Remote Image Research Articles

Remote Sensing Data-Based Satellite Image Analysis in Water Quality Detection for Public Health Data Modelling

Remote Sensing Data-Based Satellite Image Analysis in Water Quality Detection for Public Health Data Modelling

Targeting Shallow Subsurface Sampling for Mars at Oxia Planum Using Fluvial Erosion–Deposition Modeling

A model-based surface fluvial erosion and deposition approach was adapted to Martian conditions to forecast the potential locations for shallow subsurface sampling by the Rosalind Franklin ExoMars rover at Oxia Planum. While remote and on-site images show only the surface visible features, former fluvial-related accumulation sites might be hidden. During the fluvial activity, most accumulation-related areas are interesting with regard to clay-like sediments, which could adsorb organics effectively—such sites could be identified by modeling. By applying the SIMWE fluvial erosion/deposition model, substantial variability in accumulation and deposition-dominated areas with their specific pattern and spatial distribution could be outlined, indicating that sophisticated targeting of future sampling could use such a model-based approach. At the main valley-like feature, former water flow tracks were identified, as well as deposition-dominated locations, which are the best targets for shallow subsurface sampling. Joint evaluation of safety aspects like slope angle and loose sand dunes with scientific aspects provide the best sampling locations. Such model-based targeting is important as by using only orbital images, these locations could not be identified.

Standardizing Multidisciplinary Case Conferences and Improving Communication Between Referring Physicians and Radiologists: A Quality Improvement Initiative.

Purpose: Assess the effectiveness of standardizing multidisciplinary case conferences (MDCs). Methods: Anonymous electronic surveys gauged opinions of abdominal radiologists engaged in recurring MDCs. A standardized Excel template, following Cancer Care Ontario guidelines and relevant literature, was distributed to MDC managers. Physicians were instructed to send cases 36 hours prior to MDC. Template adherence was assessed at 1.5 and 8 months. A follow-up survey at 4 months evaluated the intervention's effectiveness. Results: 27/34 abdominal radiologists provided 47 baseline responses, and 12 delegated radiologists provided 23 follow-up responses. "Often/always" being provided the image's location increased from 36% (17/47) at baseline to 70% (16/23) at follow-up. Non-adherence to the 36-hour cut-off decreased from 36% (16/45) to 17% (4/23). 72% disagreed that uploading remote imaging to hospital servers is easy (33/46), similar to follow-up (18/23, 78%). In assessing the intervention, 41% noted improved standardization (9/22), another 41% considered MDCs already standardized (9/22), and 18% reported no change (4/22). Those reporting no change experienced a higher frequency of non-adherence to the 36-hour cut-off (3/4, 75%) than others (1/18, 6%), and less frequent "often/always" ratings for image location being provided (3/4, 75%) than others (2/18, 11%). 89% (25/28) of MDCs adhered to the template. Issues regarding last-minute add-on cases may be mitigated through EPIC force functions. Artificial intelligence advancements may assist in retrieving external images and patient information. Conclusion: Adherence to MDC standardization was high, allowing for more efficient preparation, potentially reducing radiologist administrative burdens. Future force functions and artificial intelligence integration into electronic patient records may further augment this.

A hybrid approach for integrating micro-satellite images and sensors network-based ground measurements using deep learning for high-resolution prediction of fine particulate matter (PM2.5) over an indian city, lucknow

The detrimental impacts of fine particulate matter (PM2.5) on human health, climate, ecosystems, crops, and building materials are well-established. However, there remain unresolved inquiries regarding the precise location of the sources of PM2.5. This study is the first attempt to use a calibrated sensors-based ambient air quality monitoring network (SAAQM network) and regulatory government monitors to train micro-satellite images for high spatial-resolution air pollution field determination of PM2.5 in Lucknow, Uttar Pradesh, India. A hybrid approach is developed to integrate three different datasets that include microsatellite images, PM2.5 ground measurements, and supporting information (meteorological parameters and geographical coordinates), to be fed into a Random Trees-Random Forest- Convolutional Neural Network (RT-RF-CNN) joint model to estimate PM2.5 concentrations at a sub-km level. The RT-RF-CNN joint model can derive PM2.5 concentrations at a spatial resolution of 500 m with statistically significant indicators such as spatial r of 0.9, a low root-mean-square error of 26.9 μg/m3 and a mean absolute error of 17.2 μg/m3. Based on our approach, the PM2.5 prediction maps using micro-satellite images (spatial resolution of 3m/pixel) and RT-RF-CNN joint model were generated for each day throughout the study period (December 2021–December 2022). The inter-grid comparison of these maps revealed the intra-urban local hotspots and coolspots at a fine-granular level seasonally, monthly, and daily. It is observed that the monsoon season has the highest number of coolspots (67%), while winter (0.1%), post-monsoon (0.5%) and summer (11%) have fewer. It is noted that the high temporal-spatial information of PM2.5 estimates from our integrated approach is not achievable by ground-based measurements and other existing satellite-based estimates alone. The findings of this study have potential applications on a diverse array, encompassing near real-time daily PM2.5 predicted maps, specific air pollution hotspot identification, PM2.5 exposure assessment at the neighbourhood level, and integration of remote sensing-based micro-satellite images and ground-based measurements.

Sustainable Agriculture-Based Climate Change Training Models using Remote Hyperspectral Image with Machine Learning Model

Sustainable Agriculture-Based Climate Change Training Models using Remote Hyperspectral Image with Machine Learning Model

Coherent ghost imaging via phase compensation

The 4-f optical illumination structure is widely used in coherent ghost imaging (CGI) system but its imaging field of view (FOV) is usually restricted by the transmission aperture of the lens. Therefore, this optical illumination structure cannot be applied to remote imaging. Based on conventional imaging illumination structure, a large-FOV CGI via phase compensation (CGI-PC) is proposed and its imaging property is investigated. The experimental results demonstrate that both the FOV and the quality of CGI can be dramatically improved when a lens is introduced into the receiving system to compensate the system’s spatial phase difference. Factors influencing the quality of CGI-PC are also discussed.

(Invited) Carbon Nanostructures for Targeted Delivery and Health Monitoring in Plants

To meet the demand for food and bioenergy of a rapidly growing global population requires novel and convergent approaches to enhance photosynthetic organism function. Nanotechnology is enabling the development of targeted and species independent tools for advancing plant bioengineering and precision agriculture.My research group has developed carbon nanotube-based sensors capable of real-time sensing of plant signaling molecules. These nanosensors enabled the creation of plant sentinels that communicate the health status of plants to electronic devices. Near-infrared (nIR) fluorescent single-walled carbon nanotubes (SWCNTs) were designed and interfaced with plant leaves to report hydrogen peroxide (H2O2), a key signaling molecule associated with the onset of plant stress. The sensor nIR fluorescence response (>900 nm) is quenched by H2O2 with selectivity against other stress-associated signaling molecules and within the plant physiological range (10–100 H2O2 μM). In vivo remote nIR imaging of H2O2 sensors enabled optical monitoring of plant health in response to stresses including UV-B light (−11%), high light (−6%), and a pathogen-related peptide (flg22) (−10%), but not mechanical leaf wounding (<3%). The sensor’s high biocompatibility was reflected on similar leaf cell death (<5%) and photosynthetic rates to controls without SWCNT. These optical nanosensors report early signs of stress and will improve our understanding of plant stress communication, provide novel tools for precision agriculture, and optimize the use of agrochemicals in the environment.We also developed targeted carbon-based nanomaterials as tools for precise chemical delivery (carbon dots, CDs) and gene delivery platforms (SWCNTs) to plant organelles (chloroplasts) and vasculature (phloem). A biorecognition approach of coating the nanomaterials with a rationally designed chloroplast targeting peptide improved the delivery of CDs with molecular baskets (TP-β-CD) for delivery of agrochemicals and of plasmid DNA coated SWCNT (TP-pATV1-SWCNT) from 47% to 70% and from 39% to 57% of chloroplasts in leaves, respectively. Plants treated with TP-β-CD (20 mg/L) and TP-pATV1-SWCNT (2 mg/L) had a low percentage of dead cells, 6% and 8%, respectively, similar to controls without nanoparticles, and no permanent cell and chloroplast membrane damage after 5 days of exposure. Orthogonally, the surfaces of carbon dot nanocarriers were functionalized with sucrose, enabling rapid and efficient foliar delivery into the plant phloem, a vascular tissue that transports sugars, signaling molecules, and agrochemicals through the whole plant. The chemical affinity of sucrose molecules to sugar membrane transporters on the phloem cells enhances the uptake of biocompatible carbon dots with β-cyclodextrin molecular baskets (suc-β-CD) that can carry a wide range of agrochemicals. The nanoparticle fluorescence emission properties allowed detection and monitoring of rapid translocation (<40 min) in the vasculature of wheat leaves by confocal and epifluorescence microscopy. The suc-β-CDs more than doubled the delivery of chemical cargoes into the leaf vascular tissue. The sucrose coating of nanoparticles approach enables unprecedented targeted delivery to roots with ≈70% of phloem-loaded nanoparticles delivered to roots.The use of plant biorecognition molecules mediated delivery provides an approach for guiding nanocarriers with their cargoes to plant organelles, cells and tissues and whole plants and engineering safer and efficient agrochemical and biomolecule delivery tools. Communication and actuation of plants mediated by nanosensors and targeted nanostructures can turn plants into technology and lead to a more precise and sustainable agriculture with reduced environmental impact.

Flow-Field Inference for Turbulent Exhale Flow Measurement.

Vision-based pulmonary diagnostics present a unique approach for tracking and measuring natural breathing behaviors through remote imaging. While many existing methods correlate chest and diaphragm movements to respiratory behavior, we look at how the direct visualization of thermal CO2 exhale flow patterns can be tracked to directly measure expiratory flow. In this work, we present a novel method for isolating and extracting turbulent exhale flow signals from thermal image sequences through flow-field prediction and optical flow measurement. The objective of this work is to introduce a respiratory diagnostic tool that can be used to capture and quantify natural breathing, to identify and measure respiratory metrics such as breathing rate, flow, and volume. One of the primary contributions of this work is a method for capturing and measuring natural exhale behaviors that describe individualized pulmonary traits. By monitoring subtle individualized respiratory traits, we can perform secondary analysis to identify unique personalized signatures and abnormalities to gain insight into pulmonary function. In our study, we perform data acquisition within a clinical setting to train an inference model (FieldNet) that predicts flow-fields to quantify observed exhale behaviors over time. Expiratory flow measurements capturing individualized flow signatures from our initial cohort demonstrate how the proposed flow field model can be used to isolate and analyze turbulent exhale behaviors and measure anomalous behavior. Our results illustrate that detailed spatial flow analysis can contribute to unique signatures for identifying patient specific natural breathing behaviors and abnormality detection. This provides the first-step towards a non-contact respiratory technology that directly captures effort-independent behaviors based on the direct measurement of imaged CO2 exhaled airflow patterns.

Patient preferences for teleradiology services and remote image interpretation: An empirical study.

This study aims to examine patient preferences for teleradiology services and remote image interpretation. In this context, this study aims to address the following research questions: (i) How do patients perceive teleradiology services, focusing specifically on comfort, quality of care and satisfaction, communication and information, and accessibility and ease of use; (ii) How do patient demographics (age, gender, urban vs. rural residence) influence the perceptions on teleradiology services? A cross-sectional survey design is adopted in this study. The survey comprises five sections targeting demographic information, comfort, and preferences regarding remote image interpretation, perceived quality of care and satisfaction, communication and information clarity, and accessibility and technology aspects using five-point Likert scale ratings. A total of 406 patients (209 males and 197 females; 170 urban residents; 174 semi-urban residents, and 62 rural residents) using teleradiology services participated in the study. Participants reported high satisfaction with remote image interpretation (3.78 ± 1.19), quality of care (3.31 ± 1.19), understanding (3.84 ± 1.43), and user-friendliness (3.67 ± 1.29). Key issues were technical problems (3.81 ± 1.35), feedback difficulties (3.19 ± 1.58), privacy concerns (2.43 ± 1.46), and low awareness (2.37 ± 1.12). Urban participants scored significantly better in comfort, preferences, and communication than those from semi-urban and rural areas. Teleradiology design and implementation should be optimized to align with patient preferences and enhance overall satisfaction.

Revolutionizing Skin Cancer Triage: The Role of Patient-Initiated Teledermoscopy in Remote Diagnosis.

Teledermatology, defined as the use of remote imaging technologies to provide dermatologic healthcare services to individuals in a distant setting, has grown considerably in popularity since its widespread implementation during the COVID-19 pandemic. Teledermoscopy employs a smartphone dermatoscope attachment paired with a smartphone camera to visualize colors and microstructures within the epidermis and superficial dermis that cannot be seen with the naked eye ABCD criteria alone. Our retrospective observational cohort and case-control study evaluated the utility of loaning a smartphone dermatoscope attachment to patients for remote triage of self-selected lesions of concern for skin cancer. The primary outcome was the number (percentage) of in-person follow-up visits required for patients who submitted lesion images, either with or without accompanying dermoscopic images. A medical record review was conducted on all Oregon Health & Science University Department of Dermatology spot check image submissions utilizing the smartphone dermatoscopes between August 2020 and August 2022. De-identified dermoscopic images of lesions that included corresponding non-dermoscopic clinical images in their submission (n = 70) were independently reviewed by a blinded expert dermoscopist. The expert used standard clinical algorithms (ABCD criteria for clinical images; dermoscopy three-point checklist for dermoscopic images) to determine whether the imaged lesion should be converted to an in-person visit for further evaluation and consideration for biopsy. Of the 70 lesions submitted with corresponding clinical and dermoscopy images, 60 met the criteria for in-person evaluation from clinical (non-dermoscopic) image review compared to 28 meeting the criteria for in-person evaluation from dermoscopic images of the same lesion. Thus, a 53% reduction in conversion to an in-person consultation with the addition of smartphone dermatoscope images in virtual lesion triage was observed (p < 0.001, McNemar's Test). Implementing patient-led teledermoscopy may reduce the frequency of in-person visits for benign lesions and consequently improve access to in-person dermatology consultations for patients with concerning and possibly malignant lesions.

Deep learning and remote photoplethysmography powered advancements in contactless physiological measurement.

In recent decades, there has been ongoing development in the application of computer vision (CV) in the medical field. As conventional contact-based physiological measurement techniques often restrict a patient's mobility in the clinical environment, the ability to achieve continuous, comfortable and convenient monitoring is thus a topic of interest to researchers. One type of CV application is remote imaging photoplethysmography (rPPG), which can predict vital signs using a video or image. While contactless physiological measurement techniques have an excellent application prospect, the lack of uniformity or standardization of contactless vital monitoring methods limits their application in remote healthcare/telehealth settings. Several methods have been developed to improve this limitation and solve the heterogeneity of video signals caused by movement, lighting, and equipment. The fundamental algorithms include traditional algorithms with optimization and developing deep learning (DL) algorithms. This article aims to provide an in-depth review of current Artificial Intelligence (AI) methods using CV and DL in contactless physiological measurement and a comprehensive summary of the latest development of contactless measurement techniques for skin perfusion, respiratory rate, blood oxygen saturation, heart rate, heart rate variability, and blood pressure.

Advancements in Remote Compressive Hyperspectral Imaging: Adaptive Sampling with Low-Rank Tensor Image Reconstruction

We advanced the practical development of compressive hyperspectral cameras for remote sensing scenarios with a design that simultaneously compresses and captures high-quality spectral information of a scene via configurable measurements. We built a prototype imaging system that is compatible with light-modulation devices that encode the incoming spectrum. The sensing approach enables a substantial reduction in the volume of data collected and transmitted, facilitating large-scale remote hyperspectral imaging. A main advantage of our sensing design is that it allows for adaptive sampling. When prior information of a survey region is available or gained, the modulation patterns can be re-programmed to efficiently sample and detect desired endmembers. Given target spectral signatures, we propose an optimization scheme that guides the encoding process. The approach severely reduces the number of required sampling patterns, with the ability to achieve image segmentation and correct distortions. Additionally, to decode the modulated data, we considered a novel reconstruction algorithm suited for large-scale images. The computational methodology leverages the multidimensional structure and redundant representation of hyperspectral images via the canonical polyadic decomposition of multiway arrays. Under realistic remote sensing scenarios, we demonstrated the efficiency of our approach with several data sets collected by our prototype camera and reconstructed by our low-rank tensor decoder.

Extraction of water bodies from high-resolution remote sensing imagery based on a deep semantic segmentation network

The precise delineation of urban aquatic features is of paramount importance in scrutinizing water resources, monitoring floods, and devising water management strategies. Addressing the challenge of indistinct boundaries and the erroneous classification of shadowed regions as water in high-resolution remote sensing imagery, we introduce WaterDeep, which is a novel deep learning framework inspired by the DeepLabV3 + architecture and an innovative fusion mechanism for high- and low-level features. This methodology first creates a comprehensive dataset of high-resolution remote sensing images, then progresses through the Xception baseline network for low-level feature extraction, and harnesses densely connected Atrous Spatial Pyramid Pooling (ASPP) modules to assimilate multi-scale data into sophisticated high-level features. Subsequently, the network decoder amalgamates the elemental and intricate features and applies dual-line interpolation to the amalgamated dataset to extract aqueous formations from the remote images. Experimental evidence substantiates that WaterDeep outperforms its existing deep learning counterparts, achieving a stellar overall accuracy of 99.284%, FWIoU of 95.58%, precision of 97.562%, recall of 95.486%, and F1 score of 96.513%. It also excels in the precise demarcation of edges and the discernment of shadows cast by urban infrastructure. The superior efficacy of the proposed method in differentiating water bodies in complex urban environments has significant practical applications in real-world contexts.

Research Progress of Using Canine Olfactory Search for Human Remains.

In the process of murder investigation, it is of great significance to find the discarded and buried human remains accurately. The main methods of searching for human remains include human visual search, aerial detection, geophysical technology, remote imaging technology and canine olfactory search technique. Canine olfactory search for human remains is a recognized time-effective and non-invasive search method, making dogs the most valuable search tool in forensic investigation. By systematically reviewing and summarizing relevant literature, and based on the theory of volatile organic compound produced by the decomposition of human remains, this paper explores the basic principle of the canine olfactory search technique for human remains. This paper also reviews the application of training canine search technique for human remains in forensic investigation by using human blood, tissue, cadaver putrefying fluid and odor substitutes as sniffing sources. The application prospect of canine olfactory search for human remains was prospected from the perspectives of detection of volatile organic compound during cadaver decay, development of odor substitutes and adsorption devices, and technology tactics used in canine training and use, to provide references for the relevant research of canine olfactory search for human remains in China.

Assessing Many Image Processing Products Retrieved from Sentinel-2 Data to Monitor Shallow Landslides in Agricultural Environments

Remote images are useful tools for detecting and monitoring landslides, including shallow landslides in agricultural environments. However, the use of non-commercial satellite images to detect the latter is limited because their spatial resolution is often comparable to or greater than landslide sizes, and the spectral characteristics of the pixels within the landslide body (LPs) are often comparable to those of the surrounding pixels (SPs). The buried archaeological remains are also often characterized by sizes that are comparable to image spatial resolutions and the spectral characteristics of the pixels overlying them (OBARPs) are often comparable to those of the pixels surrounding them (SBARPs). Despite these limitations, satellite images have been used successfully to detect many buried archaeological remains since the late 19th century. In this research context, some methodologies, which examined the values of OBARPs and SBARPs, were developed to rank images according to their capability to detect them. Based on these previous works, this paper presents an updated methodology to detect shallow landslides in agricultural environments. Sentinel-2 and Google Earth (GE) images were utilized to test and validate the methodology. The landslides were mapped using GE images acquired simultaneously or nearly simultaneously with the Sentinel-2 data. A total of 52 reference data were identified by monitoring 14 landslides over time. Since remote sensing indices are widely used to detect landslides, 20 indices were retrieved from Sentinel-2 images to evaluate their capability to detect shallow landslides. The frequency distributions of LPs and SPs were examined, and their differences were evaluated. The results demonstrated that each index could detect shallow landslides with sizes comparable to or smaller than the spatial resolution of Sentinel-2 data. However, the overall accuracy values of the indices varied from 1 to 0.56 and two indices (SAVI and RDVI) achieved overall accuracy values equal to 1. Therefore, to effectively distinguish areas where shallow landslides are present from those where they are absent, it is recommended to apply the methodology to many image processing products. In conclusion, given the significant impact of these landslides on agricultural activity and surrounding infrastructures, this methodology provides a valuable tool for detecting and monitoring landslide presence in such environments.

Instrumentation and testing for road condition monitoring – A state-of-the-art review

The sensing instrumentation and testing operations routinely implemented in-situ to monitor the condition of both paved (flexible, rigid) and unpaved roads are critically reviewed: image acquisition, appraisal of surface profile and skid resistance, measurement of stress, strain and deflection, execution of tomography tests, evaluation of environmental parameters as well as determination of traffic volume. Further, this state-of-the-art review sheds light on the financial aspects and cost spectra of the different diagnostic procedures. Finally, a systematic literature review highlights trends in recent research and documents that scientific studies have mainly focused on remote imaging, deflection and tomography measurements.

Field application of near-borehole fracture acoustic imaging using a three-dimensional scanning inversion method

Borehole acoustic remote imaging can help identify and evaluate abnormal geological bodies within tens of meters of a well, which has a very important application prospect in fracture imaging. Compared to the borehole mode wave, the echo amplitude from a near-borehole fracture is considerably small, and the currently used wave field separation method cannot completely filter the borehole mode wave. Due to the influence of the large amplitude of the remaining borehole mode wave, several artifacts are often produced near the well when using traditional migration imaging. Thus, imaging and characterizing near-borehole fractures via acoustic imaging becomes difficult. To overcome this, we propose a three-dimensional scanning inversion method that attenuates the effects of remaining refracted waves on borehole acoustic imaging. We then perform data processing and imaging of field-logging data based on this method, and multimode echoes are used to identify and evaluate fractures. Finally, a comparison of the inversion results of the borehole acoustic and borehole electrical imaging showed that the three-dimensional scanning inversion method can effectively attenuate the remaining refracted wave and better realize the clear imaging of near-borehole fractures. Different modes of echoes illuminate different areas of fractures near wells. Therefore, based on scattered S-waves, the proposed imaging method can effectively detect fractures that cross-cut a well. Moreover, based on multimode waves, the method can suppress artifacts and highlight the azimuth characteristics of near-borehole fractures. Using the proposed method, high-precision positioning and fine characterization of formation fractures in near and far wells can be achieved simultaneously. Further, near-borehole fracture parameters—such as tendency, dip, and depth—can be inverted, and the results are consistent with the inversion results obtained through borehole electrical imaging. Thus, the proposed method enables the quantitative evaluation of near-borehole fractures.

Continuous Associations between Remote Self-Administered Cognitive Measures and Imaging Biomarkers of Alzheimer’s Disease

BackgroundEasily accessible and self-administered cognitive assessments that can aid early detection for Alzheimer’s disease (AD) dementia risk are critical for timely intervention.Objectives/DesignThis cross-sectional study investigated continuous associations between Mayo Test Drive (MTD)–a remote, self-administered, multi-device compatible, web-based cognitive assessment–and AD-related imaging biomarkers.Participants/Setting684 adults from the Mayo Clinic Study of Aging and Mayo Clinic Alzheimer’s Disease Research Center participated (age=70.4±11.2, 49.7% female). Participants were predominantly cognitively unimpaired (CU; 94.0%).MeasurementsParticipants completed (1) brain amyloid and tau PET scans and MRI scans for hippocampal volume (HV) and white matter hyperintensities (WMH); (2) MTD remotely, consisting of the Stricker Learning Span and Symbols Test which combine into an MTD composite; and (3) in-person neuropsychological assessment including measures to obtain Mayo Preclinical Alzheimer’s disease Cognitive Composite (Mayo-PACC) and Global-z. Multiple regressions adjusted for age, sex, and education queried associations between imaging biomarkers and scores from remote and in-person cognitive measures.ResultsLower performances on MTD were associated with greater amyloid, entorhinal tau, and global tau PET burden, lower HV, and higher WMH. Mayo-PACC and Global-z were associated with all imaging biomarkers except global tau PET burden. MCI/Dementia participants showed lower performance on all MTD measures compared to CU with large effect sizes (Hedge’s g’s=1.65–2.02), with similar findings for CU versus MCI only (Hedge’s g’s=1.46–1.83).ConclusionMTD is associated with continuous measures of AD-related imaging biomarkers, demonstrating ability to detect subtle cognitive change using a brief, remote assessment in predominantly CU individuals and criterion validity for MTD.

Remote 3D Imaging and Classification of Pelagic Microorganisms with A Short‐Range Multispectral Confocal LiDAR

AbstractPlankton is essential to maintain healthy aquatic ecosystems since it influences the biological carbon pump globally. However, climate change‐induced alterations to oceanic properties threaten planktonic communities. It is therefore crucial to monitor their abundance to assess the health status of marine ecosystems. In situ optical tools unlock high‐resolution measurements of sub‐millimeter specimens, but state‐of‐the‐art underwater imaging techniques are limited to fixed and small close‐range volumes, requiring the instruments to be vertically dived. Here, a novel scanning multispectral confocal light detection and ranging (LiDAR) system for short‐range volumetric sensing in aquatic media is introduced. The system expands the inelastic confocal principle to multiple wavelength channels, allowing the acquisition of 4D point clouds combining near‐diffraction limited morphological and spectroscopic data that is used to train artificial intelligence (AI) models. Volumetric mapping and classification of microplastics is demonstrated to sort them by color and size. Furthermore, in vivo autofluorescence is resolved from a community of free‐swimming zooplankton and microalgae, and accurate spectral identification of different genera is accomplished. The deployment of this photonic platform alongside AI models overcomes the complex and subjective task of manual plankton identification and enables non‐intrusive sensing from fixed vantage points, thus constituting a unique tool for underwater environmental monitoring.

Sub-Band Backdoor Attack in Remote Sensing Imagery

Remote sensing datasets usually have a wide range of spatial and spectral resolutions. They provide unique advantages in surveillance systems, and many government organizations use remote sensing multispectral imagery to monitor security-critical infrastructures or targets. Artificial Intelligence (AI) has advanced rapidly in recent years and has been widely applied to remote image analysis, achieving state-of-the-art (SOTA) performance. However, AI models are vulnerable and can be easily deceived or poisoned. A malicious user may poison an AI model by creating a stealthy backdoor. A backdoored AI model performs well on clean data but behaves abnormally when a planted trigger appears in the data. Backdoor attacks have been extensively studied in machine learning-based computer vision applications with natural images. However, much less research has been conducted on remote sensing imagery, which typically consists of many more bands in addition to the red, green, and blue bands found in natural images. In this paper, we first extensively studied a popular backdoor attack, BadNets, applied to a remote sensing dataset, where the trigger was planted in all of the bands in the data. Our results showed that SOTA defense mechanisms, including Neural Cleanse, TABOR, Activation Clustering, Fine-Pruning, GangSweep, Strip, DeepInspect, and Pixel Backdoor, had difficulties detecting and mitigating the backdoor attack. We then proposed an explainable AI-guided backdoor attack specifically for remote sensing imagery by placing triggers in the image sub-bands. Our proposed attack model even poses stronger challenges to these SOTA defense mechanisms, and no method was able to defend it. These results send an alarming message about the catastrophic effects the backdoor attacks may have on satellite imagery.