and descending satellite orbits. Ultimately, the InSAR results were analyzed and compared with the data obtained from the CROPOS CORS GNSS station in Metković (MET3).

CVGSR: Stereo image Super-Resolution with Cross-View guidance

The ability to capture the complementary information of stereo images is critical to the development of stereo image super-resolution. Most existing studies have attempted to integrate reliable stereo correspondence along the polar directions through parallax attention and various fusion strategies. However, most of these approaches ignore the large parallax differences in stereo images, resulting in poor performance of convolutional-based parallax attention in capturing the long-range dependencies between images. In this paper, we propose a novel cross-view guided stereo image super-resolution network (CVGSR) for reconstructing high-resolution stereo image pairs with rich texture details by fully exploiting the complementary nature of stereo image pairs. Specifically, we first deploy a cross-view interaction module (CVIM) to explore intra/cross-view dependencies from local to global to compensate for the incomplete compatibility of information between the left and right views. This module uses a progressive cross-guiding strategy to better merge features from occluded and non-occluded regions. Based on this, an efficient attention Transformer (EAT) is improved to activate more input information and further mine the cross-view complementarity. Furthermore, we design a texture loss to optimize the visual perceptual quality of reconstructed images with sharp boundaries and rich texture details. Extensive experiments on four stereo image datasets demonstrate that the proposed CVGSR achieves a competitive and excellent performance.

Deep learning-based whole-body PSMA PET/CT attenuation correction utilizing Pix-2-Pix GAN.

Sequential PET/CT studies oncology patients can undergo during their treatment follow-up course is limited by radiation dosage. We propose an artificial intelligence (AI) tool to produce attenuation-corrected PET (AC-PET) images from non-attenuation-corrected PET (NAC-PET) images to reduce need for low-dose CT scans. A deep learning algorithm based on 2D Pix-2-Pix generative adversarial network (GAN) architecture was developed from paired AC-PET and NAC-PET images. 18F-DCFPyL PSMA PET-CT studies from 302 prostate cancer patients, split into training, validation, and testing cohorts (n = 183, 60, 59, respectively). Models were trained with two normalization strategies: Standard Uptake Value (SUV)-based and SUV-Nyul-based. Scan-level performance was evaluated by normalized mean square error (NMSE), mean absolute error (MAE), structural similarity index (SSIM), and peak signal-to-noise ratio (PSNR). Lesion-level analysis was performed in regions-of-interest prospectively from nuclear medicine physicians. SUV metrics were evaluated using intraclass correlation coefficient (ICC), repeatability coefficient (RC), and linear mixed-effects modeling. Median NMSE, MAE, SSIM, and PSNR were 13.26%, 3.59%, 0.891, and 26.82, respectively, in the independent test cohort. ICC for SUVmax and SUVmean were 0.88 and 0.89, which indicated a high correlation between original and AI-generated quantitative imaging markers. Lesion location, density (Hounsfield units), and lesion uptake were all shown to impact relative error in generated SUV metrics (all p < 0.05). The Pix-2-Pix GAN model for generating AC-PET demonstrates SUV metrics that highly correlate with original images. AI-generated PET images show clinical potential for reducing the need for CT scans for attenuation correction while preserving quantitative markers and image quality.

Application of image rendering based on improved neural networks and sensors in landscape design

Landscape planning and design is an indispensable part of modern urban construction. In landscape design, rendering technology can transform designers' imagination into actual images, allowing people to better understand the design intent. In order to solve the problem of insufficient precision in image rendering in landscape design, this paper proposes a landscape design image rendering technology based on improved neural network technology. This technology combines stereo matching algorithms, which evaluate the corrected stereo images to calculate the differences between the real scene and the underlying model. The stereo matching algorithm is based on the basic model and obtains accurate depth information from a wide range image pair, making the rendered landscape image more realistic and precise. The experiment demonstrates the effectiveness of the landscape design image rendering system through two important aspects: processing speed and parallax accuracy. The development of image rendering technology in landscape design cannot be separated from the mutual promotion of theory and practice. The technology proposed in this article has certain practical value and hopes to make more positive contributions to urban construction and cultural inheritance.

A road surface reconstruction dataset for autonomous driving

Recent developments in intelligent robot systems, especially autonomous vehicles, put forward higher requirements for safety and comfort. Road conditions are crucial factors affecting the comprehensive performance of ground vehicles. Nonetheless, existing environment perception datasets for autonomous driving lack attention to road surface areas. In this paper, we introduce the road surface reconstruction dataset, providing multi-modal, high-resolution, and high-precision data collected by real-vehicle platform in diverse driving conditions. It covers common road types containing approximately 16,000 pairs of stereo images, point clouds, and ground-truth depth/disparity maps, with accurate data processing pipelines to ensure its quality. Preliminary evaluations reveal the effectiveness of our dataset and the challenge of the task, underscoring substantial opportunities of it as a valuable resource for advancing computer vision techniques. The reconstructed road structure and texture contribute to the analysis and prediction of vehicle responses for motion planning and control systems.