Two-dimensional layered CuInP2S6 (CIPS) combines robust room-temperature-switchable ferroelectric polarization with precise control over polarization states, allowing for multiple neuromorphic devices. In this work, we present a band-engineerable ferroelectric CIPS-WS2/MoS2 heterostructure, which features polarization-regulated adaptive visual contrast across dynamic illumination conditions. Under bright-light conditions, negative polarization of CIPS induces a bidirectional (ultraviolet) UV/visible response. The UV component triggers negative photoresponse, suppressing background current and thus enhancing contrast. Under low-light conditions, positive polarization of CIPS minimizes dark current, preserving high contrast despite weak signals. Under both conditions, the contrast can be further amplified via synaptic potentiation of the CIPS layer. The device can emulate the UV/vis vision of bees, improving the flower stamen detection rate. In the proposed autonomous driving scenarios, it achieves a high target recognition accuracy under both bright and dim conditions, outperforming conventional static sensors. These results highlight the device's potential for reliable operation in extreme environments. This work demonstrates a hardware-level adaptive vision strategy with broad potential for autonomous systems, robotics, and intelligent sensing applications.

DeC-YOLO: Dynamic illumination compensation and context-transformer fusion for low-light object detection

Abstract Recent advancements in 3D Gaussian Splatting (3DGS) have made significant improvements in real‐time novel view synthesis and 3D reconstruction. 3DGS has seen significant development in driving scenarios, but existing methods are mainly designed for videos captured by autonomous vehicles. It is not suitable for the complexity and dynamic lighting challenges present in dash cam videos. Despite the progress made by previous work in dealing with reflections and occlusions, the distribution of 3D Gaussian points remains inaccurate due to the inherent complexity of dash cam scenes, causing geometric distortions in the rendered output. Additionally, uncontrolled dynamic illumination exacerbates Gaussian point density anomalies and local geometric distortions. These challenges significantly hinder the development of scene reconstruction techniques based on dash cam videos. To address these challenges, we present RDC‐GS, an innovative method featuring a point correction mechanism to eliminate distribution errors of Gaussian points during training, and a brightness‐aware illumination technique to enhance detailed representation under dynamic lighting conditions. This approach yields more robust scene reconstruction. Experiments conducted on real dash cam videos demonstrate that our method achieves a 1.5‐dB PSNR improvement over current state‐of‐the‐art techniques. Comprehensive experiments validate the efficacy of our approach across challenging scenarios.

ABSTRACT This study presents a novel LIDet-Net (low-light illumination defect detection network) framework for weld defect detection in low-light industrial environments. To address challenges of insufficient illumination, blurred defect boundaries, and small-target detection, the framework integrates a dynamic illumination enhancer Retinex-WeldNet with a multi-scale detector RE-YOLO. Retinex-WeldNet enhances brightness and suppresses noise based on the Retinex theory, while RE-YOLO employs a re-calibrated feature pyramid network (RC-FPN) to sharpen the boundary representation and improve small-target detection. Experimental results demonstrate that LIDet-Net achieves a mean average precision (mAP) of 91.1% on the low-light weld (LL-WELD) dataset, outperforming the baseline YOLOv11 model by 8.3%. Moreover, the framework demonstrates strong robustness and generalisation across steel and PCBs defect datasets, proving to be an efficient and accurate solution for industrial inspection under variable lighting.

Future lunar south pole missions face dual challenges of highly variable illumination and rugged terrain that directly constrain rover mobility and energy sustainability. To address these issues, this study proposes a dynamic illumination-constrained spatio-temporal A* (DIC3D-A*) path-planning algorithm that jointly optimizes terrain safety and illumination continuity in polar environments. Using high-resolution digital elevation model data from the Lunar Reconnaissance Orbiter Laser Altimeter, a 1300 m × 1300 m terrain model with 5 m/pixel spatial resolution was constructed. Hourly solar visibility for November–December 2026 was computed based on planetary ephemerides to generate a dynamic illumination dataset. The algorithm integrates slope, distance, and illumination into a unified heuristic cost function, performing a time-dependent search in a 3D spatiotemporal state space. Simulation results show that, compared with conventional A* algorithms considering only terrain or distance, the DIC3D-A* algorithm improves CSDV by 106.1% and 115.1%, respectively. Moreover, relative to illumination-based A* algorithms, it reduces the average terrain roughness index by 17.2%, while achieving shorter path length and faster computation than both the Rapidly-exploring Random Tree Star and Deep Q-Network baselines. These results demonstrate that dynamic illumination is the dominant environmental factor affecting lunar polar rover traversal and that DIC3D-A* provides an efficient, energy-aware framework for illumination-adaptive navigation in upcoming missions such as Chang’E-7.

Conventional High Dynamic Range (HDR) image fusion algorithms generally require two or more original images with different exposure times for synthesis, making them unsuitable for real-time processing scenarios such as video streams. Additionally, the synthesized HDR images have the same bit depth as the original images, which may lead to banding artifacts and limits their applicability in professional fields requiring high fidelity. This paper utilizes a Field Programmable Gate Array (FPGA) to support an image sensor operating in Clear HDR mode, which simultaneously outputs High Conversion Gain (HCG) and Low Conversion Gain (LCG) images. These two images share the same exposure duration and are captured at the same moment, making them well-suited for real-time HDR fusion. This approach provides a feasible solution for real-time processing of video streams. An adaptive adjustment algorithm is employed to address the requirement for high fidelity. First, the initial HCG and LCG images are fused under the initial fusion parameters to generate a preliminary HDR image. Subsequently, the gain of the high-gain images in the video stream is adaptively adjusted according to the brightness of the fused HDR image, enabling stable brightness under dynamic illumination conditions. Finally, by evaluating the read noise of the HCG and LCG images, the fusion parameters are adaptively optimized to synthesize an HDR image with higher bit depth. Experimental results demonstrate that the proposed method achieves a processing rate of 46 frames per second for 2688 × 1520 resolution video streams, enabling real-time processing. The bit depth of the image is enhanced from 12 bits to 16 bits, preserving more scene information and effectively addressing banding artifacts in HDR images. This improvement provides greater flexibility for subsequent image processing tasks. Consequently, the adaptive algorithm is particularly suitable for dynamically changing scenarios such as real-time surveillance and professional applications including industrial inspection.

This paper presents a human–machine interaction system that utilizes augmented reality (AR) as an immersive gaming environment, enabling users to interact seamlessly with virtual elements through hand gestures. Two powerful platforms (OpenCV and OpenGL) are employed for video processing, image analysis, and real-time 3D rendering. Unlike existing ArUco-based pipelines, the proposed multi-layer filtering strategy ensures reliable marker detection even under reflective surfaces and dynamic illumination, significantly enhancing system robustness. The AR framework further employs an inexpensive webcam instead of costly RGBD sensors, generating depth information mathematically through pose estimation. This approach not only reduces computational complexity but also achieves high precision in marker detection and identification, outperforming the default OpenCV ArUco library in both speed and accuracy. The system runs efficiently on low-end CPUs equipped with GPUs across multiple video resolutions without compromising performance, providing a cost-effective and accessible solution for realistic AR-based gaming and interaction.

Accurate crowd counting has become increasingly essential for public safety management and Internet of Video Things (IoVT) applications, driven by rapid population growth and urbanization. However, RGB-Thermal (RGB-T) crowd counting remains challenging due to poor recognition of small targets and degraded performance under extreme conditions such as low-light environments. To address these issues, we propose a bidirectional-guided network with scene-semantics driven fusion for RGB-T crowd counting (BigCounter) that enhances robustness and generalization in complex scenes. BigCounter comprises of three parallel branches: a primary branch, a dynamic illumination auxiliary enhancement branch (DIAEB), and a high-resolution auxiliary enhancement branch (HAEB), which respectively improve robustness under illumination variations and accuracy for small target detection. Moreover, a cross-layer scene-driven fusion module (CLSFM) and a cross-modal semantic-driven fusion module (CMSFM) are designed to strengthen structural consistency and explore semantic complementarity between modalities. Through multi-branch collaboration and semantic-aware fusion, BigCounter significantly enhances feature representation. Extensive experiments on two benchmark RGB-T datasets demonstrate that BigCounter achieves superior accuracy and generalization compared with state-of-the-art methods.

To address the issue of link stability in vehicle-to-everything visible light communication (VLC) systems being susceptible to dynamic environmental factors in complex road conditions, which directly reduce the effective communication range, this paper proposes a hybrid receiver architecture based on a linear photodiode (PD) and a single-photon avalanche diode (SPAD) array. The architecture builds a dynamic optical scheduling core, i.e., the hybrid controller, by means of a micro-electro-optical selective switch, which achieves adaptive switching of the incident signal between the linear receiving path of the PIN module and the photon detection path of the SPAD array module. Our hybrid PD/SPAD receiver advances beyond prior indoor-focused designs by addressing outdoor vehicular challenges: weather attenuation, turbulence, pointing errors, and dynamic illumination. The resulting adaptive path selection ensures stability and minimizes BER across all operating conditions. Simulations at a target BER=10-3 show that under low-light fog (visibility V=0.5km), the hybrid extends reliable range from 60 (PIN-only) to 157 m (+162%); under sunny conditions with 20 mW background illumination, it achieves 29 versus 16 m for SPAD-only (+81%); and under cloudy conditions (V=5km) it maintains >50Mb/s at 200 m while reducing outage probability by 85% relative to the best single-path baseline. By automatically selecting the optimal detection path as channel conditions vary, the proposed architecture enhances link reliability, extends operating range, and improves availability for robust V2V-VLC in complex road environments.

The three-dimensional documentation of hypogean structures poses significant methodological challenges due to the absence of natural light, confined spaces, and the presence of fragile painted surfaces. This study presents an integrated workflow for the survey of the Tomba dell’Orco (Tarquinia), combining terrestrial laser scanning, photogrammetry, and the light painting technique. Borrowed from photographic practice, light painting was employed as a dynamic lighting strategy during photogrammetric acquisition to overcome issues of uneven illumination and harsh shadows typical of underground environments. By moving handheld LED sources throughout long-exposure shots, operators produced evenly illuminated images suitable for feature extraction and high-resolution texture generation. These image datasets were subsequently integrated with laser scanning point clouds through a structured pipeline encompassing registration, optimization, and texture reprojection, culminating in web dissemination via the ATON framework. The methodological focus demonstrates that light painting provides a scalable and replicable solution for documenting complex hypogean contexts, improving the photometric quality and surface readability of 3D models while reducing acquisition time compared to static lighting setups. The results highlight the potential of dynamic illumination as an operational enhancement for 3D recording workflows in low-light cultural heritage environments.

Luminescence imaging is invaluable for studying biological and material systems, particularly when advanced protocols that exploit temporal dynamics are employed. However, implementing such protocols often requires custom instrumentation, either modified commercial systems or fully bespoke setups, which poses a barrier for researchers without expertise in optics, electronics, or software. To address this, we present a versatile macroscopic fluorescence imaging system capable of supporting a wide range of protocols, and provide detailed build instructions along with open-source software to enable replication with minimal prior experience. We demonstrate its broad utility through applications to plants, reversibly photoswitchable fluorescent proteins, and optoelectronic devices.

We developed a machine vision-based robotic system to address automation challenges in pharmaceutical pill sorting and packaging. The hardware platform integrates a high-resolution industrial camera with an HSR-CR605 robotic arm. Image processing leverages the VisionMaster 4.3.0 platform for color classification and positioning. Coordinate mapping between camera and robot is established through a three-point calibration method, with real-time communication realized via the Modbus/TCP protocol. Experimental validation demonstrates that the system achieves 95% recognition accuracy under conditions of pill overlap ≤ 30% and dynamic illumination of 50–1000 lux, ±0.5 mm picking precision, and a sorting efficiency of108 pills per minute. These results confirm the feasibility of integrating domestic hardware and algorithms, providing an efficient automated solution for the pharmaceutical industry. This work makes three key contributions: (1) demonstrating a cost-effective domestic hardware-software integration achieving 42% cost reduction while maintaining comparable performance to imported alternatives, (2) establishing a systematic validation methodology under industrially-relevant conditions that provides quantitative robustness metrics for pharmaceutical automation, and (3) offering a practical implementation framework validated through multi-scenario experiments that bridges the gap between laboratory research and production-line deployment.

Study on the psychophysiological responses under dynamic illuminance and its influence on light comfort

The performance degradation of camouflaged object detection (COD) under complex backgrounds and dynamic illumination conditions has become a challenging issue in optical imaging and detection. To address the limitation of traditional visible-light imaging methods, which easily fail due to their inability to differentiate material and surface optical properties, a polarization-driven multimodal fusion network (PMFNet) is proposed in this paper. High-precision COD is achieved through iterative enhancement of polarization features. First, a feature rectification module is designed based on polarization differences induced by the surface scattering properties of objects. Second, a polarization-guided iterative refinement mechanism is developed, dynamically correcting texture degradation in RGB modality by employing high-resolution polarization features. Finally, a polarization adaptive fusion module is introduced to achieve context-aware complementary enhancement of RGB features through refined polarization information, thus deeply fusing complementary features of the two modalities. The proposed PMFNet demonstrates robust detection performance under adverse illumination and complex background conditions. Experimental results on public datasets demonstrate that the proposed PMFNet outperforms state-of-the-art COD methods.

HighlightsWhat are the main findings?The DIVE algorithm effectively addresses uneven illumination and color distortion in turbid underwater images by combining dynamic illumination correction and visual enhancement modules.The algorithm achieves real-time processing at 25 FPS for 1920 × 1080 resolution videos, making it suitable for embedded devices and underwater robotic inspections.What is the implication of the main finding?DIVE provides a robust solution for underwater defect detection, significantly improving image quality in high-sediment environments (up to 500 g/m3).The method enhances feature extraction for concrete surface defects, such as cracks and holes, supporting applications in marine engineering and dam monitoring.Aiming at the problem of image quality degradation caused by turbid water, non-uniform illumination, and scattering effect in the surface defect detection of underwater concrete structures, firstly, the concrete surface images under different shooting distances, different sediment concentrations, and different illumination conditions were collected through laboratory experiments to simulate the concrete surface images of a reservoir dam with higher sediment concentration and deeper water depth. On this basis, an underwater image enhancement algorithm named DIVE (Dynamic Illumination and Vision Enhancement) is proposed. DIVE solves the problems of luminance unevenness and color deviation in stages through the illumination–scattering decoupling processing framework, and combines efficient computing optimization to achieve real-time processing. The lighting correction of Gaussian distributions (dynamic illumination module) was processed in stages with suspended particle scattering correction (visual enhancement module), and the bright and dark areas were balanced and color offset was corrected by local gamma correction in Lab space and dynamic decision-making of G/B channel. Through thread pool parallelization, vectorization and other technologies, the real-time requirement can be achieved at the resolution of 1920 × 1080. Tests show that DIVE significantly improves image quality in water bodies with sediment concentration up to 500 g/m3, and is suitable for complex scenes such as reservoirs, oceans, and sediment tanks.

Detecting small moving objects under challenging lighting conditions, such as overexposure and underexposure, remains a critical challenge in computer vision applications including surveillance, autonomous driving, and anti-UAV systems. Traditional RGB-based detectors often suffer from degraded object visibility and highly dynamic illumination, leading to suboptimal performance. To address these limitations, we propose a novel RGB-Event fusion framework that leverages the complementary strengths of RGB and event modalities for enhanced small object detection. Specifically, we introduce a Temporal Multi-Scale Attention Fusion (TMAF) module to encode motion cues from event streams at multiple temporal scales, thereby enhancing the saliency of small object features. Furthermore, we design a Sparse Noisy Gated Attention Fusion (SNGAF) module, inspired by the mixture-of-experts paradigm, which employs a sparse gating mechanism to adaptively combine multiple fusion experts based on input characteristics, enabling flexible and robust RGB-Event feature integration. Additionally, we present RGBE-UAV, which is a new RGB-Event dataset tailored for small moving object detection under diverse exposure conditions. Extensive experiments on our RGBE-UAV and public DSEC-MOD datasets demonstrate that our method outperforms existing state-of-the-art RGB-Event fusion approaches, validating its effectiveness and generalization under complex lighting conditions.

Direct Volume Rendering (DVR) plays an important role in scientific data visualization. To generate photo-realistic DVR results, the physical light transport throughout the volume is simulated by applying the Monte Carlo-based volumetric path tracing (VPT) approach. For real-time applications, due to the time constraint for rendering each frame, only a limited number of samples shall be taken for the computation per pixel. This can result in a significant amount of noise in the rendering results. This paper describes our optimized VPT sampling algorithm and a novel denoising technique to generate consistently high-quality realistic DVR results in real time. We develop a new shading model that can reduce estimation variance to enhance the quality of DVR results. Additionally, a hybrid acceleration structure is created by integrating both octree and macrocell to improve sampling efficiency. This allows the acquisition of sufficiently more shading samples while maintaining the desired interactive frame rate. To further eliminate remaining noise and improve temporal stability of DVR results, we develop a novel spatiotemporal denoising framework. Our denoiser decouples the estimated radiance into high-detail low-noise and low-detail high-noise components. Different denoising algorithms are separately applied to these components to reduce noise without introducing blurring artifacts. Our DVR system can consistently offer high rendering quality and good temporal stability across DVR result frames in real time. During fast user interactions and with rapid alterations of the illumination condition, our rendering method can still provide good visual comfort and representation accuracy without visible latency.

Spikelet number, a core phenotypic parameter for wheat yield composition, requires precise estimation through accurate spike contour extraction and differentiation between grain surfaces and spikelet surfaces. However, technical challenges persist in precise spike segmentation under complex field backgrounds and morphological differentiation between grain/spikelet surfaces. Building on two-year multi-angle wheat spike imagery, we propose an enhanced YOLOv9-LDS multi-scale object detection framework. The algorithm innovatively constructs a lightweight depthwise separable network (LDSNet) as backbone, balancing computational efficiency and accuracy through channel re-parameterization strategy; incorporates an Efficient Local Attention (ELA) module to build feature enhancement networks, and employs dual-path feature fusion mechanisms to strengthen edge texture responses, significantly improving discrimination of overlapping spikes and complex backgrounds. Further optimizes the loss function system by replacing traditional IoU with Scylla Intersection over Union (SIoU) metric, enhancing bounding box regression through dynamic focus factors, and adding high-resolution small-object detection layers to mitigate dense spikelet feature loss. Independent test set validation shows the improved model achieves 83.9% contour integrity recognition rate and 92.4% mAP@0.5, exceeding baseline by 3.2 and 5.3% points respectively. Ablation studies confirm LDSNet-ELA integration reduces false positives by 27.6%, while the enhanced loss function system improves small-object recall by 19.4%. The proposed framework demonstrates superior performance in complex field scenarios with dense targets and dynamic illumination. The multi-scale feature synergy enhancement mechanism overcomes traditional models' limitations in detecting overlapping spikes. This method not only enables precise spike phenotyping but also provides robust algorithmic support for intelligent field spikelet counting systems, advancing translational applications in crop phenomics.

ABSTRACTMany animals adopt camouflage strategies that involve matching their appearance to colour and texture-based features in their environment. However, these features may be difficult to estimate in habitats that are prone to dynamic lighting, which might alter the features' appearance, or disrupt the capacity of visual systems to resolve those features. In this study, we tested whether a common form of shallow underwater dynamic lighting termed ‘caustics’, consisting of moving light bands travelling along the substrate, affect the expression of skin papillae in cuttlefish (Sepia officinalis). To do so, we exposed cuttlefish individuals to rock stimuli varying in their surface texture and colouration in both caustic and non-caustic lighting and scored their papillae expression. We established a positive correlation between the degree of papillae expression and the maximum contrast cues in the visual scene, such as those derived from object surface texture or colouration, with stronger contrast cues resulting in a more pronounced papillae expression. In addition, we found that cuttlefish also expressed their papillae when exposed to caustics, and this response was adopted irrespective of the presence or absence of an object in their visual field, highlighting that increased visual contrast levels deriving from exposure to dynamic lighting alone can elicit papillae expression in cuttlefish. We discuss whether these camouflage responses might be adaptive, reducing their likelihood of being detected by predators, or alternatively could represent a constraint on visual processing.

ECR Spotlight is a series of interviews with early-career authors from a selection of papers published in Journal of Experimental Biology and aims to promote not only the diversity of early-career researchers (ECRs) working in experimental biology but also the huge variety of animals and physiological systems that are essential for the ‘comparative’ approach. Christian Drerup is an author on ‘ Visual contrast from background features and dynamic illumination contributes to three-dimensional camouflage in cuttlefish’, published in JEB. Christian conducted the research described in this article while a PhD student in James Herbert-Read's lab at University of Cambridge, Cambridge, UK. Christian is now a Postdoctoral Research Associate in the lab of Olena Riabinina at Department of Biosciences, Durham University, Durham, UK, investigating how animals adapt their behavioural strategies based on the sensory information they can perceive, especially in perceptually challenging environments.