Image Dehazing Research Articles

P‐3.20: A Novel Algorithm for Nighttime Image Dehazing

P‐3.20: A Novel Algorithm for Nighttime Image Dehazing

Haze optical-model-based nighttime image dehazing by modifying attenuation and atmospheric light.

Scene imaging is often affected by artificial light sources within a hazy environment at night, causing degraded images with low brightness, color distortion, and glow. These problems render the traditional atmospheric scattering optical model obsolete and incompatible. To address this issue, we established an optical imaging model suitable for nighttime dehazing, and an illumination component is incorporated into the attenuation term. We also introduced the near-light source coefficient to redefine the glow. Based on this model, we propose a new nighttime dehazing method. First, the rough atmospheric light is estimated using its low-frequency characteristics. Then, the glow is calculated by the near-light source coefficient. Finally, we remove the haze and illumination to get a clear image. Extensive experiments prove that our method exhibits a better color recovery effect, which effectively improves the visibility and detail. Furthermore, we believe our method outperforms other methods, both qualitatively and quantitatively.

ClarifyNet: A high-pass and low-pass filtering based CNN for single image dehazing

Dehazing refers to removing the haze and restoring the details from hazy images. In this paper, we propose ClarifyNet, a novel, end-to-end trainable, convolutional neural network architecture for single image dehazing. We note that a high-pass filter detects sharp edges, texture, and other fine details in the image, whereas a low-pass filter detects color and contrast information. Based on this observation, our key idea is to train ClarifyNet on ground-truth haze-free images, low-pass filtered images, and high-pass filtered images. Based on this observation, we present a shared-encoder multi-decoder model ClarifyNet which employs interconnected parallelization. While training, ground-truth haze-free images, low-pass filtered images, and high-pass filtered images undergo multi-stage filter fusion and attention. By utilizing a weighted loss function composed of SSIM loss and L1 loss, we extract and propagate complementary features. We comprehensively evaluate ClarifyNet on I-HAZE, O-HAZE, Dense-Haze, NH-HAZE, SOTS-Indoor, SOTS-Outdoor, HSTS, and Middlebury datasets. We use PSNR and SSIM metrics and compare the results with previous works. For most datasets, ClarifyNet provides the highest scores. On using EfficientNet-B6 as the backbone, ClarifyNet has 18 M parameters (model size of ∼71 MB) and a throughput of 8 frames-per-second while processing images of size 2048 × 1024.

RGB-Based Triple-Dual-Path Recurrent Network for Underwater Image Dehazing

In this paper, we present a powerful underwater image dehazing technique that exploits two image characteristics—RGB color channels and image features. In using RGB color channels, each color channel is decomposed into two units based on the similarities via the k-mean. This markedly improves the adaptability and identification of similar pixels, and thus reduces pixels with a weak correlation, leaving only pixels with a higher correlation. We use an infinite impulse response (IIR) in the triple-dual and parallel interaction structure to suppress hazed pixels via a pixel comparison and amplification to increase the visibility of even very minor features. This improves the visual perception of the final image, thus improving the overall usefulness and quality of the image. The softmax-weighted fusion is finally used to fuse the output color channel features to attain the final image. This preserves the color, leaving our proposed method’s output very true to the original scene’s. This is accomplished by taking advantage of adaptive learning based on the confidence levels of the pixel contribution variation in each color channel during subsequent fuses. The proposed technique both visually and objectively outperforms the existing methods in several rigorous tests.

Online knowledge distillation network for single image dehazing

Single image dehazing, as a key prerequisite of high-level computer vision tasks, catches more and more attentions. Traditional model-based methods recover haze-free images via atmospheric scattering model, which achieve favorable dehazing effect but endure artifacts, halos, and color distortion. By contrast, recent learning-based methods dehaze images by a model-free way, which achieve better color fidelity but tend to acquire under-dehazed results due to lacking of knowledge guiding. To combine these merits, we propose a novel online knowledge distillation network for single image dehazing named OKDNet. Specifically, the proposed OKDNet firstly preprocesses hazy images and acquires abundant shared features by a multiscale network constructed with attention guided residual dense blocks. After that, these features are sent to different branches to generate two preliminary dehazed images via supervision training: one branch acquires dehazed images via the atmospheric scattering model; another branch directly establishes the mapping relationship between hazy images and clear images, which dehazes images by a model-free way. To effectively fuse useful information from these two branches and acquire a better dehazed results, we propose an efficient feature aggregation block consisted of multiple parallel convolutions with different receptive. Moreover, we adopt a one-stage knowledge distillation strategy named online knowledge distillation to joint optimization of our OKDNet. The proposed OKDNet achieves superior performance compared with state-of-the-art methods on both synthetic and real-world images with fewer model parameters. Project website: https://github.com/lanyunwei/OKDNet.

Image Dehazing Technique Based on Sky Weight Detection and Fusion Transmission

<p>Computer vision techniques are widely applied to the object detection, license plate recognition, remote sensing, and outdoor monitoring system. The performance of these applications mainly relies on the high quality of outdoor image. However, an outdoor image can be led to contrast decrease, color distortion, and unclear structure by poor weather conditions and human factors such as haze, fog, and air pollution. These issues may lower down the sharpness of a photo. Despite of the single-image dehazing is used to solve these issues, it cannot achieve a satisfactory result when the method deals with the bright scene and sky area. In this article, we aim to design an adaptive dehazing technique based on fusion transmission and sky weight detection. The sky weight detection is employed to distinguish the foreground and background, while detected results are applied to the fusion strategy to calculate deep and shallow transmissions. Thus, this can get rids of the subject of over-adjustment. Experimental results have demonstrated that the new method can outperform the latest state-of-the-art methods in terms of subjective and the objective assessments.</p> <p> </p>

Polarimetric dehazing based on fusing intensity and degree of polarization

Image dehazing is an indispensable approach to address the image degeneration in hazy environment. In recent years, polarimetric methods have been widely investigated for image dehazing. In the basic algorithm of polarimetric dehazing, in addition to the dehazed intensity, the degree of polarization (DoP) can also be reconstructed with the assistance of Stokes parameters. Considering the additional information from the DoP, the polarimetric dehazing method is improved with the fusion of the intensity and the DoP. The wavelet transform is exploited to make fusion, and different fusion rules are designed for the low-frequency and the high-frequency components, respectively. The proposed method can improve the quality of polarimetric dehazed image, especially for the transparent object that is hard to be handled with previous methods. The validity of the proposed method is demonstrated in experiments.

Trends and Prospects of Techniques for Haze Removal From Degraded Images: A Survey

For the last two decades, image processing techniques have been used frequently in computer vision applications. The most challenging task in image processing is restoring images that are degraded due to various weather conditions. Mainly, the visibility of outdoor images is corrupted due to adverse atmospheric effects. The visibility of acquired images is reduced in these circumstances. Haze is an atmospheric phenomenon that reduces the clarity of an image. Due to the presence of particles such as dust, dirt, soot, or smoke, there is significant decay in the color and contrast of captured images. Haze present in acquired images causes issues in a variety of computer vision applications. Therefore, enhancing the contrast of a hazy image and restoring the visibility of the scene is essential. Since clear images are required in every application, image dehazing is an important step. Hence, many researchers are working on it. Different methods have been presented in the literature for image dehazing. This study describes various traditional and deep learning techniques of image dehazing from an analytical perspective. The main intention behind this work is to provide an intuitive understanding of the major techniques that have made a relevant contribution to haze removal. In this paper, we have covered different types of contributions toward dehazing based on the traditional method as well as deep learning approaches. With a considerable amount of instinctive simplifications, the reader is expected to have an improved ability to visualize the internal dynamics of these processes.

Single image dehazing with an independent Detail-Recovery Network

Single image dehazing is a prerequisite that affects the performance of many visually related tasks and has attracted increasing attention in recent years. However, most existing dehazing methods place more emphasis on haze removal but less on the detail recovery of the dehazed images. In this paper, we propose a single image dehazing method with an independent Detail Recovery Network (DRN), which considers capturing the details from the input image over a separate network and then integrating them into a coarse dehazed image. The overall network consists of two independent networks, named DRN and the dehazing network. Specifically, the DRN aims to recover the dehazed image details through the joint efforts of the local branch and the global branch. The local branch can obtain local detail information through the convolution layer, and the global branch can capture multi-scale global information by Smooth Dilated Convolution (SDC). In addition, we apply multi-faceted loss to improve the stability of the dehazing model. Extensive experiments on public image dehazing datasets illustrate the effectiveness of the modules in the proposed method and reveal that our method outperforms state-of-the-art dehazing methods. The code is released in https://github.com/YanLi-LY/Dehazing-DRN.

Structure‐aware dehazing of sewer inspection images based on monocular depth cues

AbstractIn sewer pipes, haze caused by the humid environment seriously impairs the quality of closed‐circuit television (CCTV) images, which leads to poor performance of subsequent pipe defects detection. Meanwhile, the complexity of sewer images, such as steep depth change and extensive textureless regions, brings great challenges to the performance or application of general dehazing algorithms. Therefore, this study estimates sewer depth maps first with the help of the water–pipewall borderlines to produce the paired dehazing dataset. Then a structure‐aware nonlocal network (SANL‐Net) is proposed with the detected borderlines and the dehazing result as two supervisory signals. SANL‐Net shows its superiority over other state‐of‐the‐art approaches with 147 in mean square error (MSE), 27.28 in peak signal to noise ratio (PSNR), 0.8963 in structural similarity index measure (SSIM), and 15.47M in parameters. Also, the outstanding performance in real image dehazing implies the accuracy of depth estimation. Experimental results indicate that SANL‐Net significantly improves the performance of defects detection tasks, such as an increase of 23.16% in mean intersection over union (mIoU) for semantic segmentation.

Image dehazing based on polarization information and deep prior learning

(objective)Generally, images taken under hazy weather have lower visibility and contrast. Therefore, how to improve their visibility and contrast and recover the image information is critical. Various dehazing algorithms have been proposed based on light intensity information. However, these methods mostly lack sufficient scattering information about atmospheric particles. Polarization supplies extra information independent of intensity, which can be used to effectively remove scattered light from haze particles, making it possible to help image dehazing. (methods)Therefore, this paper proposes a new image dehazing algorithm based on polarization information and deep prior learning (polarization-based unsupervised dehazing network, PUDN). Polarization information can provide more information about the image, while prior knowledge can constrain and optimize the network. This algorithm can remove haze without pretraining or corresponding ground-truth image. (result)Experimental results show that introducing polarization information and prior knowledge can improve the visibility and contrast of haze images, significantly improve haze image quality. Among them, the visible edge of heavy-haze image is improved by 71% on average, and the contrast is improved by 31 %. The dehazing performance is superior to that of traditional and supervised learning methods. Our method provides a new direction for image dehazing processing.

Rich feature distillation with feature affinity module for efficient image dehazing

Single-image haze removal is a long-standing hurdle for computer vision applications. Several works have been focused on transferring advances from image classification, detection, and segmentation to the niche of image dehazing, primarily focusing on contrastive learning and knowledge distillation. However, these approaches prove computationally expensive, raising concern regarding their applicability to on-the-edge use-cases. This work introduces a simple, lightweight, and efficient framework for single-image haze removal, exploiting rich “dark-knowledge” information from a lightweight pre-trained super-resolution model via the notion of heterogeneous knowledge distillation. We designed a feature affinity module to maximize the flow of rich feature semantics from the super-resolution teacher to the student dehazing network. In order to evaluate the efficacy of our proposed framework, its performance as a plug-and-play setup to a baseline model is examined. Our experiments are carried out on the RESIDE-Standard dataset to demonstrate the robustness of our framework to the synthetic and real-world domains. The extensive qualitative and quantitative results provided establish the effectiveness of the framework, achieving gains of up to 15% (PSNR) while reducing the model size by ∼20 times.

Image dehazing via enhancement, restoration, and fusion: A survey

Haze usually causes severe interference to image visibility. Such degradation on images troubles both human observers and computer vision systems. To seek high-quality images from degraded ones, a large number of image dehazing algorithms have been proposed from different perspectives like image enhancement, restoration, and fusion. Especially in recent years, with the rapid development of deep learning, CNN-based methods have already dominated the mainstream of image dehazing and gained significant progress on benchmark datasets. This paper firstly presents a comprehensive survey of existing image dehazing methods, and then conducts both qualitative and quantitative comparisons among representative methods, from classic methods to recent advanced approaches. We expect the literature survey and benchmark analysis could help readers better understand the advantages and limitations of existing dehazing methods. Moreover, a discussion on possible trends in single image dehazing is put forward to innovate further works.

Dehazing algorithm for unmanned surface vessel based on GAN-U-Net++ network

The autonomous perception of unmanned surface vessels in the natural sea environment has improved significantly; however, it is adversely affected by haze. Improving the U-Net network can retain and restore spatial information effectively and improve image dehazing. However, many dehazing algorithms that use the U-Net network as the main architecture face problems, such as distorted images and blurred contours of small distant objects. To overcome these limitations, we propose a generative adversarial network (GAN)-U-Net++ network that uses the U-Net++ network with a coordinate attention mechanism as the main part of a GAN generator. We used the coordinate attention mechanism to perform horizontal and vertical pixel level pooling to improve the attention on small targets in hazy images. The addition of the edge preservation loss function and image gradient loss function to small targets improved the sharpness of their edges. A Markov discriminator was used to evaluate the authenticity of the images. Several experiments were conducted to compare the proposed method with existing dehazing algorithms. The structural similarity value and peak signal-to-noise ratio increased by 15.74% and 29.14%, respectively.

Attention-Gate-Based Model with Inception-like Block for Single-Image Dehazing

In recent decades, haze has become an environmental issue due to its effects on human health. It also reduces visibility and degrades the performance of computer vision algorithms in autonomous driving applications, which may jeopardize car driving safety. Therefore, it is extremely important to instantly remove the haze effect on an image. The purpose of this study is to leverage useful modules to achieve a lightweight and real-time image-dehazing model. Based on the U-Net architecture, this study integrates four modules, including an image pre-processing block, inception-like blocks, spatial pyramid pooling blocks, and attention gates. The original attention gate was revised to fit the field of image dehazing and consider different color spaces to retain the advantages of each color space. Furthermore, using an ablation study and a quantitative evaluation, the advantages of using these modules were illustrated. Through existing indoor and outdoor test datasets, the proposed method shows outstanding dehazing quality and an efficient execution time compared to other state-of-the-art methods. This study demonstrates that the proposed model can improve dehazing quality, keep the model lightweight, and obtain pleasing dehazing results. A comparison to existing methods using the RESIDE SOTS dataset revealed that the proposed model improves the SSIM and PSNR metrics by at least 5–10%.

Non-Local Retinex Based Dehazing and Low Light Enhancement of Images

We know that a vast amount of research has recently been done on dehazing single images. More work is done on day-time images than night-time images. Also, enhancement of low light images is another area in which lots of research going on. In this paper, a simple yet effective unified variational model is proposed for dehazing of day and night images and low-light enhancement based on non-local global variational regularization. Given the relation between image dehazing and retinex, the haze removal process can minimize a variational retinex model. Estimating of ambient light and transmission maps is a key step in modern dehazing methods. Atmospheric light is not uniform and constant for hazy night images, as night scenes often contain multiple light sources. Often lit and non-illuminated regions have different colour characteristics and cause total variation colour distortion and halo artifacts. Our work directly implements a non-local retinal model based on the L2 norm that simulates the average activity of inhibitory and excitatory neuronal populations in the cortex to overcome this problem. This potential biological feasibility of the L2 norm of our work is divided into two parts using a filtered gradient approach, the reflection sparse prior and the reflection gradient fidelity before the observed image gradient. This unified framework of NLTV-Retinex and DCP efficiently performs low-light enhancement and dehazing of day and night images. We show results obtained using our method on daytime and night-time images and a low-light image dataset. We quantitatively and qualitatively compare our results with recently reported methods, which demonstrate the effectiveness of our method.

Single Image Dehazing Based on Two-Stream Convolutional Neural Network

Objective The haze weather environment leads to the deterioration of the visual effect of the image, and it is difficult to carry out the work of the advanced vision task. Therefore, dehazing the haze image is an important step before the execution of the advanced vision task. Traditional dehazing algorithms achieve image dehazing by improving image brightness and contrast or constructing artificial priors such as color attenuation priors and dark channel priors, but the effect is unstable when dealing with complex scenes. In the method based on convolutional neural network, the image dehazing network of the encoding and decoding structure does not consider the difference before and after the dehazing image, and the image spatial information is lost in the encoding stage. In order to overcome these problems, this paper proposes a novel end-to-end two-stream convolutional neural network for single image dehazing. Method The network model is composed of a spatial information feature stream and a high-level semantic feature stream. The spatial information feature stream retains the detailed information of the dehazing image, and the high-level semantic feature stream extracts the multi-scale structural features of the dehazing image. A spatial information auxiliary module is designed between the feature streams. This module uses the attention mechanism to construct a unified expression of different types of information, and realizes the gradual restoration of the clear image with the semantic information auxiliary spatial information in the dehazing network. A parallel residual twicing module is proposed, which performs dehazing on the difference information of features at different stages to improve the model’s ability to discriminate haze images. Result The peak signal-to-noise ratio and structural similarity are used to quantitatively evaluate the similarity between the dehazing results of each algorithm and the original image. The structure similarity and peak signal-to-noise ratio of the method in this paper reached 0.852 and 17.557dB on the Hazerd dataset, which were higher than all comparison algorithms. On the SOTS dataset, the indicators are 0.955 and 27.348dB, which are sub-optimal results. In experiments with real haze images, this method can also achieve excellent visual restoration effects.

Gradient Guided Dual-Branch Network for Image Dehazing

Recently, massive deep learning-based image dehazing methods have sprung up. These methods can effectively remove most of the haze and obtain far better results than the traditional methods. With the removal of the haze, however, edge details of the image are also lost, which is usually more noticeable in the gradient space. This paper proposes a gradient guided dual-branch network (GGDB-Net) for image dehazing. Specifically, we explore the hazy image gradient map to guide our model to focus on the hazy regions and edge restoration. We implement two parallel branches with a comprehensive loss function, which collaborate to dehaze and repair the lost edges in haze images. Moreover, the gradient-guided approach can potentially be applied to existing learning-based image dehazing models to boost their performance. Experimental results indicate that our results have good visual perceptions and are comparable to state-of-the-art methods in quantitative metrics.

Self-Adaptation Feature Attention Network with Multi-Step Fusion for Single Image Dehazing

Although single image dehazing has been widely studied as a common low-level computer vision task, it still faces serious challenges such as limited ability to dehaze real foggy pictures. We propose an efficient end-to-end self-adaptation feature attention (SAFA) network with multi-step fusion for this purpose. The proposed SAFA module can adaptively expand the receptive field to obtain the key structure information in space and extract more comprehensive and accurate features. In addition, considering the lack of connection between features acquired at low and high levels in the network, we also implement a multi-step fusion module, which makes the features of different layers in the network complementary effectively in the process of image recovery. The network structure is simplified, and the required computing resources are significantly reduced by decreasing network parameters. For multiple datasets and photographs with real haze, our method demonstrates better efficiency and availability, both quantitatively and qualitatively.

Single image dehazing algorithm based on optical diffraction deep neural networks.

Single image dehazing is a challenging task because of the hue and brightness distortion problems due to the atmospheric scattering. These problems limit the perceptual fidelity, as well as information integrity, of a given image. In this paper, we propose an image dehazing method based on the optical neural networks dehazing by simulating optical diffraction. The algorithm is trained from a large number of hazy images and their corresponding clean images. The experimental results demonstrate that the proposed method has reached an advanced level in both PSNR and SSIM dehazing performance indicators, and the amount of calculation is less than most artificial neural networks.