Image Correction Research Articles

Highly robust thermal infrared and visible image registration with canny and phase congruence detection

When a dual-optical vehicle-mounted camera is photographed, visible and thermal infrared (VI-T) images cannot be registered, and realising fine matching of the images to solve for the optimal affine parameters is thus key to this research. Visible and thermal image matching is of great interest owing to the presence of challenging scale transformations, nonlinear radial aberrations, and robustness under illumination conditions. Thus, this paper presents a novel algorithm, namely, Highly Robust Thermal Infrared and Visible Image Registration with Canny and Phase Congruence Detection (HRCP). Firstly, the image is pre-processed, and feature point detection is performed by combining adaptive thresholding GW-Canny edge detection and phase coherence (PC), which increases the edge information. Secondly, SURF and FAST feature detection at the maximum distance of PC maps can obtain stable and reliable feature points, which are robust to scale invariance and rotation invariance. Finally, the optimal affine parameter is solved using the least square method with several iterations. Compared to existing matching methods, we find that VI-T image matching is particularly sensitive to both lighting factors and scale-transformed nonlinear radial distortion and that HRCP is robust in performing matching. For the randomly selected dataset, the average image matching time is reduced by 3.265 s, the number of correct image matches increases by 44 pairs on average, the root mean square error improves to 1.830, and the average error improves to 2.815. The proposed method not only improves the inaccuracy of radiation-insensitive feature transform matching in both low and high light, but also achieves a higher accuracy and an increase in the number of feature matching points, which facilitates robust VI-T image registration.

Identification of Individuals of Two Takin Subspecies Using Biological and Ecological Criteria in Eastern Himalayas of China.

Limited background data are available on the Mishmi takin (Budorcas taxicolor taxicolor) and Bhutan takin (Budorcas taxicolor whitei) subspecies in the Eastern Himalayas of China because of the lack of systematic field investigations and research. Therefore, mature-animal ecological methods were used to evaluate these takin subspecies' phenotypic characteristics, distribution range, activity rhythm, and population size. From 2013 to 2022, 214 camera traps were installed for wild ungulate monitoring and investigation in all human-accessible areas of the Eastern Himalayas, resulting in 4837 distinguishable takin photographs. The external morphological characteristics were described and compared using visual data. Artificial image correction and related technologies were used to establish physical image models based on the differences between subspecies. MaxEnt niche and random encounter models obtained distribution ranges and population densities. Mishmi takins have a distribution area of 17,314 km2, population density of 0.1729 ± 0.0134 takins/km2, and population size of 2995 ± 232. Bhutan takins have a distribution area of 25,006 km2, population density of 0.1359 ± 0.0264 takins/km2, and population size of 3398 ± 660. Long-term monitoring data confirmed that the vertical migration within the mountain ecosystems is influenced by climate. Mishmi takins are active at 500-4500 m, whereas Bhutan takins are active at 1500-4500 m. The two subspecies were active at >3500 m from May to October yearly (rainy season). In addition, surveying combined with model simulation shows that the Yarlung Zangbo River is not an obstacle to migration. This study provides basic data that contribute to animal diversity knowledge in biodiversity hotspots of the Eastern Himalayas and detailed information and references for species identification, distribution range, and population characteristics.

Contrast-Enhanced Sonography of the Liver: How to Avoid Artifacts.

Contrast-enhanced sonography (CEUS) is a very important diagnostic imaging tool in clinical settings. However, it is associated with possible artifacts, such as B-mode US-related artifacts. Sufficient knowledge of US physics and these artifacts is indispensable to avoid the misinterpretation of CEUS images. This review aims to explain the basic physics of CEUS and the associated artifacts and to provide some examples to avoid them. This review includes problems related to the frame rate, scanning modes, and various artifacts encountered in daily CEUS examinations. Artifacts in CEUS can be divided into two groups: (1) B-mode US-related artifacts, which form the background of the CEUS image, and (2) artifacts that are specifically related to the CEUS method. The former includes refraction, reflection, reverberation (multiple reflections), attenuation, mirror image, and range-ambiguity artifacts. In the former case, the knowledge of B-mode US is sufficient to read the displayed artifactual image. Thus, in this group, the most useful artifact avoidance strategy is to use the reference B-mode image, which allows for a simultaneous comparison between the CEUS and B-mode images. In the latter case, CEUS-specific artifacts include microbubble destruction artifacts, prolonged heterogeneous accumulation artifacts, and CEUS-related posterior echo enhancement; these require an understanding of the mechanism of their appearance in CEUS images for correct image interpretation. Thus, in this group, the most useful artifact avoidance strategy is to confirm the phenomenon's instability by changing the examination conditions, including the frequency, depth, and other parameters.

Deep learning-based perspective distortion correction for outdoor photovoltaic module images

To address the growing demand for photovoltaic inspection, an increasing volume of photoluminescence, electroluminescence, and infrared image datasets is being produced for research and analysis. Most of these images exhibit perspective distortions, which introduce challenges in subsequent analysis, such as defect detection and classification. Hence, distortion correction is required during the preprocessing stage. However, manually correcting each dataset to eliminate distortions is labour-intensive. This paper addresses this issue by developing a convolutional neural network model to estimate the camera parameters of yaw, pitch, and roll. A heuristic method was referenced as a comparative benchmark, demonstrating that the developed deep learning-based approach is equally accurate and significantly faster. The findings underscore the potential of deep learning techniques in automating and refining image analysis in photovoltaic diagnostics, offering substantial improvements over traditional methods in terms of speed and scalability.

Atmospheric pollutants in Rosario, Argentina analysed through remote sensing: Wildfires and COVID-19 lockdown effects

Spatial-temporal dynamics of atmospheric pollutants can be analysed by space-based observations, contributing to environmental management and public health interventions. The influence of wildfires and anthropogenic activities on air quality is studied for the city of Rosario and its surroundings, including both urban and non-urban areas. Utilizing advanced satellite-based observations, we assess Aerosol Optical Depth (AOD) and Nitrogen Dioxide (NO2). The study employs Multi-Angle Implementation of Atmospheric Correction (MAIAC) and Moderate Resolution Imaging Spectroradiometer (MODIS) onboard NASA's Terra and Aqua satellites for AOD analysis. NO2 measurements were derived from the Ozone Monitoring Instrument (OMI) on the Aura satellite. Backward trajectory analysis using Hybrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) shows the connection between fire pixels and air masses reaching Rosario. During the COVID-19 lockdown, a period with no significant fire events in the studied region was selected. This provided, for the first time, a baseline on the AOD median value of 0.05 for Rosario city and its surroundings, while NO2 total column median values were close to 3.00 x 1015 molecules/cm2. In a business-as-usual scenario, AOD increased by approximately 52.8% (March 2022) up to 111.3% (March 2019). NO2 median values remain almost the same (March 2019) or reached a median value of 3.38 x 1015 molecules/cm2 (March 2022). During wildfire events, such as March 2023, AOD surged by around 50.9%–108.6% compared to the business-as-usual scenario (March 2019 and March 2022, respectively). NO2 median values ranged from 3.52 x 1015 molecules/cm2 (March 2023) to 3.66 x 1015 molecules/cm (March 01, 2020 to March 19, 2020). NO2 levels correlated with intense fire periods. The analysis provides valuable insights into the complex interplay between natural events, human activities, and air quality dynamics in the region.

A Complex Environmental Water-Level Detection Method Based on Improved YOLOv5m.

The existing methods for water-level recognition often suffer from inaccurate readings in complex environments, which limits their practicality and reliability. In this paper, we propose a novel approach that combines an improved version of the YOLOv5m model with contextual knowledge for water-level identification. We employ the adaptive threshold Canny operator and Hough transform for skew detection and correction of water-level images. The improved YOLOv5m model is employed to extract the water-level gauge from the input image, followed by refinement of the segmentation results using contextual priors. Additionally, we utilize a linear regression model to predict the water-level value based on the pixel height of the water-level gauge. Extensive experiments conducted in real-world environments encompassing daytime, nighttime, occlusion, and lighting variations demonstrate that our proposed method achieves an average error of less than 2 cm.

A thermal infrared imaging observation system for the measurement of overland flow velocities under controlled laboratory conditions

Flow velocity is an important physical parameter in characterizing the hydraulics of water flow. The accurate measurement of overland flow velocities has great significance for understanding and modeling sediment transport and soil erosion. In this study, a Thermal Infrared Imaging Observation System (TIIOS) based on thermal infrared imaging techniques and computer vision recognition technology was established to measure overland flow velocities under controlled laboratory conditions. TIIOS has three subsystems including thermal tracer control, image acquisition and transmission, and image storage and processing. It can be used to dynamically monitor changes in shallow flow velocity by means of automatic control of thermal tracer, instantaneous image acquisition, image correction, noise removal and centroid determination. The observation precision was high with a standard deviation of 0.004 m·s−1 for flow velocity, a temporal resolution of 1/9th s, and a spatial resolution of 2 mm. The new system achieved higher accuracy compared to the traditional dye tracing and salt tracing techniques in observing the dynamic transport process of thermal tracer movement. The relative errors of the centroid velocity measured by the TIIOS ranged from −9.83 % to 6.40 %; whereas the relative errors of centroid velocity measured by salt tracer ranged from −21.36 % to 17.22 %. Measurements by the established TIIOS under different hydraulic conditions demonstrate that the relative errors of the centroid velocity are all smaller than those of the leading-edge velocity. This observational system provided a reliable way to measure shallow flow velocity for better understanding the mechanisms of water erosion processes and showed its potential to be used in field experiments under natural conditions.

A study of somatization symptoms and low-frequency amplitude fluctuations of emotional memory in adolescent depression

Studies have revealed that somatization symptoms are associated with emotional memory in adolescents with depressive disorders. This study investigated somatization symptoms and emotional memory among adolescents with depressive disorders using low-frequency amplitude fluctuations (ALFF). Participants were categorized into the somatization symptoms (FSS) group, non-FSS group and healthy control group (HC). The correctness of negative picture re-recognition was higher in the FFS and HC group than in the non-FSS group. The right superior occipital gyrus and right inferior temporal gyrus were significantly larger in the FSS group than those in the non-FSS and HC groups. Additionally, the ALFF in the superior occipital and inferior temporal gyrus were positively correlated with CSI score. Furthermore, the ALFF values in the temporal region positively correlated with correct negative image re-recognition. The negative image re-recognition rate was positively correlated with the ALFF in the left and right middle occipital gyri. These findings indicated that somatization symptoms in adolescent depression are associated with the superior occipital gyrus and inferior temporal gyrus. Notably, somatization symptoms play a role in memory bias within depressive disorders, with middle occipital and inferior temporal gyri potentially serving as significant brain regions.

Microstructural Characterization of Eutectics using Digital Image Analysis

Metallic eutectics play an important role in casting technology and properties. For this reason, the study of eutectics microstructure is indispensable in the casting qualification. Eutectics have many similar characteristics, including the morphology, size and spatial arrangement of eutectic phases. This makes it possible to develop a method of general use based on analyzing eutectic microscopy images. The method presented in this article performs a posteriori background correction for OM images. The shape and size of phases are determined using cellular automata and machine learning. Another cellular automaton and cluster analysis characterizes the spatial arrangement of eutectic phases. It can also be used to determine the distance between objects both locally and within a given object group. The algorithm is suitable for exploring and examining the spatial clustering of objects. The methods can be included in an algorithm, so a detailed examination of the eutectic microstructure can be carried out. The method was tested on micrographs of Al-Cu, Al-Ni, Al-Si and cast irons.

Correcting Optical Aberration via Depth-Aware Point Spread Functions.

Optical aberration is a ubiquitous degeneration in realistic lens-based imaging systems. Optical aberrations are caused by the differences in the optical path length when light travels through different regions of the camera lens with different incident angles. The blur and chromatic aberrations manifest significant discrepancies when the optical system changes. This work designs a transferable and effective image simulation system of simple lenses via multi-wavelength, depth-aware, spatially-variant four-dimensional point spread functions (4D-PSFs) estimation by changing a small amount of lens-dependent parameters. The image simulation system can alleviate the overhead of dataset collecting and exploiting the principle of computational imaging for effective optical aberration correction. With the guidance of domain knowledge about the image formation model provided by the 4D-PSFs, we establish a multi-scale optical aberration correction network for degraded image reconstruction, which consists of a scene depth estimation branch and an image restoration branch. Specifically, we propose to predict adaptive filters with the depth-aware PSFs and carry out dynamic convolutions, which facilitate the model's generalization in various scenes. We also employ convolution and self-attention mechanisms for global and local feature extraction and realize a spatially-variant restoration. The multi-scale feature extraction complements the features across different scales and provides fine details and contextual features. Extensive experiments demonstrate that our proposed algorithm performs favorably against state-of-the-art restoration methods.

Anti-noise performance analysis in amplitude-modulated collinear holographic data storage using deep learning

In an amplitude-modulated collinear holographic data storage system, optical system aberration and experimental noise due to the recording medium often result in a high bit error rate (BER) and low signal-to-noise ratio (SNR) in directly read detector data. This study proposes an anti-noise performance analysis using deep learning. End-to-end convolutional neural networks were employed to analyze noise resistance in encoded data pages captured by the detector. Experimental results demonstrate that these networks effectively correct system imaging aberrations, detector light intensity response, holographic storage medium response non-uniformity, and defocusing noise from the recording objective lens. Consequently, the BER of reconstructed encoded data pages can be reduced to 1/10 of that from direct detection, while the SNR can be increased more than fivefold, enhancing the accuracy and reliability of data reading in amplitude holographic data storage systems.

AutoCorNN: An Unsupervised Physics-Aware Deep Learning Model for Geometric Distortion Correction of Brain MRI Images Towards MR-Only Stereotactic Radiosurgery.

Geometric distortions in brain MRI images arising from susceptibility artifacts at air-tissue interfaces pose a significant challenge for high-precision radiation therapy modalities like stereotactic radiosurgery, necessitating sub-millimeter accuracy. To achieve this goal, we developed AutoCorNN, an unsupervised physics-aware deep-learning model for correcting geometric distortions. Two publicly available datasets, the MPI-Leipzig Mind-Brain-Body with 318 subjects, and the Vestibular Schwannoma-SEG dataset, encompassing 242 patients were utilized. AutoCorNN integrates two 2D convolutional encoder-decoder neural networks with the forward physical model of MRI signal generation to predict undistorted MR and field map images from distorted MR input. The network is trained in an unsupervised manner by minimizing the mean absolute error between the measured and estimated k-space data, without requiring ground truth images during training or deployment. The model was evaluated on vestibular schwannoma cases. AutoCorNN achieved a peak signal-to-noise ratio (PSNR) of 41.35 ± 0.02dB, a root mean square error (RMSE) of 0.02 ± 0.003, and a structural similarity index (SSIM) of 0.99 ± 0.02 outperforming uncorrected and B0-mapping correction methods. Geometric distortions of about 1.6mm were observed at the air-tissue interfaces at the air canal and nasal cavity borders. Geometrically, distortion correction increased the target volume from 3.12 ± 0.52cc to 3.84 ± 0.54cc. Dosimetrically, AutoCorNN improved target coverage (0.96 ± 0.01 to 0.97 ± 0.02), conformity index (0.92 ± 0.03 to 0.94 ± 0.03), and reduced dose gradients outside the target. AutoCorNN achieves accurate geometric distortion correction comparable to conventional iterative methods while offering substantial computational acceleration, enabling precise target delineation and conformal dose delivery for improved radiation therapy outcomes.

Computer Vision Algorithm Design in Image Processing Based on Projective Geometry

Image processing with computer vision, particularly in the realm of projective geometry, offers remarkable potential for various applications. Through the lens of projective geometry, images can be transformed, augmented, and reconstructed with precision, facilitating tasks such as image rectification, 3D reconstruction, and object tracking. Landmark estimation in computer vision is a vital task with broad applications across various domains. This process involves identifying key points or landmarks within images, enabling tasks such as facial recognition, object tracking, and gesture recognition. This paper, proposed a novel approach for landmark estimation in computer vision using Projective Geometry Landmark Estimation (PGLM). The proposed model aims to estimate the landmark features by a projective geometry model. With the estimation of the geometry features landmarks related to the facial, object, and medical images are computed. The PGLM model uses the point features for the location of the landmark features. In order to compare PGLM's performance to that of more conventional classification methods like Random Forest, K-Nearest Neighbors (KNN), and Support Vector Machine (SVM), simulation analysis is carried out. From what we can see, PGLM routinely beats these alternatives when we compare their accuracy, precision, recall, and F1 score. The findings stated the effectiveness of PGLM as a promising approach for landmark estimation in image processing tasks, paving the way for further advancements in this domain.

Optimization of handheld spectrally encoded coherence tomography and reflectometry for point-of-care ophthalmic diagnostic imaging.

Handheld optical coherence tomography (HH-OCT) systems enable point-of-care ophthalmic imaging in bedridden, uncooperative, and pediatric patients. Handheld spectrally encoded coherence tomography and reflectometry (HH-SECTR) combines OCT and spectrally encoded reflectometry (SER) to address critical clinical challenges in HH-OCT imaging with real-time en face retinal aiming for OCT volume alignment and volumetric correction of motion artifacts that occur during HH-OCT imaging. We aim to enable robust clinical translation of HH-SECTR and improve clinical ergonomics during point-of-care OCT imaging for ophthalmic diagnostics. HH-SECTR is redesigned with (1)optimized SER optical imaging for en face retinal aiming and retinal tracking for motion correction, (2)a modular aluminum form factor for sustained alignment and probe stability for longitudinal clinical studies, and (3)one-handed photographer-ergonomic motorized focus adjustment. We demonstrate an HH-SECTR imaging probe with micron-scale optical-optomechanical stability and use it for in vivo human retinal imaging and volumetric motion correction. This research will benefit the clinical translation of HH-SECTR for point-of-care ophthalmic diagnostics.

Benthic habitat mapping for estimating seagrass carbon stock across Takabonerate Islands, Indonesia

Data on benthic habitat cover are essential for monitoring and managing coastal areas, providing valuable ecosystem services for local communities. Moreover, seagrass extent is crucial for estimating carbon stock using Tier-1 calculations. On the other side, remote sensing technology has proven effective in mapping benthic habitat cover, including seagrass, over large local areas. Therefore, this research aims to map specific benthic habitats in a local area to obtain high-accuracy estimates of seagrass extent for carbon stock estimation. The study was conducted in the Takabonerate Islands, Indonesia, using Sentinel-2A imagery. The study involved field data processing, which enabled the mapping of eight benthic habitat classes using Sentinel-2A images, along with the steps for Sentinel-2A image processing for benthic habitat mapping. The benthic habitat map was produced following image correction steps, including atmospheric, sunglint, and water column corrections, followed by image classification using a Neural Net Classifier (NNC). Seagrass carbon stock estimation was performed using the Tier-1 equation from IPCC 2013. The study revealed a total benthic habitat area of 57,538.04 ha, with a seagrass area of 3127.42 ha for carbon stock estimation. The potential seagrass carbon stock in the study area was estimated to range from 337.76 GgC/ha to 31.27 GgC/ha using the Tier-1 equation. The overall accuracy of the benthic habitat map was 80.5 %, calculated using a confusion matrix table. The spatial information from this study can be used by local coastal managers and governments for baseline and monitoring purposes. The key interpretations from this study provide valuable findings that can be adapted and improved for future local mapping purposes.

Deep-learning-based Attenuation Correction for 68Ga-DOTATATE Whole-body PET Imaging: A Dual-center Clinical Study.

Attenuation correction is a critical phenomenon in quantitative positron emission tomography (PET) imaging with its own special challenges. However, computed tomography (CT) modality which is used for attenuation correction and anatomical localization increases patient radiation dose. This study was aimed to develop a deep learning model for attenuation correction of whole-body 68Ga-DOTATATE PET images. Non-attenuation-corrected and computed tomography-based attenuation-corrected (CTAC) whole-body 68Ga-DOTATATE PET images of 118 patients from two different imaging centers were used. We implemented a residual deep learning model using the NiftyNet framework. The model was trained four times and evaluated six times using the test data from the centers. The quality of the synthesized PET images was compared with the PET-CTAC images using different evaluation metrics, including the peak signal-to-noise ratio (PSNR), structural similarity index (SSIM), mean square error (MSE), and root mean square error (RMSE). Quantitative analysis of four network training sessions and six evaluations revealed the highest and lowest PSNR values as (52.86±6.6) and (47.96±5.09), respectively. Similarly, the highest and lowest SSIM values were obtained (0.99±0.003) and (0.97±0.01), respectively. Additionally, the highest and lowest RMSE and MSE values fell within the ranges of (0.0117±0.003), (0.0015±0.000103), and (0.01072±0.002), (0.000121±5.07xe-5), respectively. The study found that using datasets from the same center resulted in the highest PSNR, while using datasets from different centers led to lower PSNR and SSIM values. In addition, scenarios involving datasets from both centers achieved the best SSIM and the lowest MSE and RMSE. Acceptable accuracy of attenuation correction on 68Ga-DOTATATE PET images using a deep learning model could potentially eliminate the need for additional X-ray imaging modalities, thereby imposing a high radiation dose on the patient.

Correction Compensation and Adaptive Cost Aggregation for Deep Laparoscopic Stereo Matching

Perception of digitized depth is a prerequisite for enabling the intelligence of three-dimensional (3D) laparoscopic systems. In this context, stereo matching of laparoscopic stereoscopic images presents a promising solution. However, the current research in this field still faces challenges. First, the acquisition of accurate depth labels in a laparoscopic environment proves to be a difficult task. Second, errors in the correction of laparoscopic images are prevalent. Finally, laparoscopic image registration suffers from ill-posed regions such as specular highlights and textureless areas. In this paper, we make significant contributions by developing (1) a correction compensation module to overcome correction errors; (2) an adaptive cost aggregation module to improve prediction performance in ill-posed regions; (3) a novel self-supervised stereo matching framework based on these two modules. Specifically, our framework rectifies features and images based on learned pixel offsets, and performs differentiated aggregation on cost volumes based on their value. The experimental results demonstrate the effectiveness of the proposed modules. On the SCARED dataset, our model reduces the mean depth error by 12.6% compared to the baseline model and outperforms the state-of-the-art unsupervised methods and well-generalized models.

Deep underwater image compression for enhanced machine vision applications

Underwater image compression is fundamental in underwater visual applications. The storage resources of autonomous underwater vehicles (AUVs) and underwater cameras are limited. By employing effective image compression methods, it is possible to optimize the resource utilization of these devices, thereby extending the operational time underwater. Current image compression methods neglect the unique characteristics of the underwater environment, thus failing to support downstream underwater visual tasks efficiently. We propose a novel underwater image compression framework that integrates frequency priors and feature decomposition fusion in response to these challenges. Our framework incorporates a task-driven feature decomposition fusion module (FDFM). This module enables the network to understand and preserve machine-friendly information during the compression process, prioritizing task relevance over human visual perception. Additionally, we propose a frequency-guided underwater image correction module (UICM) to address noise issues and accurately identify redundant information, enhancing the overall compression process. Our framework effectively preserves machine-friendly features at a low bit rate. Extensive experiments across various downstream visual tasks, including object detection, semantic segmentation, and saliency detection, consistently demonstrated significant improvements achieved by our approach.

Illumination correction for dyed products with spotted hyena optimizer and regularized extreme learning machine

Dyed products exhibit varying colors when illuminated by different light sources. Human eyes possess color constancy, enabling them to automatically adjust and overlook color discrepancies, effectively self-correcting visually. However, these variations are perceived as distinct colors by computers. To eliminate the effects of lighting on the dyed products, illumination correction is needed to make the dyed products show their original color. To achieve this goal, this article presents a novel algorithm, WOA-SHO-RELW (Whale Optimization Algorithm-Spotted Hyena Optimizer-Regularized Extreme Learning Machine), designed to emulate color constancy and applies it to the illumination correction of dyed products. We employ the WOA to optimize the initial population of SHO, carefully selecting a group of suitable initial parameters. Subsequently, SHO is utilized to optimize the hidden biases and input weights of RELM, with the overarching aim of enhancing the model’s effectiveness and performance, thereby ensuring more precise illumination correction for images of dyed products. The comparison results with other algorithms show that our WOA-SHO-RELM has better comprehensive performances.

Enhancements in image quality and block detection performance for Reinforced Soil-Retaining Walls under various illuminance conditions

To ensure continuous monitoring of reinforced soil-retaining walls (RSWs) even under low-illuminance conditions, such as during the night, it is imperative to evaluate the performance of deep learning-based detection. In this study, we constructed a laboratory RSW model and generated a dataset with varying illuminance levels to assess the impact of image enhancement and block detection. Various image enhancement methods were applied to improve image quality and evaluate their effect on deep learning. RGB optimization (RO) was proposed to optimize RGB intensity and compared with gamma correction, histogram equalization, and low-light image enhancement with illumination map estimation. RO demonstrated outstanding image enhancement performance, as evidenced by lightness order error, peak signal-to-noise ratio, and structural similarity index measure, ensuring high image quality. The trained RO model using Mask R-CNN exhibited excellent accuracy, recall, and F1 score, delivering remarkable detection performance under low illuminance conditions, resulting in a 7.44 % improvement in the F1 score. Image enhancement techniques that maintain similarity, such as lightness order error and structural similarity, across varying illuminance conditions contribute to enhancing the block detection performance of Mask R-CNNs.