Multispectral Camera Research Articles (Page 1)

Biological vision systems excel at acquiring and processing information, but there is often a trade-off between these capabilities. For instance, mantis shrimp possess exceptional spectral sensing but poor color perception due to limited neural processing. Taking the best of both worlds, the mantis shrimp's spectral detection ability and the human-like visual processing power are integrated to achieve full-color perception. Using an aerosol-liquid-solid spraying technique, an array of high-quality, excess ion migration enhanced perovskite narrowband photodetectors spanning the ultraviolet to visible spectrum is developed. These detectors enable a computational multispectral imaging system that captures seven spectral images in one shot. A deep-learning-based color fusion network is designed to efficiently translate multispectral inputs into an RGB representation, significantly enhancing color recognition of these mantis shrimp-inspired multispectral cameras and affording the capability to overcome metamerism. These perovskite intelligent camera leverages the strengths of biological vision and demonstrate a novel approach to multispectral imaging that could advance applications in machine vision, remote sensing, and medical imaging.

MCPS: A multispectral composite projection stealth technology for invisibility under multispectral camera detection

Abstract. Identifying emergent aquatic vegetation (EAV) species is important to monitor environmental changes and support in decision-making. With the advent of uncrewed aerial vehicle (UAV), it has been possible to acquire high resolution images and assist in this task. At the same time, the high level of detail of these images can be considered noise for some segmentation algorithms and testing different smoothing and subsampling variations can be very relevant. Hence, the aim of this study is to analyse the performance of three segmentation algorithms (region growing, SLIC superpixel and watershed) to generate “homogeneous” regions in high resolution images, considering blur and scale change effects. To do this, images of a lake impacted by EAV were captured using a multispectral camera on board of UAV with 8 mm ground sample distance. After image processing, an orthomosaic was produced and three clippings were extracted from it to be segmented and tested empirically with variations of subsampling (1 cm, 1.6 cm, 2 cm, and 3 cm) and standard deviation smooth filter application (σ = 2, 4, and 8). The results showed that region growing and watershed algorithms are the most affected by high spatial resolution, and greatly benefits from the smoothing and subsampling applied, i.e., reducing the amount of detail, while superpixel algorithm created more consistent and uniform results, especially after smoothing, as evidenced by the quantitative evaluation based on segment entropy, characterized by kurtosis.

Sugar content is a crucial indicator of grape ripeness and grading, and developing non-contact and non-destructive sugar content detection devices is essential for grape-picking robots and sorting platforms. Spectroscopy, which can detect the chemical composition of grapes, has become a key technology for developing non-destructive testing devices. In this paper, we collected 2,880 randomly labeled multispectral images of Sunshine Rose grapes with a Changguang Yuchen MS600 PRO multispectral camera and measured the sugar content (in Brix values) of the labeled grapes with a handheld refractometer, using data exclusively from this grape variety. To address noise and misalignment issues in the multispectral images, we proposed preprocessing methods including Gaussian denoising and ECC (Enhanced Correlation Coefficient) algorithm registration. Based on a ResNet-50 residual network, we constructed a grape sugar content prediction regression model Improved-Res with SE (Squeeze-and-Excitation) attention modules, DSC (Depthwise Separable Convolutions), and Inception modules. The model's performance was evaluated by MSE (Mean Squared Error), MAE (Mean Absolute Error), and R2 (R-Square) metrics. We compared the performance of four feature extraction methods combined with four traditional machine learning models, as well as seven deep learning models. The results showed that among traditional machine learning methods, the combination of color histogram feature extraction and the XGBoost regression achieved the best performance, with MSE, MAE, and R2 of 1.35, 0.90 Brix, and 0.78, respectively. Among deep learning methods, the ResNet-50 model demonstrated the best performance, with MSE, MAE, and R2 of 0.95, 0.96 Brix, and 0.84, respectively. Effective improvements of SE attention module, depthwise separable convolutions, and Inception module in the ResNet-50 model was confirmed through ablation experiments: the proposed Improved-Res model achieved MSE, MAE, and R2 of 0.49, 0.55 Brix, and 0.92, respectively, which significantly outperformed traditional machine learning methods and classical deep learning models.

Post-fire soil and vegetation changes can intensify erosion and sediment yield by altering the factors controlling the runoff–infiltration balance. Erosion barriers (EBs) are widely used in hydrological and forest restoration to mitigate erosion, reduce sediment transport, and promote vegetation recovery. However, precise spatial assessments of their effectiveness remain scarce, requiring validation through operational methodologies. This study evaluates the impact of EB on post-fire vegetation recovery at two temporal and spatial scales: (1) Remotely Piloted Aircraft System (RPAS) imagery, acquired at high spatial resolution but limited to a single acquisition date coinciding with the field flight. These data were captured using a MicaSense RedEdge-MX multispectral camera and an RGB optical sensor (SODA), from which NDVI and vegetation height were derived through aerial photogrammetry and digital surface models (DSMs). (2) Sentinel-2 satellite imagery, offering coarser spatial resolution but enabling multi-temporal analysis, through NDVI time series spanning four consecutive years. The study was conducted in the area of the Luna Fire (northern Spain), which burned in July 2015. A paired sampling design compared upstream and downstream areas of burned wood stacks and control sites using NDVI values and vegetation height. Results showed slightly higher NDVI values (0.45) upstream of the EB (p < 0.05), while vegetation height was, on average, ~8 cm lower than in control sites (p > 0.05). Sentinel-2 analysis revealed significant differences in NDVI distributions between treatments (p < 0.05), although mean values were similar (~0.32), both showing positive trends over four years. This study offers indirect insight into the functioning and effectiveness of EB in post-fire recovery. The findings highlight the need for continued monitoring of treated areas to better understand environmental responses over time and to inform more effective land management strategies.

In the lower and middle sections of the Yangtze River Basin Region (YRBR) in China, challenges posed by climate change and delayed harvesting of preceding crops have hindered the timely sowing of wheat, leading to an increasing prevalence of late-sown wheat fields. This trend has emerged as a significant impediment to achieving high and stable production of wheat in this area. During the growing seasons of 2022–2023 and 2023–2024, an unmanned aerial vehicle (UAV)-based multispectral camera was used to monitor different wheat materials at various growth stages under normal sowing treatment (M1) and late sowing with increased plant density (M2). By assessing yield loss, the wheat tolerance to late sowing was quantified and categorized. The correlation between the differential vegetation indices (D-VIs) and late sowing resistance was examined. The findings revealed that the J2-Logistic model demonstrated optimal classification performance. The precision values of stable type, intermediate type, and sensitive type were 0.92, 0.61, and 1.00, respectively. The recall values were 0.61, 0.92, and 1.00. The mean average precision (mAP) of the model was 0.92. This study proposes a high-throughput and low-cost evaluation method for wheat tolerance to late sowing, which can provide a rapid predictive tool for screening suitable varieties for late sowing and facilitating late-sown wheat breeding.

The responsibility for risk assessment and user safety in forested and recreational areas lies with the property owner. This study shows that unmanned aerial vehicles (UAVs), combined with remote sensing and GIS analysis, effectively support the identification of high-risk trees, particularly those with reduced structural stability. UAV-based surveys successfully detect 78% of dead or declining trees identified during ground inspections, while significantly reducing labor and enabling large-area assessments within a short timeframe. The study covered an area of 6.69 ha with 51 reference trees assessed on the ground. Although the multispectral camera also recorded the red-edge band, it was not included in the present analysis. Compared to traditional ground-based surveys, the UAV-based approach reduced fieldwork time by approx. 20–30% and labor costs by approx. 15–20%. Orthomosaics generated from images captured by commercial multispectral drones (e.g., DJI Mavic 3 Multispectral) provide essential information on tree condition, especially mortality indicators. UAV data collection is fast and relatively low-cost but requires equipment capable of capturing high-resolution imagery in specific spectral bands, particularly near-infrared (NIR). The findings suggest that UAV-based monitoring can enhance the efficiency of large-scale inspections. However, ground-based verification remains necessary in high-traffic areas where safety is critical. Integrating UAV technologies with GIS supports the development of risk management strategies aligned with the principles of precision forestry, enabling sustainable, more proactive and efficient monitoring of tree-related hazards.

Imaging in controlled agriculture helps maximize plant growth by saving labor and optimizing resources. By monitoring specific plant traits, growers can prevent crop losses by correcting environmental conditions that lead to physiological disorders like leaf tipburn. This study aimed to identify morphometric and spectral markers for the early detection of tipburn in two Romaine lettuce (Lactuca sativa) cultivars (‘Chicarita’ and ‘Dragoon’) using an image-based system with color and multispectral cameras. By monitoring tipburn in treatments using melatonin, lettuce cultivars, and with and without supplemental lighting, we enhanced our system’s accuracy for high-resolution tipburn symptom identification. Canopy geometrical features varied between cultivars, with the more susceptible cultivar exhibiting higher compactness and extent values across time, regardless of lighting conditions. These traits were further used to compare simple linear, logistic, least absolute shrinkage and selection operator (LASSO) regression, and random forest models for predicting leaf fresh and dry weight. Random forest regression outperformed simpler models, reducing the percentage error for leaf fresh weight from ~34% (LASSO) to ~13% (RMSE: 34.14 g to 17.32 g). For leaf dry weight, the percentage error decreased from ~20% to ~12%, with an explained variance increase to 94%. Vegetation indices exhibited cultivar-specific responses to supplemental lighting. ‘Dragoon’ consistently had higher red-edge chlorophyll index (CIrededge), enhanced vegetation index, and normalized difference vegetation index values than ‘Chicarita’. Additionally, ‘Dragoon’ showed a distinct temporal trend in the photochemical reflectance index, which increased under supplemental lighting. This study highlights the potential of morphometric and spectral traits for early detection of tipburn susceptibility, optimizing cultivar-specific environmental management, and improving the accuracy of predictive modeling strategies.

Monitoring on irrigated rice field is crucial to anticipate crop damage due to drought and flooding. However, conventional observation methods present complex challenges such as observer bias, point-based data, and limited observational coverage. Observations assisted by multispectral (MS) camera technology mounted on unmanned aerial vehicles (UAVs/drones) were able to address these issues but are still underdeveloped. This study reports the potential use of multispectral cameras mounted on a drone to assess the effects of drought and inundation on rice crops using vegetation indices such as the Normalized Difference Water Index (NDWI), Normalized Difference Vegetation Index (NDVI), and Normalized Difference Red-Edge Index (NDRE). Observations were conducted three times during the vegetative (10 days after transplanting), generative (45 days after transplanting), and ripening (80 days after transplanting) stages on Ciherang and IR64 rice varieties. The rice plants were subjected to different irrigation treatments: drying, flooding, and control. The results showed that the NDRE index had a poor response to vegetation during the vegetative stage but improved during the generative and maturation stages. Meanwhile, the growth of weeds among the rice plants affected the NDVI and NDWI values, especially under dry conditions, making these indices less reliable in such scenarios.

The development of remote sensing technologies using unmanned aerial vehicles (UAVs) and sensors of various measurement methods makes it possible to contactlessly detect various anomalies. Traditional agricultural methods are insufficient to meet the growing demand. The reliability of information after its processing is of primary importance. This reliability is achieved by a combination of measurement methods, the correct choice of contactless measurement tools and the use of machine learning methods for data processing and training on known anomalies. Despite a fairly large selection of tools, there is a limited number of such tools for UAVs. First of all, this is due to the weight and size characteristics of these tools. The most widely used research tools are metal detectors, ground penetrating radars, magnetometers, radiosondes and optical-spectral devices, such as multispectral cameras. The article develops an open architecture that can be used as a basis for building UAVs for a wide range of specific applications. A minimum set of necessary tools has been selected, based on the measurements and processing of which there will be enough information to identify anomalies in the soil with a high degree of reliability. The scope of application of the obtained results is quite extensive: from precision farming, including monitoring crop diseases and chemical non-contact soil analysis to identifying foreign inclusions and anomalies in areas where military actions took place.

This study evaluated the impact of biochar on the growth of strawberry plants, combining visual and proximal sensing monitoring. The plants were rooted in soil enriched with biochar, derived from pyrolysis of soft wood at 550 °C and applied in two doses (2 and 15 g/L), and after physical activation with CO2 at 900 °C; there was also a treatment with no biochar (unaltered). Visual monitoring was based on data logging twice per week of plants’ height and number of flowers and ripe fruits. Proximal sensing monitoring involved a system including a low-cost multispectral camera and a Raspberry Pi 4. The camera acquired nadiral images hourly in three spectral bands (550, 660, and 850 nm), allowing calculation of the normalized difference vegetation index (NDVI). After three months, control plants reached a height of 12.3 ± 0.4 cm, while those treated with biochar and activated biochar grew to 18.03 ± 1.0 cm and 17.93 ± 1.2 cm, respectively. NDVI values were 0.15 ± 0.11 for control plants, increasing to 0.26 ± 0.03 (+78%) with biochar and to 0.28 ± 0.03 (+90%) with activated biochar. In conclusion, biochar application was beneficial for strawberry plants’ growth according to both visual and proximal-sensed measures. Further research is needed to optimize the integration of visual and proximal sensing monitoring, also enhancing the measured parameters.

An important global research topic is the analysis of the crop status, viability, and disease status of vegetables, fruits, and plants in agricultural areas. The Normalized Vegetation Difference Index (NDVI) is commonly used to analyze these conditions by using near-infrared (NIR) features in satellite images or multispectral cameras, such as Lansat-8, to produce NDVI maps. However, these methods have limitations such as high cost and difficulty in accessing images. To address these limitations, this study proposes a new neural network-based index called nNDVI, which uses a Multi-Layer Perceptron (MLP), an Artificial Neural Network (ANN), to convert the NDVI value from standard RGB images. The nNDVI allows for the analysis of vegetation in agricultural areas using low-cost RGB cameras. The MLP model was trained with R (red), G (green), and B (blue) values as input, and real NDVI values for the Swiss forest and Togo farm images were obtained with the MicaSenseAltum camera. The results of testing the model on the dataset showed an accuracy of 92.013% when comparing the nNDVI values obtained with the RGB cameras to the actual NDVI values. Thus, the proposed method demonstrates the ability to use nNDVI maps obtained using low-cost RGB cameras as an alternative to NDVI maps obtained using high-cost multispectral cameras. Overall, this study makes a valuable contribution to the field of agricultural research by presenting a cost-effective and accessible method for analyzing vegetation in agricultural areas.

Situation awareness is essential for ensuring safety in hazardous environments, where timely and accurate information is critical for decision-making. Unmanned Aerial Vehicles (UAVs) have emerged as valuable tools in enhancing situation awareness by providing real-time data and monitoring capabilities in high-risk areas. This study explores the integration of advanced technologies, focusing on imaging and sensor technologies such as thermal, spectral, and multispectral cameras, deployed in critical zones. By merging these technologies into UAV platforms, responders gain access to essential real-time information while reducing human exposure to hazardous conditions. This study presents case studies and practical applications, highlighting the effectiveness of these technologies in a range of hazardous situations.

Several approaches are currently being used by law enforcement to locate the remains of victims. Yet, traditional methods are invasive and time-consuming. Unmanned Aerial Vehicle (UAV)-based remote sensing has emerged as a potential tool to support the location of human remains and clandestine graves. While offering a non-invasive and low-cost alternative, UAV-based remote sensing needs to be tested and validated for forensic case work. To assess current knowledge, a systematic review of 19 peer-reviewed articles from four databases was conducted, focusing specifically on UAV-based remote sensing for human remains and clandestine grave location. The findings indicate that different sensors (colour, thermal, and multispectral cameras), were tested across a range of burial conditions and models (human and mammalian). While UAVs with imaging sensors can locate graves and decomposition-related anomalies, experimental designs from the reviewed studies lacked robustness in terms of replication and consistency across models. Trends also highlight the potential of automated detection of anomalies over manual inspection, potentially leading to improved predictive modelling. Overall, UAV-based remote sensing shows considerable promise for enhancing the efficiency of human remains and clandestine grave location, but methodological limitations must be addressed to ensure findings are relevant to real-world forensic cases.

The sustainable intensification of forage production in Mediterranean climates requires technological solutions that optimize the use of agricultural inputs. This study aimed to evaluate the performance of proximal optical sensors in recommending and monitoring variable rate nitrogen fertilization in winter forage crops cultivated under Mediterranean conditions. A handheld multispectral active sensor (HMA), a multispectral camera on an unmanned aircraft vehicle (UAV), and one passive on-the-go sensor (OTG) were used to generate real-time nitrogen (N) application prescriptions. The sensors were assessed for their correlation with agronomic parameters such as plant fresh matter (PFM), plant dry matter (PDM), plant N content (PNC), crude protein (CP) in%, crude protein yield (CPyield) per unit of area, and N uptake (NUp). The real-time N fertilization stood out by promoting a 15.23% reduction in the total N fertilizer applied compared to a usual farmer-fixed dose of 150 kg ha−1, saving 22.90 kg ha−1 without compromising crop productivity. Additionally, NDVI_OTG showed moderate simple linear correlation with PFM (R2 = 0.52), confirming its effectiveness in prescription based on vegetative vigor. UAV_II (NDVI after fertilization) showed even stronger correlations with CP (R2 = 0.58), CPyield (R2 = 0.53), and NUp (R2 = 0.53), highlighting its sensitivity to physiological responses induced by N fertilization. Although the HMA sensor operates via point readings, it also proved effective, with significant correlations to NUp (R2 = 0.55) and CPyield (R2 = 0.53). It is concluded that integrating sensors enables both precise input prescription and efficient monitoring of plant physiological responses, fostering cost-effectiveness, sustainability, and improved agronomic efficiency.

Abstract The ubiquitous hydration features observed by Zhurong rover provided new insights into Mars aqueous paleoenvironments. However, the impact of the Martian dust was not previously discussed. Here, we conduct a joint analysis of the Multispectral Camera and Short‐Wave Infrared data to constrain the surface composition. The results show that these hydration features are robust against instrumental biases and associated with dusty surfaces. The 1.9 feature is shared between Zhurong and Perseverance landing sites, suggesting that it may be relatively common for Mars dust. The discrepancy between in situ and orbital data could be mainly due to atmospheric effects. The 2.2 band is more specific to the Zhurong landing site, and spectrally consistent with hydrated silica regarding the band shape and position. We propose two possible processes for the origin of such hydrous components at Zhurong landing site, aeolian deposits from nearby cones and/or in situ aqueous alteration products.

Abstract In variety testing and breeding of wheat (Triticum aestivum L.), it is crucial to know the timing of phenological stages and the senescence behavior of genotypes to select for locally adapted varieties. Sound knowledge of the timing of phenological stages also allows for a more meaningful interpretation of measurements such as yield, quality, or disease ratings. In the presence of stresses, only a combined characterization of phenology and environmental conditions can allow for insights into unraveling stress resistance and stress avoidance. Capturing these traits visually in the field is very time‐consuming. Here, a semimobile PhenoCam setup was used to track phenology and senescence from ear emergence to full maturity. PhenoCams mounted on field masts took images of wheat plot trials on a daily basis. In a partial least squares regression analysis, the temporal features of multiple vegetation indices were combined in one model to track phenology and senescence. The method was compared with visual reference methods and repeated drone flights with a multispectral camera. The Pearson's correlation between visual reference methods and PhenoCam predictions was stronger than 0.8, often above 0.9, for most stages. An economic analysis showed that PhenoCams are economically interesting, especially for observing remote experimental sites. Thus, PhenoCams offer a cost‐effective replacement for visual ratings of phenology and senescence, particularly in the context of multienvironment trials.

ABSTRACT In this paper we present a method for measuring the temperature at the onset of melting in the laser-heated diamond anvil cell based on the analysis of speckle dynamics. The method is implemented using a tandem acousto-optic imaging spectrometer and a multispectral camera. We determine the position of points on the melting line of nickel monoaluminide at pressures of 15, 33 and 44 GPa. We show that the experimental points are slightly lower than those calculated using molecular dynamics by the one-phase and two-phase approaches. We consider two variants of the approximation of the experimental points on the phase diagram and assume a polymorphic transition below the melting line. Using the equation of state for nickel monoaluminide, we establish a relationship between the threshold pressure of shock compression at which melting starts and the initial porosity of the samples.

Abstract Mueller Matrix polarimetry (MMP) characterizes changes in light polarization after interacting with a medium, providing insights into tissue microstructure. Combined with multispectral (MS) imaging cameras, MS-MMP offers a novel way to quickly and safely acquire tissue surface information. Machine learning methodologies enable new diagnostic methods by automating tasks on fresh tissue biopsies, though this requires extensive and diverse data. To achieve this, we propose a user-friendly MS-MMP imager with a simple interface and fast acquisition time operated by laboratory technicians and residents. We show that our system, when operated by laboratory staff over several months, yields highquality data in large amounts and with positive feedback of its inclusion in a clinically compliant workflow. This positive outcome is promising for such systems to be used for large data collection initiatives.

Abstract The development of high‐throughput phenotyping platforms to capture time‐series data on large, diverse populations holds promise for crop researchers and breeders investigating growth‐related traits. We used imagery from unoccupied aerial vehicles (UAVs) with red/green/blue (RGB) and multispectral cameras flown over multiple site‐years in Saskatchewan, Canada, and Metaponto, Italy, to gather data for crop height, area, and volume in a lentil diversity panel (324 genotypes). The temporal nature of the UAV image‐derived data enabled the modeling of growth curves for volume, height, and area, something that would be impractical under traditional phenotyping procedures in such a large population grown in multiple environments. A principal component analysis and hierarchical clustering revealed differential growth patterns across contrasting environments, with large variations in temperature and photoperiod, within our lentil diversity panel. Combining this analysis with genome‐wide genotyping data, we identified markers, from an exome capture array (267,845 single nucleotide polymorphisms), associated with crop growth that could be used for marker‐assisted selection. Our study demonstrates the potential for UAV‐based imaging to obtain large‐scale time‐series data across multiple environments to model growth curves and investigate genotype‐by‐environment interactions. In addition, we can now use phenotypic traits that were once impractical to collect and derive novel phenotypes to improve our understanding of crop growth and the genetics underlying adaptation in lentil, approaches that will be useful for both researchers and breeders.