Depth Estimation Method Research Articles

Depth Estimation Based on MMwave Radar and Camera Fusion with Attention Mechanisms and Multi-Scale Features for Autonomous Driving Vehicles

Autonomous driving vehicles have strong path planning and obstacle avoidance capabilities, which provide great support to avoid traffic accidents. Autonomous driving has become a research hotspot worldwide. Depth estimation is a key technology in autonomous driving as it provides an important basis for accurately detecting traffic objects and avoiding collisions in advance. However, the current difficulties in depth estimation include insufficient estimation accuracy, difficulty in acquiring depth information using monocular vision, and an important challenge of fusing multiple sensors for depth estimation. To enhance depth estimation performance in complex traffic environments, this study proposes a depth estimation method in which point clouds and images obtained from MMwave radar and cameras are fused. Firstly, a residual network is established to extract the multi-scale features of the MMwave radar point clouds and the corresponding image obtained simultaneously from the same location. Correlations between the radar points and the image are established by fusing the extracted multi-scale features. A semi-dense depth estimation is achieved by assigning the depth value of the radar point to the most relevant image region. Secondly, a bidirectional feature fusion structure with additional fusion branches is designed to enhance the richness of the feature information. The information loss during the feature fusion process is reduced, and the robustness of the model is enhanced. Finally, parallel channel and position attention mechanisms are used to enhance the feature representation of the key areas in the fused feature map, the interference of irrelevant areas is suppressed, and the depth estimation accuracy is enhanced. The experimental results on the public dataset nuScenes show that, compared with the baseline model, the proposed method reduces the average absolute error (MAE) by 4.7–6.3% and the root mean square error (RMSE) by 4.2–5.2%.

The Depth Estimation and Visualization of Dermatological Lesions: Development and Usability Study.

Thus far, considerable research has been focused on classifying a lesion as benign or malignant. However, there is a requirement for quick depth estimation of a lesion for the accurate clinical staging of the lesion. The lesion could be malignant and quickly grow beneath the skin. While biopsy slides provide clear information on lesion depth, it is an emerging domain to find quick and noninvasive methods to estimate depth, particularly based on 2D images. This study proposes a novel methodology for the depth estimation and visualization of skin lesions. Current diagnostic methods are approximate in determining how much a lesion may have proliferated within the skin. Using color gradients and depth maps, this method will give us a definite estimate and visualization procedure for lesions and other skin issues. We aim to generate 3D holograms of the lesion depth such that dermatologists can better diagnose melanoma. We started by performing classification using a convolutional neural network (CNN), followed by using explainable artificial intelligence to localize the image features responsible for the CNN output. We used the gradient class activation map approach to perform localization of the lesion from the rest of the image. We applied computer graphics for depth estimation and developing the 3D structure of the lesion. We used the depth from defocus method for depth estimation from single images and Gabor filters for volumetric representation of the depth map. Our novel method, called red spot analysis, measures the degree of infection based on how a conical hologram is constructed. We collaborated with a dermatologist to analyze the 3D hologram output and received feedback on how this method can be introduced to clinical implementation. The neural model plus the explainable artificial intelligence algorithm achieved an accuracy of 86% in classifying the lesions correctly as benign or malignant. For the entire pipeline, we mapped the benign and malignant cases to their conical representations. We received exceedingly positive feedback while pitching this idea at the King Edward Memorial Institute in India. Dermatologists considered this a potentially useful tool in the depth estimation of lesions. We received a number of ideas for evaluating the technique before it can be introduced to the clinical scene. When we map the CNN outputs (benign or malignant) to the corresponding hologram, we observe that a malignant lesion has a higher concentration of red spots (infection) in the upper and deeper portions of the skin, and that the malignant cases have deeper conical sections when compared with the benign cases. This proves that the qualitative results map with the initial classification performed by the neural model. The positive feedback provided by the dermatologist suggests that the qualitative conclusion of the method is sufficient.

Fusing Events and Frames with Coordinate Attention Gated Recurrent Unit for Monocular Depth Estimation.

Monocular depth estimation is a central problem in computer vision and robot vision, aiming at obtaining the depth information of a scene from a single image. In some extreme environments such as dynamics or drastic lighting changes, monocular depth estimation methods based on conventional cameras often perform poorly. Event cameras are able to capture brightness changes asynchronously but are not able to acquire color and absolute brightness information. Thus, it is an ideal choice to make full use of the complementary advantages of event cameras and conventional cameras. However, how to effectively fuse event data and frames to improve the accuracy and robustness of monocular depth estimation remains an urgent problem. To overcome these challenges, a novel Coordinate Attention Gated Recurrent Unit (CAGRU) is proposed in this paper. Unlike the conventional ConvGRUs, our CAGRU abandons the conventional practice of using convolutional layers for all the gates and innovatively designs the coordinate attention as an attention gate and combines it with the convolutional gate. Coordinate attention explicitly models inter-channel dependencies and coordinate information in space. The coordinate attention gate in conjunction with the convolutional gate enable the network to model feature information spatially, temporally, and internally across channels. Based on this, the CAGRU can enhance the information density of the sparse events in the spatial domain in the recursive process of temporal information, thereby achieving more effective feature screening and fusion. It can effectively integrate feature information from event cameras and standard cameras, further improving the accuracy and robustness of monocular depth estimation. The experimental results show that the method proposed in this paper achieves significant performance improvements on different public datasets.

MSDFNet: multi-scale detail feature fusion encoder–decoder network for self-supervised monocular thermal image depth estimation

Abstract Currently available thermal image depth estimation methods are difficult to efficiently extract fine multi-scale feature information from thermal images and suffer from the problem of blurring details at the edges of the estimated depth map. To address these challenges, this paper proposes MSDFNet, a multi-scale detail feature fusion encoder–decoder network, for self-supervised monocular thermal image depth estimation. The model is based on a channel expansion hourglass residual lightweight feature encoder, which can capture rich and fine-grained multi-scale feature information with low computational effort. MSDFNet utilizes a detail feature weight evaluation decoder to fuse cross-scale features and reevaluate the importance of each feature, thereby emphasizing critical edge information at multiple scales. Additionally, MSDFNet incorporates a depth consistency loss function, which provides self-supervised signals for the detailed features of thermal images and improves the optimization of network performance. The method is applied to the ViViD++ and MS2 datasets and achieves state-of-the-art depth estimation performance compared to existing state-of-the-art algorithms. In the Indoor Dark scenario of the ViViD++ dataset, the Abs Rel, Sq Rel, RMSE, and RMSE log error metric values of MSDFNet are reduced by 6.71%, 11.92%, 9.09%, and 5.73%, respectively, while the accuracy metric values δ < 1.25 i , i = 1 , 2 , 3 were improved by 4.18%, 1.13%, and 0.2%, respectively. In addition, MSDFNet proves its excellent generalization ability on the MS2 dataset. The Abs Rel and RMSE error values in the night scene are reduced by 45.6% and 30.09%, respectively, and the accuracy δ < 1.25 i , i = 1 , 3 is improved by 20.95% and 1.33%, respectively. The Abs Rel and RMSE values in the rainy day scenario are reduced by 1.33% and 1.21%, respectively, and the accuracy δ < 1.25 i , i = 1 , 3 is improved by 0.24% and 0.83%, respectively.

Text2NeRF: Text-Driven 3D Scene Generation With Neural Radiance Fields.

Text-driven 3D scene generation is widely applicable to video gaming, film industry, and metaverse applications that have a large demand for 3D scenes. However, existing text-to-3D generation methods are limited to producing 3D objects with simple geometries and dreamlike styles that lack realism. In this work, we present Text2NeRF, which is able to generate a wide range of 3D scenes with complicated geometric structures and high-fidelity textures purely from a text prompt. To this end, we adopt NeRF as the 3D representation and leverage a pre-trained text-to-image diffusion model to constrain the 3D reconstruction of the NeRF to reflect the scene description. Specifically, we employ the diffusion model to infer the text-related image as the content prior and use a monocular depth estimation method to offer the geometric prior. Both content and geometric priors are utilized to update the NeRF model. To guarantee textured and geometric consistency between different views, we introduce a progressive scene inpainting and updating strategy for novel view synthesis of the scene. Our method requires no additional training data but only a natural language description of the scene as the input. Extensive experiments demonstrate that our Text2NeRF outperforms existing methods in producing photo-realistic, multi-view consistent, and diverse 3D scenes from a variety of natural language prompts. Our code and model are available at https://github.com/eckertzhang/Text2NeRF.

A monocular three-dimensional object detection model based on uncertainty-guided depth combination for autonomous driving

Three-Dimensional (3D) object detection is a crucial task for enhancing safety and efficiency in autonomous driving. However, estimating depth from monocular images remains a challenging task. Most existing monocular 3D object detection methods rely on additional auxiliary data sources to compensate for the lack of spatial information in monocular images. Nevertheless, these methods bring substantial computational overhead and time-consuming preprocessing steps. To address this issue, we propose a novel depth estimation method for monocular images that does not rely on any auxiliary information. Leveraging both the texture and geometric cues of detected objects, our method generates two depth estimates for each object based on the extracted Region of Interest (RoI) features: a direct depth estimate and a height-based depth estimate with uncertainty modeling. Our model dynamically assigns weights to these depth estimates based on their respective uncertainties and combines them to obtain the final depth. During the training process, the model assigns higher weights to depth branches with higher uncertainties, as these estimates exhibit greater tolerance to errors. As the combined depth network introduces increased complexity, we utilize Group Normalization (GN) to better capture spatial information in the prediction branch outputs. Furthermore, we leverage the Two-Dimensional (2D) information of objects to predict the residual in the 2D center after downsampling, aiding in the regression of 3D center. On the KITTI benchmark, our model achieves an average precision (AP) of 16.65 % and 23.19 % on 3D and bird's-eye view (BEV) detection for the moderate category, surpassing the state-of-the-art (SOTA) models in each category.

New subsurface structural insights of Northeast Vietnam: Advanced implications from high-resolution magnetic data

Northeast Vietnam is covered by sedimentary rocks, and its subsurface structures can be determined by geophysical studies. However, the structural patterns in Northeast Vietnam have so far been poorly studied and there is a lack of addressing the complex geological features of this area. Outlining such geological features is commonly achieved by applying various edge detection methods to magnetic fields. In this research, we applied nine enhancement techniques to magnetic data of Northeast Vietnam to identify geological features in this area. The methods are tested on synthetic data before applying to RTP (reduction to pole) magnetic data that are obtained from the multiple stage RTP technique. Four semiautomated depth estimation methods are also used to obtain the depth of magnetic structures in Northeast Vietnam. The results indicate that the normalized total gradient and power transform of the tilt angle of the total horizontal gradient can produce geological features more clearly than others. The findings show that the study area is dominated by NE-SW and ENE-WSW trending lineaments, and most of magnetic sources are around 1–2.5 km deep in the area. The NE-SW and ENE-WSW trending structures are possibly related to the left-lateral motion along the Ailaoshan–Red River shear zone. Our findings also show that Northeast Vietnam may be considered as a southwestern continuation of the South China Block. Our results provide a better picture of the structural features of Northeast Vietnam and overcome the gap in understanding the tectonic and geological framework of this important area.

An improved ResNet method for urban flooding water depth estimation from social media images

Social media images featuring high timeliness, large data volume, and low cost, is utilized for flooding depth estimation. Common flood depth estimation methods are to segment each reference instance, resulting in local water depths and contradictory results. This study added Coordinate Attention into Residual Neural Network (ResNet), and proposed Coordinate Attention-Residual Neural Network (CA-ResNet) to acquire global submerged water depths. The 2021 Henan rainstorm in China was used as case study and a flooding dataset from Chinese Sina Microblog was constructed, with total 5676 images and 4189 flooded. All the flooded depths class 0–9 reflected from the images were annotated manually. The proposed CA-ResNet method was compared with VGGNet, ResNet, GoogleNet, and EfficientNet, both accuracy and F1 score improved by 1% − 3%. The average accuracy of level 0 – 8 within the error range of CA-ResNet reached 83.14%. It is proved that CA-ResNet could achieve flood depth estimation, facilitating efficient decision support for flood responding.

Underwater image color restoration based on depth estimation

Underwater images are often accompanied by color casts, and color restoration of underwater images remains a challenging problem. To tackle these degradation challenges, we present a depth estimation method based on image segmentation, and completes image color restoration on this basis. Specifically, we introduce image segmentation techniques to partition the image into discrete blocks. Our findings indicate a linear correlation between the average b component value in the Lab color space for each block and the corresponding image depth. Subsequently, the segmented images are sorted by depth to facilitate the identification of image blocks that are optimal for estimating backscatter. Local lighting estimation is then performed on these blocks to calculate the image transmission map, thereby completing the color restoration of the underwater image. The color restoration achieved through this method outperforms some advanced techniques. Our image color restoration method also demonstrates good accuracy and stability across different datasets.

Perspective Transform-based Depth Estimation of Monocular Camera for Electrocution Threat Determination of Construction Machinery

Abstract. This paper discusses the application of perspective transform-based monocular camera depth estimation method in monitoring the safety distance between construction machinery and power lines. The traditional distance measurement relies on manual operation, which is inefficient and not precise enough, while the monocular depth estimation method based on deep learning can improve the accuracy, but faces high equipment cost and relies on a large amount of training data. In contrast, the depth estimation method based on perspective transformation proposed in this paper utilizes mathematical and physical principles to establish a mathematical model between the pixel distances of an image and the actual scene distances through perspective projection theory, which simplifies the hardware requirements and reduces the data dependence and improves the computational efficiency and applicability. Through experimental validation under different construction scenes and exposure conditions, the method demonstrates high accuracy and robustness, proving its practicality in construction safety management. This research provides a new technical direction and practical application possibility in the field of intelligent monitoring and 3D reconstruction.

3D reconstruction and depth estimation method for local anomalies of rail surface based on multi-view stereo matching

Abstract Detecting rail surface anomalies has become crucial for ensuring the safety of train operations. However, manual detection methods are time-consuming and labor-intensive. Although traditional detection models based on one-dimensional signals or two-dimensional images can effectively identify the existence of defects, they are difficult to use to obtain local depth characteristic information, leading to difficulties in accurately assessing the extent of damage in abnormal regions. Rail track maintenance strategies are typically developed based on the detected different depth ranges of defects. To address this issue, this paper proposes a local depth estimation method based on the point cloud of the target rail surface reconstructed by PatchmatchNet. Firstly, according to the acquisition method of multi-view images and the sampling environment at the site, a practical data collection protocol is established for generating a multi-view dataset of rail surface. Next, dense point cloud of the target rail surface were reconstructed by PatchmatchNet. Subsequently, a method for estimating the local anomaly depth is developed based on the point cloud. Experimental results demonstrate that the proposed method achieves a minimum error of approximately 5.5% within a maximum depth range of 0.35mm, when compared to depth measurements obtained using a structured light camera. Finally, this paper presents a more comprehensive three-dimensional representation of the local abnormal physical structure, surpassing traditional one-dimensional and two-dimensional forms. This enhanced representation offers richer information for the effective determination of whether the anomalies constitute defects and aids in distinguishing true defects from other surface irregularities. Such advancement significantly improves the precision of defect assessment and supports the development of highly precise repair strategies.

SPDepth: Enhancing Self-Supervised Indoor Monocular Depth Estimation via Self-Propagation

Due to the existence of low-textured areas in indoor scenes, some self-supervised depth estimation methods have specifically designed sparse photometric consistency losses and geometry-based losses. However, some of the loss terms cannot supervise all the pixels, which limits the performance of these methods. Some approaches introduce an additional optical flow network to provide dense correspondences supervision, but overload the loss function. In this paper, we propose to perform depth self-propagation based on feature self-similarities, where high-accuracy depths are propagated from supervised pixels to unsupervised ones. The enhanced self-supervised indoor monocular depth estimation network is called SPDepth. Since depth self-similarities are significant in a local range, a local window self-attention module is embedded at the end of the network to propagate depths in a window. The depth of a pixel is weighted using the feature correlation scores with other pixels in the same window. The effectiveness of self-propagation mechanism is demonstrated in the experiments on the NYU Depth V2 dataset. The root-mean-squared error of SPDepth is 0.585 and the δ1 accuracy is 77.6%. Zero-shot generalization studies are also conducted on the 7-Scenes dataset and provide a more comprehensive analysis about the application characteristics of SPDepth.

An approximate method of estimating the depth of objects using wavelet-based transformation

Modern three-dimensional computer vision technologies are actively developing, offering new solutions based on the analysis of environmental details. One of the key tasks is to determine the depth of location of objects in images, which requires efficient and fast methods of their processing. This study proposes a new depth estimation method for 3D computer vision based on wavelet transform optimization. Solutions are offered to simplify the disparity calculation, which is traditionally used for the construction of depth maps, with the possibility of describing contours with adjustable detail based on the wavelet transform. The main advantage of this method is the increase in performance due to the allocation of the Haar wavelet carrier length in the extremum search area. The simulation confirmed the effectiveness of the proposed approach for constructing depth maps, which allows us to recommend it for use in unmanned aerial vehicles (UAVs) in conditions of limited computing and energy resources.

Learning Domain Invariant Features for Unsupervised Indoor Depth Estimation Adaptation

Predicting depth maps from monocular images has made an impressive performance in the past years. However, most depth estimation methods are trained with paired image-depth map data or multi-view images (e.g., stereo pair and monocular sequence), which suffer from expensive annotation costs and poor transferability. Although unsupervised domain adaptation methods are introduced to mitigate the reliance on annotated data, rare works focus on the unsupervised cross-scenario indoor monocular depth estimation. In this article, we propose to study the generalization of depth estimation models across different indoor scenarios in an adversarial-based domain adaptation paradigm. Concretely, a domain discriminator is designed for discriminating the representation from source and target domains, while the feature extractor aims to confuse the domain discriminator by capturing domain-invariant features. Further, we reconstruct depth maps from latent representations with the supervision of labeled source data. As a result, the feature extractor learned features possess the merit of both domain-invariant and low source risk, and the depth estimator can deal with the domain shift between source and target domains. We conduct the cross-scenario and cross-dataset experiments on the ScanNet and NYU-Depth-v2 datasets to verify the effectiveness of our method and achieve impressive performance.

Evaluation of gold mineralisation potential using AHP systems and weighted overlay analysis

The demand for sustainable development goals and the absence of systematic development and organised exploration for gold has prompted this study to integrate magnetic and radiometric datasets with lithology to evaluate the gold mineralisation potential in the Ilesha schist belt. This study considers 3168.72 km2 of the Ilesha schist belt in southwestern Nigeria, a frontier belt for gold deposits. The high-resolution airborne magnetic and radiometric datasets were processed using enhancement techniques, including the analytical signal, lineament density, and K/Th ratio. CET grid analysis, Euler deconvolution, and analytical signal depth estimation methods were used to aid the interpretation. The spatial integration and interpolation were performed using the Analytical Hierarchy Process (AHP) and weighted overlay analytical tools within the ArcGIS environment. The dominant structural controls for potential mineralisation are ENE–WSW and ESE–WNW trends. The depth of the magnetic sources revealed by the analytical signal ranged from 63.17 to 629.47 m, while depths ranging from 47.32 to 457.22 m were obtained from Euler deconvolution. The delineated highly magnetic edge sources, dense lineaments, radiometrically highlighted alteration zones, and lithological hosts for gold mineralisation were integrated to establish the gold mineralisation potential map. The AHP deductions reveal that 10.52% of the study site is within the high mineralisation potential class, a remarkable 60.39% falls within the moderate class, a significant portion (28.86%) falls within the poor class, and 0.23% is considered unfavourable. The result was optimised by validation using known mines, with 94% (i.e., 15 out of 16 mining sites) plotting within the high mineralisation potential class. This assessment provides invaluable insight for stakeholders and policymakers to embark on gold exploration and exploitation and promote sustainable mineral development.

Long-term reprojection loss for self-supervised monocular depth estimation in endoscopic surgery

Aim: Depth information plays a key role in enhanced perception and interaction in image-guided surgery. However, it is difficult to obtain depth information with monocular endoscopic surgery due to a lack of reliable cues for perceiving depth. Although there are reprojection loss-based self-supervised learning techniques to estimate depth and pose, the temporal information from the adjacent frames is not efficiently utilized to handle occlusion in surgery. Methods: We design long-term reprojection loss (LT-RL) self-supervised monocular depth estimation techniques by integrating longer temporal sequences into reprojection to learn better perception and to address occlusion artifacts in image-guided laparoscopic and robotic surgery. For this purpose, we exploit four temporally adjacent source frames before and after the target frame, where conventional reprojection loss uses two adjacent frames. The pixels that are visible in the target frame but occluded in the immediate two adjacent frames will produce the inaccurate depth but a higher chance to appear in the four adjacent frames during the calculation of minimum reprojection loss. Results: We validate LT-RL on the benchmark surgical datasets of Stereo correspondence and reconstruction of endoscopic data (SCARED) and Hamlyn to compare the performance with other state-of-the-art depth estimation methods. The experimental results show that our proposed technique yields 2%-4% better root-mean-squared error (RMSE) over the baselines of vanilla reprojection loss. Conclusion: Our LT-RL self-supervised depth and pose estimation technique is a simple yet effective method to tackle occlusion artifacts in monocular surgical video. It does not add any training parameters, making it flexible for integration with any network architecture and improving the performance significantly.

Matched cross‐spectrum phase processing for source depth estimation in deep water

AbstractZhou et al. (2022) proposed a method for instantaneous source depth estimation in deep water using matched beam intensity processing (MBIP) that leverages depth‐related oscillatory features arising from Lloyd's mirror interference in the frequency domain (Appl. Acoust., 2022, 186, 108493). However, the efficacy of MBIP diminishes with a random source spectrum. To counteract the effect of a random source spectrum, a vertical line array (VLA) positioned near the sea bottom is employed and a cross‐spectrum is generated from the conventional beamforming output, which is achieved by dividing the VLA into upper‐half and lower‐half arrays. The authors find that the cross‐spectrum phase interference pattern is determined by the source depth and is unaffected by the source spectrum. Consequently, a matched cross‐spectrum phase processing method is proposed, which demonstrates appreciable depth estimation accuracy compared to that of the existing techniques, as evidenced by the results from both simulated and experimental data.

A Self-Supervised Network-Based Smoke Removal and Depth Estimation for Monocular Endoscopic Videos.

In minimally invasive surgery videos, label-free monocular laparoscopic depth estimation is challenging due to smoke. For this reason, we propose a self-supervised collaborative network-based depth estimation method with smoke-removal for monocular endoscopic video, which is decomposed into two steps of smoke-removal and depth estimation. In the first step, we develop a de-endoscopic smoke for cyclic GAN (DS-cGAN) to mitigate the smoke components at different concentrations. The designed generator network comprises sharpened guide encoding module (SGEM), residual dense bottleneck module (RDBM) and refined upsampling convolution module (RUCM), which restores more detailed organ edges and tissue structures. In the second step, high resolution residual U-Net (HRR-UNet) consisting of a DepthNet and two PoseNets is designed to improve the depth estimation accuracy, and adjacent frames are used for camera self-motion estimation. In particular, the proposed method requires neither manual labeling nor patient computed tomography scans during the training and inference phases. Experimental studies on the laparoscopic data set of the Hamlyn Centre show that our method can effectively achieve accurate depth information after net smoking in real surgical scenes while preserving the blood vessels, contours and textures of the surgical site. The experimental results demonstrate that the proposed method outperforms existing state-of-the-art methods in effectiveness and achieves a frame rate of 94.45fps in real time, making it a promising clinical application.

Reconstructing the local structures of Chinese ancient architecture using unsupervised depth estimation

Digitalization of ancient architectures is one of the effective means for the preservation of heritage structures, with 3D reconstruction based on computer vision being a key component of such digitalization techniques. However, Chinese ancient architectures are located in mountainous areas, and existing 3D reconstruction methods fall short in restoring the local structures of these architectures. This paper proposes a self-attention-guided unsupervised single image-based depth estimation method, providing innovative technical support for the reconstruction of local structures in Chinese ancient architectures. First, an attention module is constructed based on features extracted from architectural images learned by the encoder, and then embedded into the encoder-decoder to capture the interdependencies across local features. Second, a disparity map is generated using the loss constraint network, including reconstruction matching, smoothness of the disparity, and left-right disparity consistency. Third, an unsupervised architecture based on binocular image pairs is constructed to remove any potential adverse effects due to unknown scale or estimated pose errors. Finally, with the known baseline distance and camera focal length, the disparity map is converted into the depth map to perform the end-to-end depth estimation from a single image. Experiments on the our architecture dataset validates our method, and it performs well also well on KITTI.

Depth Image Rectification Based on an Effective RGB–Depth Boundary Inconsistency Model

Depth image has been widely involved in various tasks of 3D systems with the advancement of depth acquisition sensors in recent years. Depth images suffer from serious distortions near object boundaries due to the limitations of depth sensors or estimation methods. In this paper, a simple method is proposed to rectify the erroneous object boundaries of depth images with the guidance of reference RGB images. First, an RGB–Depth boundary inconsistency model is developed to measure whether collocated pixels in depth and RGB images belong to the same object. The model extracts the structures of RGB and depth images, respectively, by Gaussian functions. The inconsistency of two collocated pixels is then statistically determined inside large-sized local windows. In this way, pixels near object boundaries of depth images are identified to be erroneous when they are inconsistent with collocated ones in RGB images. Second, a depth image rectification method is proposed by embedding the model into a simple weighted mean filter (WMF). Experiment results on two datasets verify that the proposed method well improves the RMSE and SSIM of depth images by 2.556 and 0.028, respectively, compared with recent optimization-based and learning-based methods.