Poor Lighting Conditions Research Articles

CROSE: Low-light enhancement by CROss-SEnsor interaction for nighttime driving scenes

An increasing number of image perception models are being utilized in the field of autonomous driving. During nighttime driving, the visual perception capabilities of a single-modality RGB sensor become compromised. To address this issue, image fusion methods that utilize multi-modality data to provide a more comprehensive visual representation for nighttime scenes. Nevertheless, the existing image fusion methods are limited by poor generalization abilities and lighting conditions, resulting in inadequate handling of the contrast and spectral corruption. In this paper, we propose a cross-sensor low-light enhancement algorithm that provides more accurate visual perceptions. Our approach employs multiple sensors as preceptors for nighttime driving scenes. Compared with typical fusion algorithms that use a simple one-stage workflow for cross-sensor data, our proposed method adopts an advanced content-enhancement strategy by recursively and interactively scaling up informative pixels. More specifically, our approach employs information measurement to describe the information and then utilizes a cross-sensor content enhancement module to dynamically enhance the mutual information between the infrared spectrum and the RGB streams. Our experiments show that CROSE effectively preserves texture details, resulting in clearer and more solid fusion results for nighttime driving scenes.

Crosstalk of sucrose, potassium, and magnesium on growth, yield, and carbohydrate partitioning in bell pepper (Capsicum annuum L.) under poor light conditions

Bell pepper has received high demand due to its high nutritional value and contained antioxidant compounds. This research aimed to evaluate the effects of sucrose and some minerals on morpho-physiological characteristics, yield parameters, and carbohydrate partitioning in the bell pepper. The pot experiment was outlined hydroponically in a factorial format based on a completely randomized design. The first factor, sucrose, was sprayed in two concentrations (0 and 0.5% w/v) and the second factor, mineral nutrition, consisted of four levels (control, magnesium nitrate 0.5% w/v, potassium nitrate 0.5% w/v, magnesium nitrate 0.25% w/v + potassium nitrate 0.25% w/v). Results indicated that sucrose 5 + combined magnesium & potassium nitrate was superior in all of the biomass and yield parameters. In this combined treatment, fruit yield per plant and fruit vitamin C increased up to 44.03% and 47.71% in comparison with control. The highest leaf-soluble carbohydrate was recorded in plants treated with sucrose 5 + combined magnesium & potassium nitrate which was not significantly different from other contained sucrose treatments. The highest fruit soluble carbohydrate was observed in sucrose 5 + combined magnesium & potassium nitrate treatment as well, but its difference was not significant from sucrose 5 + potassium nitrate. The highest leaf potassium concentration was seen in three treatments that combined sucrose and minerals. The treatment of combined magnesium & potassium nitrate contributed to more leaf magnesium concentration compared with other nutrient treatments. In total, sucrose spray alongside combined magnesium & potassium nitrate led to higher carbohydrate accumulation in fruit, causing yield increment and nutritional value enhancement in bell peppers.

MDPose: Human Skeletal Motion Reconstruction Using WiFi Micro-Doppler Signatures

Motion tracking systems based on optical sensors typically suffer from poor lighting conditions, occlusion, limited coverage, and may raise privacy concerns. More recently, radiofrequency (RF) based approaches using commercial WiFi devices have emerged which offer low-cost ubiquitous sensing whilst preservin privacy. However, RF sensing systems typically output range-Doppler maps, time-frequency spectrograms, cross-range plots etc which cannot represent human motion intuitively and usually requires further processing. In this study, we propose MDPose, a novel framework for human skeletal motion reconstruction base on WiFi micro-Doppler signatures. MDPose provides an effective solution to represent human activity by reconstructing a skeleton model with 17 key points, which can assist with the interpretation of conventional RF sensing outputs in a more understandable way. Specifically, MDPose is implemented over three sequential stage to address a series of challenges: First, a denoising algorithm is employed to remove any unwanted noise that may affect feature extraction and enhance weak Doppler measurements. Secondly, a convolutional neural network (CNN)-recurrent neural network (RNN) architecture is applied to learn temporal spatial dependenc from clean micro-Doppler signatures and restore velocity information to key points under the supervision of the motion capture (Mocap) system. Finally, a pose optimisation mechanism based on learning optimisation vectors is employed to estimate the initial state of the skeleton and to limit additional errors. We hav conducted a comprehensive set of tests in a variety of environments using numerous subjects with a single receiver radar system to demonstrate the performance of MDPose, and report 29.4mm mean absolute error over all key points positions on several common daily activities, which has performance comparable to that of state-ofthe- art RF-based pose estimation systems.

An unmanned aerial vehicle light detection and ranging Simultaneous Localisation And Mapping algorithm based on factor graph optimisation for tunnel 3D mapping

AbstractThe current mature Simultaneous Localisation And Mapping (SLAM) algorithms, when applied to tunnel scenarios with point cloud degradation and poor lighting conditions, often lead to a sharp increase in the estimated attitude error of the unmanned aerial vehicle (UAV), or even prevent the UAV from moving autonomously due to severe feature degradation. To address the above problems, the authors propose a SLAM algorithm based on factor graph optimisation, Iterative Closest Point and Normal Distributions Transform algorithms. A front‐end point cloud registration module and a back‐end construction algorithm based on filtering and graph optimisation are designed. To verify the effectiveness of the proposed algorithm, experiments are conducted on KITTI dataset and real tunnel scenes, and compared with LiDAR Odometry and Mapping (LOAM) and lightweight and ground optimised (LeGO)‐LOAM algorithms. The results show that the average processing time of the proposed method is about 75 ms, which can meet the real‐time requirements of autonomous aerial vehicles. Compared with LOAM and LeGO‐LOAM in the real tunnel experiment, the proposed method shows the tunnel 3D map construction.

Deep Cross-view Reconstruction GAN based on Correlated Subspace for Multi-view Transformation.

In scenarios where identifying face information in the visible spectrum (VIS) is challenging due to poor lighting conditions, the use of near-infrared (NIR) and thermal (TH) cameras can provide viable alternatives. However, the unique data distribution of images captured by these cameras compared to VIS images presents challenges in matching face identities. To address these challenges, we propose a novel image transformation framework. The framework includes feature extraction from the input image, followed by a transformation network that generates target domain images with perceptual fidelity. Additionally, a reconstruction network preserves original information by reconstructing the original domain image from the extracted features. By considering the correlation between features from both domains, our framework utilizes paired data obtained from the same individual. We apply this framework to two well-established image-to-image transformation models, pix2pix and CycleGAN, known as CRC-pix2pix and CRC-CycleGAN respectively. The versatility of our approach allows extension to other models based on pix2pix or CycleGAN architectures. Our models generate high-quality images while preserving the identity information of the original face. Performance evaluation on TFW and BUAA NIR-VIS datasets demonstrates the superiority of our models in terms of generated image face matching and evaluation metrics such as SSIM, MSE, PSNR, and LPIPS. Moreover, we introduce the CQUPT-VIS-TH dataset, which enriches the paired dataset with thermal-visual face data capturing various angles and expressions.

Intelligent Target Classification Algorithm for 77G Radar Based on Correction Data Set

Traffic participant classification is crucial in autonomous driving perception. Millimeter wave radio detection and ranging radar is a cost-effective and powerful method to perform the task in adverse traffic scenarios, especially in bad inclement weather (e.g. fog, snow and rain) and poor lighting conditions. This paper presents an intelligent target classification algorithm for 77G radar based on correction data set. First, in order to handle the problem that the original data set may easily be interfered by obstacles, the angle information is filtered by analyzing the spatial information of radar signals, which means the interference clutter of obstacles can be effectively removed. Second, the primary data set is corrected using the significant difference between the micro-Doppler of the human body and car. Finally, the characteristic information of the radar signal is extracted, including distance, speed, orientation, micro-Doppler and reflection intensity, and the obtained data sets containing three types of targets (vehicles, human bodies and obstacles) are generated. The generated dynamic and static data sets are collected by sufficient experiments to construct deep learning classification models. The results show that the classification accuracy is improved by the measure of data set correction.

SSTM: Semantic Segmentation-based Template Matching method for heterogenous images

Global navigation satellite systems (GNSS) have disadvantages such as signal interference, loss, and spoofing. This has led to the development of new methods to complement or replace satellite navigation. Absolute visual localization is one of the main methods of vision-based localization. The goal is to localize the current view of the UAV in a reference satellite map or in a georeferenced image of a previous flight. During nighttime, due to poor lighting conditions, infrared images are used as a substitute for visible light images. However, the significant differences between infrared and visible light images pose challenges in matching them. Therefore, this paper proposes a Semantic Segmentation-based Template Matching (SSTM) method for heterogeneous images. The method proposed in this paper breaks the general idea of using a single feature extraction algorithm for heterogeneous image matching. In this paper, the matching of UAV infrared images with visible light satellite remote sensing images is realized by unifying two different image feature extraction modules of the feature description language and the matching scheme of this feature description language. The advantages of this method are demonstrated through comparisons with other advanced techniques. With the development of semantic segmentation techniques, it is believed that this modular “translation” method can achieve matching between arbitrary sensors.

Isolation and characterization of bacteria from the gut of a mesopelagic copepod Cephalophanes reflugens (Copepoda: Calanoida)

Abstract The mesopelagic calanoid copepod Cephalophanes reflugens is characterized by its large eyes with semi-parabolic reflectors. It is hypothesized that the eyes aid in finding crustacean carcasses as food in poor deep-sea light and oligotrophic conditions. Here, the crustacean carcasses-associated luminous, chitin-degrading bacteria may play a vital role in detecting and digesting foods by the copepod. In this study, we collected adult female C. reflugens (n = 9) from the mesopelagic layers off southwestern Japan in the Pacific Ocean and obtained a total of 47 bacterial isolates from the gut contents. None of the isolates were luminescent, but almost half of the isolates were chitinolytic. Based on the 16S rRNA gene sequences, these chitinolytic bacteria were identified as members of the genus Vibrio, Photobacterium and Pseudoalteromonas. In culture experiments at 15°C, most of the chitinolytic isolates efficiently utilized colloidal chitin as the sole carbon source and degraded the carapace strip of shrimp (Litopenaeus vannamei). Some isolates showed these chitinoclastic activities even at low temperatures (<5°C), equivalent to that of the deep-sea waters. The present results imply that chitin-degrading bacteria significantly contribute to chitin digestion in the gut of C. reflugens, whereas the involvement of luminous bacteria in the feeding mechanism remains unidentified.

RETRACTED ARTICLE: Dual Siamese Anchor Points Adaptive Tracker with Transformer for RGBT Tracking

Due to environmental conditions, such as rainy days, foggy days, and dim lighting, objects in visible light images are not prominently displayed, leading to an easy loss of targets during tracking. In recent years, many RGB visible light trackers have achieved significant success in addressing visual tracking challenges. However, these trackers perform poorly when tracking targets under special conditions, such as occlusions and low-light scenarios. In contrast, objects in thermal infrared images are more distinct in poor lighting conditions. Given this characteristic, researchers have shown increased interest in the development of trackers that combine thermal infrared and visible light imagery. However, some mainstream RGBT (red–green–blue and thermal) algorithms, such as MANET and ADNET, are based on the anchor-based theory, requiring consideration of anchor box sizes and introducing a substantial number of hyperparameters. This can lead to suboptimal performance when tracking dynamically changing targets. Moreover, these models rely on convolutional neural networks for feature extraction, which have limitations in capturing global features. In this paper, we introduce a novel training network model called DAPAT, which combines the anchor-free concept with Transformer theory. DAPAT differs from previous models in several ways. Specifically, we have designed a straightforward model to extract precise global features from template and search images. We have also incorporated two enhancement modules into the model to improve template and search images of different sizes while suppressing the impact of non-target images. We employ a dual-stream feature fusion network to reduce the loss of image feature information due to feature correlation operations. Finally, we compare the performance of the tracking model proposed in this paper with some advanced RGBT trackers on three data sets (RGBT234, RGBT210, and GTOT). The test results demonstrate that our tracker exhibits improvements in robustness and success rate, among other performance aspects.

Hierarchical waste detection with weakly supervised segmentation in images from recycling plants

Reducing environmental pollution with household waste and emissions from the computing clusters is an urgent technological problem. In our work, we explore both of these aspects: the deep learning application to improve the efficiency of waste recognition on recycling plant’s conveyor, as well as carbon dioxide emission from the computing devices used in this process. To conduct research, we developed an unique open WaRP dataset that demonstrates the best diversity among similar industrial datasets and contains more than 10,000 images with 28 different types of recyclable goods (bottles, glasses, card boards, cans, detergents, and canisters). Objects can overlap, be in poor lighting conditions, or significantly distorted. On the WaRP dataset, we study training and evaluation of cutting-edge deep neural networks for detection, classification and segmentation tasks. Additionally, we developed a hierarchical neural network approach called H-YC with weakly supervised waste segmentation. It provided a notable increase in the detection quality and made it possible to segment images, learning only having class labels, not their masks. Both the suggested hierarchical approach and the WaRP dataset have shown great industrial application potential.

Study on the Recognition of Coal Miners' Unsafe Behavior and Status in the Hoist Cage Based on Machine Vision.

The hoist cage is used to lift miners in a coal mine's auxiliary shaft. Monitoring miners' unsafe behaviors and their status in the hoist cage is crucial to production safety in coal mines. In this study, a visual detection model is proposed to estimate the number and categories of miners, and to identify whether the miners are wearing helmets and whether they have fallen in the hoist cage. A dataset with eight categories of miners' statuses in hoist cages was developed for training and validating the model. Using the dataset, the classical models were trained for comparison, from which the YOLOv5s model was selected to be the basic model. Due to small-sized targets, poor lighting conditions, and coal dust and shelter, the detection accuracy of the Yolov5s model was only 89.2%. To obtain better detection accuracy, k-means++ clustering algorithm, a BiFPN-based feature fusion network, the convolutional block attention module (CBAM), and a CIoU loss function were proposed to improve the YOLOv5s model, and an attentional multi-scale cascaded feature fusion-based YOLOv5s model (AMCFF-YOLOv5s) was subsequently developed. The training results on the self-built dataset indicate that its detection accuracy increased to 97.6%. Moreover, the AMCFF-YOLOv5s model was proven to be robust to noise and light.

Research on heart rate extraction method based on mobile phone video.

As an important indicator of human health, heart rate is related to the diagnosis of cardiovascular diseases. In recent years, extracting the heart rate from the mobile phone image has become a research hotspot. However, the illumination intensity of the background, frame rate of the video, and resolution of the image influence heart rate detection accuracy. To overcome these problems, this study proposed a novel heart rate extraction method based on mobile video. Firstly, the mobile phone camera is engaged to record the finger video, the region of interest (ROI) is extracted through the iterative threshold, and the pulse signal is obtained according to the grayscale change of the resolution within the ROI. Then, a low-pass and a high-pass Butterworth filters are exploited to filter out the noise and interframes from the extracted pulse signal. Finally, an improved adaptive peak extraction algorithm is proposed to detect the pulse peaks and the heart rate derived from the difference in pulse peaks. The experimental results show that light intensity, frame rate and resolution all have an influence on the heart rate extraction accuracy, with the most obvious influence of light, the average accuracy of the experiment can reach 99.32% under good lighting conditions, while only 72.23% under poor lighting conditions. In terms of frame rate, increasing the frame rate from 30fps to 60fps, the accuracy is improved by 0.9%. For the resolution, increasing the resolution from 1080p to 2160p, the accuracy is improved by 1.12%. While comparing the proposed method with existing methods, the proposed method has a higher accuracy rate, which has important practical value and application prospects in telemedicine and daily monitoring.

Target Detection Based on Two-Stream Convolution Neural Network With Self-Powered Sensors Information

With the rapid development of artificial intelligence, neural network is widely used in various fields. Target detection algorithm is mainly based on neural network, but the accuracy of target detection algorithm is greatly related to the complexity of scene and texture. A target detection algorithm based on RGB-D image from the perspective of the lightweight of target detection network model and the integration of depth map to overcome the weak environmental illumination with self-powered sensors information is proposed. This paper analyzes the network model structure of YOLOv4 and MobileNet, compares the variation of parameter number between depthwise separable convolution and convolutional neural network, and combines the advantages of YOLOv4 network and MobileNetv3 network. The main network of three effective feature layers in YOLOv4 is replaced by MobileNetv3 network for initial feature layer extraction to strengthen the feature extraction network. At the same time, the standard convolution models in the network are replaced by depthwise separable convolution. The proposed method is compared with YOLOv4 and YOLOv4-MobileNetv3 in this paper, and the experimental results show that the proposed network retains its original advantages in accuracy, but the size of the network model is about 23% of that of YOLOv4 network model, and the processing speed is about 42% higher than that of YOLOv4 network model, and the detection accuracy can still reach 83% in the environment with poor lighting conditions.

A Survey of Deep Learning-Based Low-Light Image Enhancement.

Images captured under poor lighting conditions often suffer from low brightness, low contrast, color distortion, and noise. The function of low-light image enhancement is to improve the visual effect of such images for subsequent processing. Recently, deep learning has been used more and more widely in image processing with the development of artificial intelligence technology, and we provide a comprehensive review of the field of low-light image enhancement in terms of network structure, training data, and evaluation metrics. In this paper, we systematically introduce low-light image enhancement based on deep learning in four aspects. First, we introduce the related methods of low-light image enhancement based on deep learning. We then describe the low-light image quality evaluation methods, organize the low-light image dataset, and finally compare and analyze the advantages and disadvantages of the related methods and give an outlook on the future development direction.

An exploratory investigation into the influence of risk factors on driver injury severity in angle crashes: A random parameter bivariate ordered probit model approach

Objective Angle crashes have been acknowledged as a concerning issue in the traffic safety field, though there is limited understanding of the contributions of risk factors to injury severity. This article aims to examine the impact of risk factors and unobserved heterogeneity on the severity of driver injuries in angle collisions by utilizing angle crash data in the United States from 2016 to 2021. Methods The relationship between risk factors and driver injury severities in angle crashes was investigated using a random parameter bivariate ordered probit model (RPBOP) with 4 categories of injury severity classified as outcome variables, including no injury, possible injury, minor injury, and serious jury. Risk factors were considered as explanatory variables, classified as driver characteristics, vehicle characteristics, road characteristics, environmental characteristics, time characteristics, and crash characteristics. Bayesian inference was used to assess the unobserved heterogeneity in risk factors, and marginal effects were computed to analyze the effect of each factor on injury outcomes. Results The findings demonstrate that risk factors have varying effects on driver involvement in angle crashes. Certain factors exhibited unobserved heterogeneity, including young drivers (ages 25–44), older drivers (over age 59), road grade, and collision point orientation. On the other hand, other factors, such as female gender, motorcycles, intersections, speed limit (>50 mph), poor lighting conditions, adverse weather, urban areas, and workdays, were shown to significantly increase the likelihood of driver injury in angle collisions, as well as increase susceptibility to fatal injury. Conclusions This article offers new insights into reducing driver injuries in angle crashes and has the potential to inform policy development aimed at preventing such incidents. Further research could utilize multisource data fusion and investigate the spatiotemporal stability of risk factors to enhance the generalizability of angle collision prevention strategies.

Multi-Resolution Edge-aware Lighting Enhancement Network

Images taken under poor lighting conditions generally present problems such as low brightness, noise, and color distortion, so dealing with low-light images is challenging. To solve this problem, current mainstream methods focus on designing special deep networks. For example, many methods are based on Retinex-based models. However, the solution of the Retinex model is an ill-conditioned problem in mathematics, so these methods are difficult to obtain the expected result. In this paper, following the new paradigm of embedding prior knowledge in deep learning, we design a simple yet surprisingly effective neural network, called Multi-Resolution Edge-aware Lighting Enhancement Network (MELENet). Since resolution and acutance are two of the key factors related to the visual quality of natural images, MELENet with a dual-branch structure aims to incorporate multi-resolution edge information into deep learning for guiding lighting enhancement. Additionally, by adopting special loss functions, the model suppresses noise and prevents color distortion while obtaining sound lighting enhancement effects. Experiments have verified that such a simple solution exhibits outstanding performance compared to existing mainstream lighting enhancement methods. We also directly apply the proposed model to image de-hazing and obtain state-of-the-art results.

IMPLEMENTATION OF IMAGE PROCESSING IN THE RECOGNITION OF OFFICIAL VEHICLE LICENSE PLATES

Vehicle license plates are identifiers used to uniquely identify vehicles. However, to identify vehicle license plates there are several problems encountered, namely the different formats of vehicle license plates that make license plate recognition more complicated, vehicle license plates often contain visually similar combinations of letters and numbers (for example the letter "O" and the number "0" or the letter "I" and the number "1"), . in poor lighting conditions license plates may not be clearly visible. To solve this problem, image recognition, image processing, and pattern recognition technologies can be used. The three technologies can be used to recognize characters on vehicle license plates, but cannot yet be used to recognize the colors contained on vehicle license plates. The purpose of this research is to identify and record vehicle license plate numbers quickly and accurately, monitor the presence of vehicles in a supervised area, assist in managing parking, reduce the need for human interaction in the vehicle identification process, The methods used to recognize motor vehicle plates are edge detection and character segmentation which involves image processing to detect the edges of the vehicle plate, followed by segmentation of individual characters in the plate. Another method used is optical character recognition which involves using an optical sensor to take an image of a vehicle plate, then using character recognition techniques to identify the numbers and letters on the plate. The result of this research is that the motor vehicle number recognition system can work in various lighting conditions and poor weather conditions and can monitor and control vehicles in the parking area. The finding obtained from this research is that no method has been used for color recognition on motor vehicle plates.

A Review of Document Image Enhancement Based on Document Degradation Problem

Document image enhancement methods are often used to improve the accuracy and efficiency of automated document analysis and recognition tasks such as character recognition. These document images could be degraded or damaged for various reasons including aging, fading handwriting, poor lighting conditions, watermarks, etc. In recent years, with the improvement of computer performance and the continuous development of deep learning, many methods have been proposed to enhance the quality of these document images. In this paper, we review six tasks of document degradation, namely, background texture, page smudging, fading, poor lighting conditions, watermarking, and blurring. We summarize the main models for each degradation problem as well as recent work, such as the binarization model that can be used to deal with the degradation of background textures, lettering smudges. When facing the problem of fading, a model for stroke connectivity can be used, while the other three degradation problems are mostly deep learning models. We discuss the current limitations and challenges of each degradation task and introduce the common public datasets and metrics. We identify several promising research directions and opportunities for future research.

Understanding how indoor environmental classroom conditions influence academic performance in higher education

PurposeThis study aims to qualitatively examine the relationship between the indoor environmental quality (IEQ), lecturers’ and students’ perceived internal responses and academic performance.Design/methodology/approachTo capture user experiences with the IEQ in classrooms, semi-structured interviews with 11 lecturers and three focus group discussions with 24 students were conducted, transcribed, coded and analyzed using direct content analysis.FindingsThe findings show that lecturers and students experience poor thermal, lighting, acoustic and indoor air quality (IAQ) conditions that may influence their ability to teach and learn. Maintaining acceptable thermal and IAQ conditions was difficult for lecturers, as opening windows or doors caused noise disturbances. In uncomfortable conditions, lecturers may decide to give a break earlier or shorten a lecture. When students experienced discomfort, it may affect their ability to concentrate, their emotional status and their quality of learning.Research limitations/implicationsThe findings originate from a relatively small sample, which might have limited the number and variety of identified associations between environment and users.Practical implicationsMaintaining acceptable air and thermal conditions will mitigate the need to open windows and doors. Keeping doors and windows closed will prevent noise disturbances and related distractions. This will support the quality of learning in classrooms. This study reveals the end users’ perspectives and preferences, which can inspire designers of new school buildings in higher education.Originality/valueThis study emphasizes the importance of creating and maintaining optimal IEQ conditions to support the quality of teaching and learning. These conditions are particularly relevant when classroom occupancy rates are high or outdoor conditions are unfavourable.

Vision in the Vertical Axis: How Important Are Visual Cues in Foraging and Navigation?

In both terrestrial and aquatic environments, a large number of animal behaviors rely on visual cues, with vision acting as the dominant sense for many fish. However, many other streams of information are available, and multiple cues may be incorporated simultaneously. Being free from the constraints of many of their terrestrial counterparts, fish have an expanded range of possible movements typified by a volume rather than an area. Cues such as hydrostatic pressure, which relates to navigation in a vertical plane, may provide more salient and reliable information to fish as they are not affected by poor light conditions or turbidity. Here, we tested banded tetra fish (Astyanax fasciatus) in a simple foraging task in order to determine whether visual cues would be prioritized over other salient information, most notably hydrostatic pressure gradients. We found that in both vertical and horizontal arrays there was no evidence for fish favoring one set of cues over the other, with subjects making choices at random once cues were placed into conflict. Visual cues remained as important in the vertical axis as they were in the horizontal axis.