Daytime Images Research Articles

Analysis of applying machine learning methods in rock classification problems on core samples

Relevance. Core studies in the oil and gas industry allow us to obtain some filtration and capacity characteristics of rocks, as well as to give an idea of the composition and structure of the subsoil. This information is extremely important in the early stages of field development, as it allows us to formulate a primary version of the development project, which is then refined during the course of field drilling. However, core analysis and description are an extremely labor-intensive and human-influenced works that require automation. Thus, core image research is a popular task in the oil and gas industry that requires high accuracy and care during work; especially considering the volume of images that have to be analyzed. Aim. To review and analyze existing algorithms for classifying rocks from core images based on machine learning methods; as well as to use the information obtained to formulate recommendations for the development of these algorithms. Methods. Machine learning methods, including neural networks. Results. The work analyzed existing approaches to the study of core samples. The main advantages and disadvantages of each of them were noted and, based on the findings, a plan and requirements for conducting further research on core samples using machine learning were developed. As a result, using a convolutional neural network on the U-Net architecture, the authors have trained a model to solve the problem of segmentation of core samples in daytime images and presented the results of the model operation.

Nighttime scene understanding with label transfer scene parser

Semantic segmentation plays a crucial role in traffic scene understanding, especially in nighttime conditions. This paper tackles the task of semantic segmentation in nighttime scenes. The largest challenge of this task is the lack of annotated nighttime images to train a deep learning-based scene parser. The existing annotated datasets are abundant in daytime conditions but scarce in nighttime due to the high cost. Thus, we propose a novel Label Transfer Scene Parser (LTSP) framework for nighttime scene semantic segmentation by leveraging daytime annotation transfer. Our framework performs segmentation in the dark without training on real nighttime annotated data. In particular, we propose translating daytime images to nighttime conditions to obtain more data with annotation in an efficient way. In addition, we utilize the pseudo-labels inferred from unlabeled nighttime scenes to further train the scene parser. The novelty of our work is the ability to perform nighttime segmentation via daytime annotated labels and nighttime synthetic versions of the same set of images. The extensive experiments demonstrate the improvement and efficiency of our scene parser over the state-of-the-art methods with a similar semi-supervised approach on the benchmark of Nighttime Driving Test dataset. Notably, our proposed method utilizes only one-tenth of the amount of labeled and unlabeled data in comparison with the previous methods. Code is available at https://github.com/danhntd/Label_Transfer_Scene_Parser.git.

Long-Range Imaging of Alpha Emitters Using Radioluminescence in Open Environments: Daytime and Night-Time Applications.

Alpha emitters like plutonium pose severe health risks when ingested, damaging DNA and potentially causing cancer. Traditional detection methods require proximity within millimeters of the contamination source, presenting safety risks and operational inefficiencies. Long-range detection through alpha radioluminescence (RL) offers a promising alternative. However, most of the previous experiments have been carried out under controlled conditions that preclude the overwhelming effect of ambient light. This study demonstrates the successful detection of a 3 MBq alpha emitter in an open environment using a compact alpha camera. This camera incorporates a deep-cooled CCD and a low f-number lens system designed to minimize the blue shift effects of filters. Night-time imaging was achieved with a dual-filter system using a sandwich filter assembly centered at 337 nm and 343 nm for capturing alpha RL and subtracting background light, respectively. At night, the alpha source was detected from 1 m away within one minute, and the lowest detection limit can be calculated as 75 kBq. The system was also evaluated under simulated urban lighting conditions. For daytime imaging, a stack of tilted 276 nm short pass filters minimized sunlight interference, enabling the detection of the alpha source at 70 cm within 10 min under indirect sunlight. This research highlights the viability of long-range optical detection of alpha emitters for environmental monitoring in real-world settings.

Development of a Device for Monitoring Erosion in the Field.

Monitoring erosion is an important part of understanding the causes of this geotechnical and geological phenomenon. In order to monitor them, it is necessary to develop equipment that is sophisticated enough to resist the sun and water without damage, that is self-mechanized, and that can support the amount of data collected. This article introduces a rain-triggered field erosion monitoring device composed of three main modules: control, capture, and sensing. The control module comprises both hardware and firmware with embedded software. The capture module integrates a camera for recording, while the sensing module includes rain sensors. By filming experimental soil samples under simulated rain events, the device demonstrated satisfactory performance in terms of activation and deactivation programming times, daytime image quality without artificial lighting, and equipment protection. The great differences about this monitoring device are its ease of use, low cost, and the quality it offers. These results suggest its potential effectiveness in capturing the progression of field erosive processes.

Surface urban heat island analysis based on local climate zones using ECOSTRESS and Landsat data: A case study of Valencia city (Spain)

Extreme heat events in cities are becoming more intense, frequent and prolonged. The spatiotemporal dynamics of surface temperature are closely related to land cover and atmospheric conditions, especially for urban areas with extensive surface heterogeneity. Two high spatial resolution satellites, ECOSTRESS and Landsat, provide land surface temperature (LST) for a typical Mediterranean climate city, Valencia, Spain. In total of 17 images throughout two heatwaves in the period of July and August from 2022 to 2023 were selected to monitor LST and surface urban heat island (SUHI). The Local Climate Zone (LCZ) scheme and hotspot analysis were applied to analyze the patterns of LST and SUHI during and after the heatwaves. Daytime images show LCZ 2/3/6/8 and LCZ C/E/F exhibit elevated SUHI intensity, meanwhile the LCZ 4, A/B and G have lower SUHI. The SUHI intensity of compact buildings (LCZ 1/2/3), LCZ 8/E/G are significantly higher at night, especially on heatwave days, the SUHI of all types increased by 0.1–0.9 °C. Furthermore, the meteorological parameters were introduced trying to explain the obvious diurnal difference of SUHI during heatwaves. This research serves as a proxy for understanding the spatiotemporal dynamics of SUHI during heatwave events, offering valuable insights for city planners and policymakers to enhance thermal comfortable level and effectively cope with extreme weather.

Automatic identification and quantification of volcanic hotspots in Alaska using HotLINK: the hotspot learning and identification network

An increase in volcanic thermal emissions can indicate subsurface and surface processes that precede, or coincide with, volcanic eruptions. Space-borne infrared sensors can detect hotspots—defined here as localized volcanic thermal emissions—in near-real-time. However, automatic hotspot detection systems are needed to efficiently analyze the large quantities of data produced. While hotspots have been automatically detected for over 20 years with simple thresholding algorithms, new computer vision technologies, such as convolutional neural networks (CNNs), can enable improved detection capabilities. Here we introduce HotLINK: the Hotspot Learning and Identification Network, a CNN trained to detect hotspots with a dataset of −3,800 satellite-based, Visible Infrared Imaging Radiometer Suite (VIIRS) images from Mount Veniaminof and Mount Cleveland volcanoes, Alaska. We find that our model achieves an accuracy of 96% (F1-score 0.92) when evaluated on −1,700 unseen images from the same volcanoes, and 95% (F1-score 0.67) when evaluated on −3,000 images from six additional Alaska volcanoes (Augustine Volcano, Bogoslof Island, Okmok Caldera, Pavlof Volcano, Redoubt Volcano, Shishaldin Volcano). In comparison with an existing threshold-based hotspot detection algorithm, MIROVA (Coppola et al., Geological Society, London, Special Publications, 2016, 426, 181–205), our model detects 22% more hotspots and produces 12% fewer false positives. Additional testing on −700 labeled Moderate Resolution Imaging Spectroradiometer (MODIS) images from Mount Veniaminof demonstrates that our model is applicable to this sensor’s data as well, achieving an accuracy of 98% (F1-score 0.95). We apply HotLINK to 10 years of VIIRS data and 22 years of MODIS data for the eight aforementioned Alaska volcanoes and calculate the radiative power of detected hotspots. From these time series we find that HotLINK accurately characterizes background and eruptive periods, similar to MIROVA, but also detects more subtle warming signals, potentially related to volcanic unrest. We identify three advantages to our model over its predecessors: 1) the ability to detect more subtle volcanic hotspots and produce fewer false positives, especially in daytime images; 2) probabilistic predictions provide a measure of detection confidence; and 3) its transferability, i.e., the successful application to multiple sensors and multiple volcanoes without the need for threshold tuning, suggesting the potential for global application.

Satellite Time-Series Analysis for Thermal Anomaly Detection in the Naples Urban Area, Italy

Naples is the most densely populated Italian city (7744 inhabitants per km2). It is located in a particular geological context: the presence of Mt Vesuvius characterizes the eastern part, and the western part is characterized by the presence of the Phlegrean Fields, making Naples a high-geothermal-gradient region. This endogenous heat, combined with the anthropogenic heat due to intense urbanization, has defined Naples as an ideal location for Surface Urban Heat Island (SUHI) analysis. SUHI analysis was effectuated by acquiring the Land Surface Temperature (LST) over Naples municipality by processing Landsat 8 (L8) Thermal Infrared Sensor (TIRS) images in the 2013–2023 time series by employing Google Earth Engine (GEE). In GEE, two different approaches have been followed to analyze thermal images, starting from the Statistical Mono Window (SMW) algorithm, which computes the LST based on the brightness temperature (Tb), the emissivity value, and the atmospheric correction coefficients. The first one is used for the LST retrieval from daytime images; here, the emissivity component is derived using, firstly, the Normalized Difference Vegetation Index (NDVI) and then the Vegetation Cover Method (VCM), defining the Land Surface Emissivity (LSɛ), which considers solar radiation as the main source of energy. The second approach is used for the LST retrieval from nighttime images, where the emissivity is directly estimated from the Advance Spaceborne Thermal Emission Radiometer database (ASTER-GED), as, during nighttime without solar radiation, the main source of energy is the energy emitted by the Earth’s surface. From these two different algorithms, 123 usable daytime and nighttime LST images were downloaded from GEE and analyzed in Quantum GIS (QGIS). The results show that the SUHI is more concentrated in the eastern part, characterized by intense urbanization, as shown by the Corine Land Cover (CLC). At the same time, lower SUHI intensity is detected in the western part, defined by the Land Cover (LC) vegetated class. Also, in the analysis, we highlighted 40 spots (10 hotspots and 10 coldspots, both for daytime and nighttime collection) that present positive or negative temperature peaks for all the time series. Due to the huge amount of data, this work considered only the five representative spots that were most representative for SUHI analysis and determination of thermal anomalies in the urban environment.

ISP-Teacher:Image Signal Process with Disentanglement Regularization for Unsupervised Domain Adaptive Dark Object Detection

Object detection in dark conditions has always been a great challenge due to the complex formation process of low-light images. Currently, the mainstream methods usually adopt domain adaptation with Teacher-Student architecture to solve the dark object detection problem, and they imitate the dark conditions by using non-learnable data augmentation strategies on the annotated source daytime images. Note that these methods neglected to model the intrinsic imaging process, i.e. image signal processing (ISP), which is important for camera sensors to generate low-light images. To solve the above problems, in this paper, we propose a novel method named ISP-Teacher for dark object detection by exploring Teacher-Student architecture from a new perspective (i.e. self-supervised learning based ISP degradation). Specifically, we first design a day-to-night transformation module that consistent with the ISP pipeline of the camera sensors (ISP-DTM) to make the augmented images look more in line with the natural low-light images captured by cameras, and the ISP-related parameters are learned in a self-supervised manner. Moreover, to avoid the conflict between the ISP degradation and detection tasks in a shared encoder, we propose a disentanglement regularization (DR) that minimizes the absolute value of cosine similarity to disentangle two tasks and push two gradients vectors as orthogonal as possible. Extensive experiments conducted on two benchmarks show the effectiveness of our method in dark object detection. In particular, ISP-Teacher achieves an improvement of +2.4% AP and +3.3% AP over the SOTA method on BDD100k and SHIFT datasets, respectively. The code can be found at https://github.com/zhangyin1996/ISP-Teacher.

Visible, NIR, SWIR, and eSWIR urban and rural scene contrast comparison.

The spectral information contained in the reflective imaging bands can be exploited for specific tasks. Whether targeting or mapping, the visible (VIS), near-infrared (NIR), shortwave infrared (SWIR), extended shortwave infrared (eSWIR) bands perform very differently for every application. For any imaging project, high contrast is very important for good imagery. High contrast leads to more recognizable features within a scene and easier identifiable objects. For mapping, good background scene contrast gives prominent features more detail and their locations can be easily identified. For targeting, low background scene contrast reduces clutter, making it easier to detect objects of interest. The VIS, NIR, SWIR, and eSWIR bands are popular reflective bands to design daytime imaging systems for either task. Deciding on which band will have the best contrast for a specific task is one of the first things to study when designing an imaging system. By measuring urban and rural scenes in terms of equivalent reflectivity (ER), a direct comparison of these four bands can show the utility they provide. The systems used to measure scene contrast are designed to have the same spatial resolution and field of view (FOV). With these instantaneous FOV (IFOV) matched systems, the variance and 1D power spectral densities (PSDs) provide a quantitative comparison for the contrast among the four bands. The ER differences and resulting contrast measured among these four bands show that the eSWIR has the highest contrast in both urban and rural scenes.

Daytime HyWFS approach for daylight adaptive optics wavefront sensing.

Bright daylight photon noise and the saturation of wavefront sensors pose challenges to high-resolution daytime imaging. In this paper, a daytime hybrid wavefront sensor (HyWFS) approach for real-time wavefront sensing in daylight adaptive optics (AO) is described. The Shack-Hartmann wavefront sensor (SHWFS) algorithm is used to efficiently compensate large-scale wavefronts, while the pyramid wavefront sensor (PyWFS) algorithm offers highly sensitive correction of small wavefronts. Daylight closed-loop AO experiments were performed using the daytime HyWFS approach with both algorithms, respectively. The experiment results indicate that the proposed approach provides accurate daylight AO correction and allows for a simple switch between the two algorithms without increasing system complexity. The daytime HyWFS approach can serve as an alternative for daylight natural guide star AO, enabling high-resolution observation of resident space objects no longer limited to dawn and dusk.

MONOCULAR DEPTH ESTIMATION FOR NIGHT-TIME IMAGES

Abstract. Depth estimation plays a pivotal role in numerous computer vision applications. However, depth estimation networks trained exclusively on daytime images tend to yield poor performance when applied to nighttime scenarios due to domain differences and variations in scene characteristics. In order to address this limitation, we conducted experiments involving the creation of a synthetic nighttime dataset by employing image translation techniques through a generative network. Subsequently, we utilized the generated images to fine-tune the depth estimation network, aiming to investigate the potential for enhancing task performance using generated data. We evaluated our approach by testing with the generated data, and we observed a noticeable improvement in the depth estimation task both before and after fine-tuning. Consequently, our approach yields results that are comparable to those achieved by networks specifically designed for daytime prediction. These findings highlight the effectiveness of utilizing synthetic data to enhance the performance of depth estimation tasks, particularly in nighttime settings.

Improving Nighttime Driving-Scene Segmentation via Dual Image-Adaptive Learnable Filters

Semantic segmentation on driving-scene images is vital for autonomous driving. Although encouraging performance has been achieved on daytime images, the performance on nighttime images are less satisfactory due to the insufficient exposure and the lack of labeled data. To address these issues, we present an add-on module called dual image-adaptive learnable filters (DIAL-Filters) to improve the semantic segmentation in nighttime driving conditions, aiming at exploiting the intrinsic features of driving-scene images under different illuminations. DIAL-Filters consist of two parts, including an image-adaptive processing module (IAPM) and a learnable guided filter (LGF). With DIAL-Filters, we design both unsupervised and supervised frameworks for nighttime driving-scene segmentation, which can be trained in an end-to-end manner. Specifically, the IAPM module consists of a small convolutional neural network with a set of differentiable image filters, where each image can be adaptively enhanced for better segmentation with respect to the different illuminations. The LGF is employed to enhance the output of segmentation network to get the final segmentation result. The DIAL-Filters are light-weight and efficient and they can be readily applied for both daytime and nighttime images. Our experiments show that DAIL-Filters can significantly improve the supervised segmentation performance on ACDC_Night and NightCity datasets, while it demonstrates the state-of-the-art performance on unsupervised nighttime semantic segmentation on Dark Zurich and Nighttime Driving testbeds. Codes and models are available at https://github.com/wenyyu/IA-Seg.

Developing a contextual model of poverty prediction using data science and analytics – The case of Shelby County

This study builds on the existing poverty literature and leverages data from disparate sources including both big data sources such as satellite images and traditional data sources such as the federal, state, and local agencies to develop a context-specific poverty prediction model using design science. We examine whether and to what extent infrastructure development as measured from the satellite images as well as spatial spillovers helps predict the poverty rate of a given census tract. We also develop and implement a Vector Autoregression (VAR) based ensemble model that combines predictions from daytime and nighttime imaging with adjacent tracts' poverty rates and other economic and demographic factors identified in prior literature. Our results show that daytime imaging and spatial network features have significant predictive value and that a combination of these features gives the best predictive power. In addition, we find that the skewness of poverty rates among adjacent census tracts, not the average, is a significant predictor showing the importance of distribution of poverty around a region. Our work has major implications for researchers using deep learning and network analysis for policy development and decision making.

MS-IRTNet: Multistage information interaction network for RGB-T semantic segmentation

The complementary information from RGB and thermal images can remarkably boost semantic segmentation performance. Existing RGB-T segmentation methods usually use simple interaction strategies to extract complementary information from RGB and thermal images, which ignores recognizability features from different imaging mechanisms. To address these problems, we propose a multistage information interaction network for RGB-T semantic segmentation called MS-IRTNet. MS-IRTNet has a dual-stream encoder structure that can extract multistage feature information. To better interact with multimodal information, we design a gate-weighted interaction module (GWIM) and a feature information interaction module (FIIM). GWIM can learn multimodal information weights in different channels, while FIIM integrates and fuses weighted RGB and thermal information into a single feature map. Finally, multistage interactive information is fed into the decoder for semantic prediction. Our method achieves 60.5 mIoU on the MFNet dataset, outperforming state-of-the-art methods. Notably, MS-IRTNet also achieved state-of-the-art results in tests of daytime images (51.7 mIoU) and nighttime images (62.5 mIoU). The code and pre-trained models are available at https://github.com/poisonzzw/MS-IRTNet.

Identifying the operational status of container terminals from high-resolution nighttime-light satellite image for global supply chain network optimization

Container terminals are cargo gateways in the global maritime supply chain network. Major container terminals generally operate throughout the year, but do not operate at night, when container vessels are not calling at ports, or when there is no need to handle containers. Terminal congestion can delay containers’ shipping schedules, which impacts the supply chain network. To optimize global logistics, it is therefore important to understand fully the daily operational status of container terminals. A vessels’ automatic identification system data are not sufficient to determine whether containers are being handled in container terminals at night. Remote sensing, especially nighttime-light (NTL) imagery, might solve this problem. Recently, high-resolution images for the CE-SAT-IIB satellite with a pixel resolution of 5.1 m became available to observe NTL. This study assessed the operational status of container terminals based on satellite image taken at night. Eight terminals in the Port of Tokyo, Japan, were selected for the study. A Sentinel-2A image recorded during the day on 7 April 2021, and a CE-SAT-IIB image recorded during the night on 6 April 2021, were obtained. The digital numbers (DNs) of each red-, green-, and blue-(RGB) band image were analyzed, revealing that the red, green, and blue bands, in that order, had higher DNs in the Sentinel-2A daytime image and the CE-SAT-IIB NTL image at all terminals. One of the eight terminals had a low DN in the CE-SAT-IIB RGB image because its lights were off at the time the image was taken. The operational status of the terminals could be verified from the CE-SAT-IIB image by setting the DN threshold to the green or red bands. We also found that the CE-SAT-IIB image could distinguish white-light-emitting diode (LED) lamps from high-pressure sodium lamps based on color differences in the DNs of the RGB bands. If high-resolution NTL sensors were placed onboard microsatellites, a high-frequency observation constellation network could be constructed using a combination of NTL data and daytime images. This study showed the benefits and usefulness of NTL images of ports; the results will contribute to the overall optimization of the global maritime supply chain network.

Nonlinear Deblurring for Low-Light Saturated Image.

Single image deblurring has achieved significant progress for natural daytime images. Saturation is a common phenomenon in blurry images, due to the low light conditions and long exposure times. However, conventional linear deblurring methods usually deal with natural blurry images well but result in severe ringing artifacts when recovering low-light saturated blurry images. To solve this problem, we formulate the saturation deblurring problem as a nonlinear model, in which all the saturated and unsaturated pixels are modeled adaptively. Specifically, we additionally introduce a nonlinear function to the convolution operator to accommodate the procedure of the saturation in the presence of the blurring. The proposed method has two advantages over previous methods. On the one hand, the proposed method achieves the same high quality of restoring the natural image as seen in conventional deblurring methods, while also reducing the estimation errors in saturated areas and suppressing ringing artifacts. On the other hand, compared with the recent saturated-based deblurring methods, the proposed method captures the formation of unsaturated and saturated degradations straightforwardly rather than with cumbersome and error-prone detection steps. Note that, this nonlinear degradation model can be naturally formulated into a maximum-a posterioriframework, and can be efficiently decoupled into several solvable sub-problems via the alternating direction method of multipliers (ADMM). Experimental results on both synthetic and real-world images demonstrate that the proposed deblurring algorithm outperforms the state-of-the-art low-light saturation-based deblurring methods.

Use of Satellite Images to Determine the Temperature of Urban Surfaces for Landscape Management Purposes, Case Study Bratislava (Slovak Republic)

This contribution deals with the use of data obtained from Landsat 8 satellite imaging to identify surface temperature variability in the example of the city of Bratislava, with an emphasis on identifying hotspots outside the built-up area, for example, on agricultural land—locations which are part of the European Network of Protected Areas. Surface temperature variability is presented in two time periods, on the daytime image taken on 26 July 2021 and on the nighttime image from 28 June 2021. Surface temperature is projected in a profile cut of the area. It vertically illustrates the temperatures of individual types of surfaces. Surfaces are classified by Urban Atlas classes. Areas reflecting the spatial distribution of the residential development in the city of Bratislava have been identified by satellite images in the studied area, and they represent a phenomenon of the urban heat island. Such areas were also identified outside the built-up area, in agricultural areas. The results of our research show that it is important to deal with UHI outside the built-up areas of cities and to orient the attention the territory planning and also to the proposal of measures for the management of these areas. Especially if these areas also include territories of the European system of protected areas, as it is in the case of Bratislava city (e.g., SPA029 Sysľovské polia). The results of reducing the impacts of climate change in cities concern not only the residents. In spatial planning, it is also necessary to address the management of non-built-up areas—localities with a quasi-natural character (e.g., areas with diverse vegetation cover). In order to recognize UHI within residential areas, it is essential to identify areas with significant differences between daytime and nighttime surface temperatures. Large differences between night and daytime surface temperatures can be seen in areas outside the built-up area in Bratislava on arable land where the difference is up to 8.0 °C (in the continuous housing class where the proportion of impermeable surfaces is higher than 80% with a temperature difference of 7.6 °C). Identification of overheated surfaces in the territory makes an important basis for modification of the landscape management and management of nature protection areas. It is important to propose measures related to the reduction in the negative impacts of climate change on the landscape and biodiversity.

A One-Stage Domain Adaptation Network With Image Alignment for Unsupervised Nighttime Semantic Segmentation.

In this paper, we tackle the problem of semantic segmentation for nighttime images that plays an equally important role as that for daytime images in autonomous driving, but is also much more challenging due to very poor illuminations and scarce annotated datasets. It can be treated as an unsupervised domain adaptation (UDA) problem, i.e., applying other labeled dataset taken in the daytime to guide the network training meanwhile reducing the domain shift, so that the trained model can generalize well to the desired domain of nighttime images. However, current general-purpose UDA approaches are insufficient to address the significant appearance difference between the day and night domains. To overcome such a large domain gap, we propose a novel domain adaptation network "DANIA" for nighttime semantic image segmentation by leveraging a labeled daytime dataset (the source domain) and an unlabeled dataset that contains coarsely aligned day-night image pairs (the target daytime and nighttime domains). These three domains are used to perform a multi-target adaptation via adversarial training in the network. Specifically, for the unlabeled day-night image pairs, we use the pixel-level predictions of static object categories on a daytime image as a pseudo supervision to segment its counterpart nighttime image. We also include a step of image alignment to relieve the inaccuracy caused by the misalignment between day-night image pairs by estimating a flow to refine the pseudo supervision produced by daytime images. Finally, a re-weighting strategy is applied to further improve the predictions, especially boosting the prediction accuracy of small objects. The proposed DANIA is a one-stage adaptation framework for nighttime semantic segmentation, which does not train additional day-night image transfer models as a separate pre-processing stage. Extensive experiments on Dark Zurich and Nighttime Driving datasets show that our DANIA achieves state-of-the-art performance for nighttime semantic segmentation.

Effective Denoising Architecture for Handling Multiple Noises

Object detection, one of the core research topics in computer vision, is extensively used in various industrial activities. Although there have been many studies of daytime images where objects can be easily detected, there is relatively little research on nighttime images. In the case of nighttime, various types of noises, such as darkness, haze, and light blur, deteriorate image quality. Thus, an appropriate process for removing noise must precede to improve object detection performance. Although there are many studies on removing individual noise, only a few studies handle multiple noises simultaneously. In this paper, we propose a convolutional denoising autoencoder (CDAE)-based architecture trained on various types of noises. We also present various composing modules for each noise to improve object detection performance for night images. Using the exclusively dark (ExDark) Image dataset, experimental results show that the Sequential filtering architecture showed superior mean average precision(mAP) compared to other architectures.

SPN2D-GAN: Semantic Prior Based Night-to-Day Image-to-Image Translation

Existing image-to-image translation approaches can deal with simple scenes or styles effectively, such as summer-to-winter, horses-to-zebra, and photo-to-map. Although a great progress has been made by GAN-based methods recently, the performance of night-to-day (N2D) translation remains unsatisfactory due to imbalanced/poor visibility, and thus leading to translation ambiguity. To improve the quality of N2D translation, we propose an unpaired translation scheme based on a semantic prior generator, namely SPN2D-GAN, in a weakly-supervised manner with consideration of both image and semantic information. Specifically, we design a novel N2D generator, which can adopt the semantic information of images as prior knowledge to generate more reasonable and realistic results. Also, we suggest adjusting the brightness of nighttime images to boost the visibility, so that the generator can better extract content information. Moreover, the proposed SPN2D-GAN translates images by enforcing the distribution of daytime images in both image and semantic domains on final outputs. Besides, the cycle consistency is employed to preserve the fidelity between translations from two directions. Extensive experimental results are provided to reveal the effectiveness of our design, and demonstrate its superior performance over other state-of-the-art N2D translation approaches both quantitatively and qualitatively. Our code is available at <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/brightknight177/SPN2DGAN</uri> .