RGB Sensor Research Articles

In-situ ergonomics assessment of industrial manual tasks through AR-headset integrated sensor-systems

This paper presents an approach that is based on the use of sensors integrated into Augmented Reality (AR) headsets to acquire data for in-situ ergonomics assessment of industrial tasks. In the context of human factors evaluation, it is typically required to collect data related to anthropometric characteristics of the operators e.g. height, length of arms, width of shoulders etc., as well as data related to the execution of manufacturing tasks including working postures, task repetition, task duration, and so forth. To this end, it is needed to develop a plug and play system to collect ergonomics data from the field. The proposed approach relies on head-mounted AR devices and their integrated RGB and Depth sensors, image and point cloud processing, and skeleton tracking techniques to calculate the required data, which are later visualized in a user-friendly interface. The proposed approach was tested in a case study where the objective was to identify the proper mechanical adjustments of an exoskeleton on the spot for comfortable fitting for each operator. Five operators participated in the testing that was held under varying lighting conditions. The solution showcased improved efficiency in decision-making compared to the commonplace procedures for skeleton fitting, faster setup time compared to typical camera-based approaches, and satisfactory performance.

An ICT-based highly fluorescent isoquinoline scaffold for selective Hg(II) detection in real-water samples: Development of a smart, low-cost RGB-Arduino electronic platform

Selective detection and quantification of Hg2+ ions is crucial to minimize health and environmental risks. Fluorescent organic small-molecule probes have been expeditiously utilized owing to their unique set of improved properties. However, isoquinoline core has not been extensively explored as a fluorescence platform partly due to synthetic challenges. Herein, a serendipitously discovered synthetic route to access a small yet highly functionalized novel isoquinoline-based probe, IQ is reported. The synthesis is achieved through the in-situ generation of ammonia, followed by intermolecular [5C + 1 N] aza-annulation reaction with a ketendithioacetal-based precursor, P-IQ. IQ displayed excellent recognition ability towards Hg2+ ions in H2O:ACN (99:1, v/v) via ICT-off fluorescent quenching behavior. Comparative FT-IR, 1H/13C NMR, mass spectral studies, and DFT analyses were carried out to validate the suggested mechanisms. Reversible studies confirm the secondary recognition effect of in-situ generated (IQ + Hg2+) complex on cysteine. The binding constant and LOD were estimated to be 3.7 × 104 M−1 and 0.86 µM, respectively. Further, IQ was utilized to evaluate the mercury ion content in real water samples demonstrating its effectiveness in water quality monitoring. The practical utility of IQ was further explored by developing TLC strips, Whatman filter-paper strips, and a low-cost, portable Arduino-based platform. Arduino microcontroller is interfaced with an RGB sensor to detect color changes and quantify mercury concentration w.r.t. RGB values.

The Detection of Tree of Heaven (Ailanthus altissima) Using Drones and Optical Sensors: Implications for the Management of Invasive Plants and Insects

Tree of heaven (Ailanthus altissima) is a highly invasive tree species in the USA and the preferred host of an invasive insect, the spotted lanternfly (Lycorma delicatula). Currently, pest managers rely solely on ground surveys for detecting both A. altissima and spotted lanternflies. This study aimed to develop efficient tools for A. altissima detection using drones equipped with optical sensors. Aerial surveys were conducted to determine the optimal season, sensor type, and flight altitudes for A. altissima detection. The results revealed that A. altissima can be detected during different seasons and at specific flight heights. Male inflorescences were identifiable using an RGB sensor in the spring at <40 m, seed clusters were identifiable in summer and fall at <25 m using an RGB sensor, and remnant seed clusters were identifiable in the winter at <20 m using RGB and thermal sensors. Combining all seasonal data allowed for the identification of both male and female A. altissima. This study suggests that employing drones with optical sensors can provide a near real-time and efficient method for A. altissima detection. Such a tool has the potential to aid in the development of effective strategies for monitoring spotted lanternflies and managing A. altissima.

An open-source machine-learning application for predicting pixel-to-pixel NDVI regression from RGB calibrated images

The Normalized Difference Vegetation Index (NDVI) is the most common index to measure vegetation in agriculture and create classified prescription maps used for several purposes. It is obtained as the ratio between information from the visible and near infrared spectral bands. The calculation of NDVI often uses very expensive hyperspectral and multispectral optical sensors. This study aims to develop an efficient model to extract NDVI data from RGB images. During the work were acquired images, of different plant species at different vegetation status, through a snapshot hyperspectral camera (Specim IQ) featuring natively superimposed RGB sensor whose images were calibrated a posteriori (sRGB). NDVI was predicted from sRGB images using a Shallow-regressive neural network that performs a pixel-pixel regression. The model has been then tested on around 1000 drone images acquired by a 6x Sentera sensor to test the efficiency of the applied model. The model has been used to estimate the numerical value of NDVI and the classes produced by a clustering method (k-means). The results of this work showed that the calibrated images had a very high correlation with the NDVI in validation (r = 0.91), maintaining good performances (r = 0.71) when applied to a set of data acquired with a different non-co-registered sensor. Therefore, for the first time this work has shown that the health of vegetation through the NDVI could be calculated using unexpensive RGB device adopting a pixel-pixel regression AI approach. The approach shows its importance especially for small-sized farms where profit would not allow for the budget to access multispectral cameras and heavy carriers (i.e., UAV).

Novel Multi-View RGB Sensor for Continuous Motion Analysis in Kinetic Chain Exercises: A Pilot Study for Simultaneous Validity and Intra-Test Reliability.

As the number of musculoskeletal disorders caused by smartphone usage, sedentary lifestyles, and active sports activities increases, there is a growing demand for precise and accurate measurement and evaluation of issues such as incorrect compensation patterns, asymmetrical posture, and limited joint operation range. Urgent development of new inspection equipment is necessary to address issues such as convenience, economic feasibility, and post-processing difficulties. Using 4DEYE®, a new multi-view red, green, and blue (RGB) sensor-based motion analysis equipment, and the VICON® ratio, which are infrared-based markers, we conducted a comparative analysis of the simultaneous validity of the joint angle (trajectory) and reliability. In this study, five healthy participants who could perform movements were selected for the pilot study and two movements (Y-balance and side dip) were analyzed. In addition, the ICC (Intraclass Correlation Coefficient) was analyzed using the SPSS (Statistical Package for the Social Sciences) V.18 while the number of data frames of each equipment was equalized using the MATLAB program. The results revealed that side dips, which are open kinetic chain exercises (intraclass correlation coefficient ICC(2.1), 0.895-0.996), showed very high concordance with the Y-balance test, a closed kinetic chain exercise (ICC(2.1), 0.678-0.990). The joint measurement results were similar regardless of the movement in the open or closed kinetic chain exercise, confirming the high reliability of the newly developed multiview RGB sensor. This is of great significance because we obtained important and fundamental results that can be used in various patterns of exercise movements in the future.

Enhancing leaf area index and biomass estimation in maize with feature augmentation from unmanned aerial vehicle-based nadir and cross-circling oblique photography

Rapid and accurate estimation of plant phenotypes plays a vital role in effective breeding and management of maize crops. Unmanned aerial vehicle (UAV) platforms are emerging as a valuable tool in the assessment of crop phenotypes, offering a promising avenue to enhance the accuracy and efficiency of phenotypic analysis. This study executed two UAV photography methods to acquire aerial images for the estimation of leaf area index (LAI) and above-ground biomass (AGB) in summer maize. Nadir photography was implemented to obtain data from three sources including multi-spectral (MS), RGB, and thermal infrared (TIR) sensors. Additionally, RGB imaging-based cross-circling oblique (CCO) photography was implemented to obtain accurate 3D point cloud data. The nadir photography data was processed to generate orthomosaic image and extract features including vegetation index, canopy cover, canopy height, canopy temperature, and textural information. Features including canopy occupation volume, plant area index, canopy cover, and canopy height were extracted from the CCO photography-derived point cloud data. Furthermore, a data augmentation method, called linear regression-based feature augmentation (LRFA), was proposed for augmenting features extracted from nadir and CCO photography. The results showed that the introduction of CCO photography improved the accuracy of LAI and AGB estimation during the big trumpet stage and LAI estimation during the milk stage compared to the integrated multi-source nadir photography data. Notably, the LRFA derived features outperformed raw features in the majority of modeling scenarios, with the random forest achieved an average accuracy (R2) improvement of 6.1% for LAI estimation and 3.7% for AGB estimation. This study highlights the significance of combining different photography technologies and feature augmentation method for estimating maize phenotypes, providing novel opportunities for crop growth monitoring in modern agriculture.

A Remote Sensing Approach for Assessing Daily Cumulative Evapotranspiration Integral in Wheat Genotype Screening for Drought Adaptation.

This study considers critical aspects of water management and crop productivity in wheat cultivation, specifically examining the daily cumulative actual evapotranspiration (ETa). Traditionally, ETa surface energy balance models have provided estimates at discrete time points, lacking a holistic integrated approach. Field trials were conducted with 22 distinct wheat varieties, grown under both irrigated and rainfed conditions over a two-year span. Leaf area index prediction was enhanced through a robust multiple regression model, incorporating data acquired from an unmanned aerial vehicle using an RGB sensor, and resulting in a predictive model with an R2 value of 0.85. For estimation of the daily cumulative ETa integral, an integrated approach involving remote sensing and energy balance models was adopted. An examination of the relationships between crop yield and evapotranspiration (ETa), while considering factors like year, irrigation methods, and wheat cultivars, unveiled a pronounced positive asymptotic pattern. This suggests the presence of a threshold beyond which additional water application does not significantly enhance crop yield. However, a genetic analysis of the 22 wheat varieties showed no correlation between ETa and yield. This implies opportunities for selecting resource-efficient wheat varieties while minimizing water use. Significantly, substantial disparities in water productivity among the tested wheat varieties indicate the possibility of intentionally choosing lines that can optimize grain production while minimizing water usage within breeding programs. The results of this research lay the foundation for the development of resource-efficient agricultural practices and the cultivation of crop varieties finely attuned to water-scarce regions.

Utility of Spectral Filtering to Improve the Reliability of Marine Fauna Detections from Drone-Based Monitoring.

Monitoring marine fauna is essential for mitigating the effects of disturbances in the marine environment, as well as reducing the risk of negative interactions between humans and marine life. Drone-based aerial surveys have become popular for detecting and estimating the abundance of large marine fauna. However, sightability errors, which affect detection reliability, are still apparent. This study tested the utility of spectral filtering for improving the reliability of marine fauna detections from drone-based monitoring. A series of drone-based survey flights were conducted using three identical RGB (red-green-blue channel) cameras with treatments: (i) control (RGB), (ii) spectrally filtered with a narrow 'green' bandpass filter (transmission between 525 and 550 nm), and, (iii) spectrally filtered with a polarising filter. Video data from nine flights comprising dolphin groups were analysed using a machine learning approach, whereby ground-truth detections were manually created and compared to AI-generated detections. The results showed that spectral filtering decreased the reliability of detecting submerged fauna compared to standard unfiltered RGB cameras. Although the majority of visible contrast between a submerged marine animal and surrounding seawater (in our study, sites along coastal beaches in eastern Australia) is known to occur between 515-554 nm, isolating the colour input to an RGB sensor does not improve detection reliability due to a decrease in the signal to noise ratio, which affects the reliability of detections.

3차원 공간 스캔을 위한 ToF-Stereo 융합 센서 시스템 설계

In this paper, we propose a ToF-Stereo fusion sensor system for 3D space scanning that increases the recognition rate of 3D objects, guarantees object detection quality, and is robust to the environment. The ToF-Stereo sensor fusion system uses a method of fusing the sensing values of the ToF sensor and the Stereo RGB sensor, and even if one sensor does not operate, the other sensor can be used to continuously detect an object. Since the quality of the ToF sensor and the Stereo RGB sensor varies depending on the sensing distance, sensing resolution, light reflectivity, and illuminance, a module that can adjust the function of the sensor based on reliability estimation is placed. The ToF-Stereo sensor fusion system combines the sensing values of the ToF sensor and the Stereo RGB sensor, estimates the reliability, and adjusts the function of the sensor according to the reliability to fuse the two sensing values, thereby improving the quality of the 3D space scan.

Leaf area index and aboveground biomass estimation of an alpine peatland with a UAV multi-sensor approach

ABSTRACT Aboveground biomass (AGB) can serve as an indicator when estimating various biogeochemical processes in peatlands, an ecosystem which provides countless ecosystem services and plays a key role in climate regulation. While remote sensing has been extensively employed to assess AGB across vast areas in forested peatlands its application to small and treeless peatlands, which are typical of the alpine regions, has been limited. Due to the characteristics of peatlands, innovative approaches capable of capturing their fine-scale, highly heterogeneous and short-stature vegetation cover are needed. Likewise, other key requirements include an ability to overcome site accessibility barriers, cost-effective acquisition of datasets and minimizing damage of these protected habitats. Hence, the utilization of Unmanned Aerial Vehicles (UAVs) offers a viable means for mapping AGB in alpine peatlands. In this study, the AGB of the Val di Ciampo alpine peatland (Veneto Region, Italy) was estimated by combining datasets derived from in situ vegetation samples as well as UAV-based LiDAR, hyperspectral and RGB sensors. A limited number of vegetation samples were used to reduce the impact of the study on the ecosystem. The results indicate that a linear regression can model the relationship between AGB and Leaf Area Index (LAI) with a significant explanatory ability (R2 = 0.72; p < 0.001). Several indices derived from digital terrain model (DTM) morphologies, hyperspectral data, and orthophotos were tested using a multiple regression approach to determine their potential to enhance the model’s performance. Among these only the Double Difference (DD) index, derived from hyperspectral data, was found to slightly improve the model’s explanatory ability (R2 = 0.76). Overall, the findings of this study suggest that UAV LiDAR data provides the most reliable solution for estimating AGB in alpine peatlands, while the inclusion of hyperspectral data provides only a minor improvement in accuracy.

Addressing the sample volume dependency of the colorimetric glucose measurement on microfluidic paper-based and thread/paper-based analytical devices using a novel low-cost analytical viewpoint

The colorimetric method is widely exploited for glucose measurement on microfluidic paper-based and microfluidic thread/paper-based analytical devices, μPADs and μTPADs, respectively. However, two significant challenges still hinder the real-world applications: the variation of the generated color based on sample volume variation, and the consistent dependency on an imaging camera - usually a smartphone camera - for color readout. The latter also suffers from multiple limitations, such as the variable ambient light conditions and imaging variation among different smartphone brands and models. This manuscript suggests a new device demonstrating a novel viewpoint to the colorimetric method to address both aforementioned challenges. The presented device, Volume-Independent Autonomous box (VIA box), continuously monitors the transient color of the µPAD upon the introduction of the sample till color stabilization, in contrast to similar studies where merely the final stabilized color is read. The interpretation of the transient color profile with respect to time results in the measurement of the glucose level independent of sample volume. In addition, the VIA box monitors µPAD’s color using an ordinary RGB sensor instead of an imaging camera in similar studies. Since merely RGB values are recorded in any measurement instance, the computational cost is extremely low compared to the relevant literature where image processing techniques are used. VIA box exploits a simple low-end microcontroller to continuously monitor the transient color, interpret the color profile, and show the glucose level. The samples used in this manuscript are made of artificial sweat samples, including the known glucose range in real sweat samples.

Remotely sensed and ground measurements reveal intraspecific differences in early season needle unfolding and senescence, but lack of variability in litter flammability of Pinus halepensis

Pinus halepensis Mill. is a conifer typical of Mediterranean pinewoods adapted to drought and fire. Although the timing of seasonal events in P. halepensis is strongly associated with these stressors and other environmental factors, needle phenology and litter flammability are insufficiently characterized at the intraspecific level. These traits can be informative of the existence of locally adapted populations across the species’ distribution range, and remote sensing approaches are promising tools to infer phenological changes in forest trees. In this study, we investigated intraspecific differentiation of P. halepensis related to needle phenology at the onset of the fire season (May-June) through vegetation indexes (VIs) obtained from unmanned aerial vehicles. We collected drone images using multispectral and RGB sensors for 56 adult populations of P. halepensis categorized into five ecotypes growing in two common gardens located in Spain under contrasting conditions (dry-continental versus wet-coastal). In the dry trial, we additionally monitored the temporal variation of RGB-derived indexes using two flights spaced over a one-month period. We also performed four consecutive ground measurements of needle pigments and fuel moisture content and obtained litter flammability traits for a subset of populations. Most in situ measurements and flammability traits did not show population differentiation. Regarding RGB-derived VIs, we did not detect obvious temporal patterns which differed among populations. However, we observed greener canopies in June than in May, which are indicative of the pace of current-year needle development. We also deduced an earlier phenology (i.e., earlier current-year needle unfolding) in the dry-continental trial. Ecotypic differentiation was found for some vegetation indexes related to needle unfolding (i.e., TCARI/OSAVI) and old needle senescence (i.e., PRSI). These differences were generally consistent across trials and time, and fundamentally indicated that sub-humid ecotypes typical of the eastern Mediterranean showed earlier needle unfolding and old needle senescence than semiarid ecotypes thriving in western continental Mediterranean areas. The intraspecific divergence observed in early season phenology is potentially related to the existence of contrasting life-history strategies present at the intraspecific level for the species. In particular, some semi-arid ecotypes, which are known to optimize the trade-off between carbon gain and water loss, exhibited an early old needle senescence (high likelihood of crown fire development) coupled with a late needle unfolding (conservative water use). These results indicate a possible trade-off between drought and fire resistance for the species, which suggest the importance of considering the intraspecific characteristics of Aleppo pine in forest management actions.

Multi-modal sorting in plastic and wood waste streams

Addressing the escalating waste crisis necessitates innovative waste management strategies, particularly valorisation techniques, the efficiency of which is dictated by the purity of the feedstock. In order to mitigate the segregation challenges encountered in complex and non-homogeneous waste streams, this work proposes a vision-based architecture ample for effective sorting of parts based on shape and material-related properties. The proposed work encapsulates a novel deep learning multi-modal approach, in which multiple parallel auto-encoders are used to extract spatio-spectral information from the RGB and multi-spectral sensors and project them in a common latent space. By decoding the latent space representations, the class of each object is picked out, thus guiding the robotic sub-system accordingly. To support the proposed deep architecture, a dataset, called Multispectral Mixed Waste Dataset (MMWD) was produced, containing multi-spectral data from the visible (16 bands), near-infrared (25 bands) regions of the electromagnetic spectrum and RGB (3 bands) data. The dataset includes the following seven plastic and wood wastes: Polypropylene (PP), PolyEthylene Terephthalate (PET), Low-Density PolyEthylene (LDPE), High-Density PolyEthylene (HDPE), Medium Density Fibreboard (MDF), Melamine Faced Chipboards (MFC), and Oak Veneer samples. For the localisation of waste along the conveyor belt, YOLO v8 is used to achieve 99.5% mean average precision (mAP50). In the classification task, where the multi-modal approach was followed, the overall accuracy achieved is 96% with the prediction recall being greater than 95% for the majority of classes under examination.

Nearly monodisperse ZnS nanoparticles for portable recognition, enrichment and detection of Cd(II)

As cadmium ions (Cd2+) pose a serious threat to environment and human health, it is significant to develop portable, fast-response and sensitive Cd2+ sensors for both real-time and on-site collection of such critical information. In this study, nearly monodisperse ZnS nanoparticles (NPs) were prepared and used to recognize, enrich and detect Cd2+ ions. ZnS NPs demonstrated fast adsorption kinetics towards Cd2+ ions (removing 99.3 % of Cd2+ in 12 min), surpassing previously reported multiple Cd2+ adsorption materials. Upon adsorbing Cu2+ and Cd2+ ions via substitution reaction, ZnS NPs gain strong peroxidase activity, and demonstrate the ability to selectively detect Cu2+ and Cd2+ ions among common metal ions. By further conjugating with our previous HSnS NFs (Hollow SnS nanoflower) nanosensor to eliminate the interference of Cu2+, the ZnS NPs nanosensor could selectively detect Cd2+ ions with recoveries above 80 %. The ZnS NPs nanosensor can cover a wide range of Cd2+ concentration from 0.5 to 100 μg L−1 with an extremely low detection limit of 0.38 μg L−1. When coupling to a portable RGB sensor, as low as 0.5 μg L−1 Cd2+ ions could be reached on-site for river water, giving an outstanding recovery of 99.4 %. The current results provide a new avenue to the on-site detection of Cd2+ ions, and would doubtlessly contribute to the future construction of portable and intelligent sensor arrays towards complex environmental samples.

Differentiating nutritional and water statuses in Hass avocado plantations through a temporal analysis of vegetation indices computed from aerial RGB images

Maximizing crop production efficiently and sustainably through plant health monitoring is key for global food security. Monitoring large areas with remote sensing technologies such as unmanned aerial vehicles (UAVs) with sensors deals with time and money issues; however, the usage of advanced sensors such as hyperspectral, multispectral and thermal cameras limit their usage among all the stakeholders. In this study we explore different vegetation indices (VIs) extracted from aerial RGB images acquired in different flights to differentiate the nutritional and water statuses of Hass avocado plantations. We used an image processing workflow consisting of image selection through a convolutional neural network (CNN) model, tree crown segmentation, color correction and feature extraction to automate the computation of VIs from RGB images. To compare the performance of VIs in the differentiation of nutritional and water statuses, we proposed a comparison metric called Mean Distance between Vegetation Indices (MDVI), analyzed the evolution of the extracted features, and studied their relationships with gold standard Normalized Difference Vegetation Index (NDVI) measurements. Since the extracted features from each group vary from flight to flight due to multiple factors such as the light intensity of each season and the phenological stage of the plant, the proposed comparison metric leverages the differences between the features extracted from each group, thus reducing these temporal effects. We found that Modified Green Red Vegetation Index (MGRVI) allows a better differentiation of nutritional and water statuses. Furthermore, the correlation coefficients of this VI in the three statuses and NDVI for nitrogen group range between 0.63 and 0.85, indicating a positive strong relationship. The results of this work show that MGRVI has a potential to be used as a correlation variable in studies that only use RGB sensors in order to monitor the nutritional and water status of crops.

Measuring visual and non-visual lighting metrics in building environments with RGB sensors

Today, a spectroradiometer can measure both visual lighting metrics (illuminance, chromaticity coordinates or color temperature) and non-visual lighting metrics (Circadian Stimulus Model (CS2021) by M. Rea et al. Melanopic Equivalent Daylight (D65) Illuminance (mEDI) in CIE S 026/E: 2018) very well. Unfortunately, their cost and size make them unsuitable for practical measurements after installation. And some commercially available miniature RGB sensors can measure some visual parameters, but the question is whether they can accurately determine both visual and non-visual characteristics for practical evaluation or real-time luminaire control. Therefore, in this work, the mixed light between incident daylight and typical LED (500 lx horizontal illuminance, 4000 K) in a typical office was measured using the CSS45 spectrometer and the AS 73211 RGB sensor. Then 410/516 spectra were used to process the RGB sensor. And the remaining 20% were used to verify the obtained correlations. The correlation coefficient R2 of 0.892 and the RMSE value of 15.436 lx for the relationship between the mEDI values of the RGB sensor and the spectrophotometer and their average color difference Δu'v' of 3.78.10−5 were very good evidence for the processing quality. For verification, the correlation quality parameter pair (R2/RMSE) of CS2021 and Ev also reach the very good values of 0.842/0.015 and 0.78/29.15, respectively. Based on these achievements, it can be confirmed that a commercially available RGB sensor can measure the visual and non-visual parameters of lighting systems accurately enough after installation if it is properly processed as recommended in this work.

Learning based compressive snapshot spectral light field imaging with RGB sensors.

The application of multidimensional optical sensing technologies, such as the spectral light field (SLF) imager, has become increasingly common in recent years. The SLF sensors provide information in the form of one-dimensional spectral data, two-dimensional spatial data, and two-dimensional angular measurements. Spatial-spectral and angular data are essential in a variety of fields, from computer vision to microscopy. Beam-splitters or expensive camera arrays are required for the usage of SLF sensors. The paper describes a low-cost RGB light field camera-based compressed snapshot SLF imaging method. Inspired by the compressive sensing paradigm, the four dimensional SLF can be reconstructed from a measurement of an RGB light field camera via a network which is proposed by utilizing a U-shaped neural network with multi-head self-attention and unparameterized Fourier transform modules. This method is capable of gathering images with a spectral resolution of 10 nm, angular resolution of 9 × 9, and spatial resolution of 622 × 432 within the spectral range of 400 to 700 nm. It provides us an alternative approach to implement the low cost SLF imaging.

Artichoke deep learning detection network for site-specific agrochemicals UAS spraying

Input optimization is a distinguishing characteristic of Precision Agriculture approaches, helping reduce the environmental impact and costs and increase vegetable production quality. Thanks to the high automation evolution of Unmanned Aerial Systems (UAS), a new approach derived from their combination with Deep Learning techniques is leading to significant improvements in agricultural management practices. The study aims at artichoke plants detection and georeferencing as a first step for an on-the-fly, real time, UAS spraying system, and use the gathered information to monitor crop development through a multi-temporal approach. A commercial UAS, equipped with an RGB sensor, acquired images of the artichoke field located in Sardinia (Italy) during the 2021–2022 season in different crop growth stages. The FPN (Feature Pyramid Network), trained and compared with the YOLOv5 (You Only Look Once) network, showed a high detection level with an average F1 score of around 90%, and satisfactory off-line performances on the Nvidia Jetson Nano board. YOLOv5 achieved the best overall result. The FPN recorded a lower recall, which is desirable to achieve a minimum number of detection errors and limit the leakage of agrochemicals on false-positive targets. The multi-temporal approach influenced detection performances, with an inverse response of precision and recall metrics. The growing index trend showed a distinct value in October, peaking at the beginning of December as expected. The proposed approach contributes to designing future automatic and reliable site-specific UAS agrochemicals application and the classification of management zones.

Rapid monitoring of tea plants under cold stress based on UAV multi-sensor data

In the context of climate change, extreme weather events, represented by frost injury, are increasingly having a negative impact on the growth of tea plants. This has brought huge losses to the tea industry. Traditionally, the freezing injury of tea plants in the field was assessed by vision. This is labor-intensive and subjective. In this research, multimodal remote sensing data from different periods of natural overwintering tea plantations were collected by using unmanned aerial vehicles (UAV) equipped with multispectral (MS), thermal infrared (TIR) and RGB sensors. And the physiological data of tea leaves on the same day were obtained to construct a tea cold injury score (TCIS). Then, a convolutional neural networks-gate recurrent unit (CNN-GRU) model was improved for estimating TCIS. To better compare the performance of CNN-GRU, a single GRU model and three classical machine learning models were also used for comparison. The study found that: (1) The multimodal data fusion was superior to the unimodal data. The best prediction results were achieved for the combined bimodal MS + RGB data (Rp2 = 0.862, RMSEP = 0.138, RPD = 2.220); (2) The CNN-GRU hybrid model was superior to the other four baseline models. The best effect was achieved based on the multivariate input of MS + RGB (Rp2 = 0.862) or MS + RGB + TIR (Rp2 = 0.850); (3) The accuracy of the model after removing soil features was lower than that of the model without background removal. Therefore, the TCIS-CNN-GRU model combined with multi-source remote sensing data can objectively and accurately evaluate the cold injury phenotype of tea plants, making the CNN-GRU model more scientific and promising.

A Comparison of Monocular Visual SLAM and Visual Odometry Methods Applied to 3D Reconstruction

Pure monocular 3D reconstruction is a complex problem that has attracted the research community’s interest due to the affordability and availability of RGB sensors. SLAM, VO, and SFM are disciplines formulated to solve the 3D reconstruction problem and estimate the camera’s ego-motion; so, many methods have been proposed. However, most of these methods have not been evaluated on large datasets and under various motion patterns, have not been tested under the same metrics, and most of them have not been evaluated following a taxonomy, making their comparison and selection difficult. In this research, we performed a comparison of ten publicly available SLAM and VO methods following a taxonomy, including one method for each category of the primary taxonomy, three machine-learning-based methods, and two updates of the best methods to identify the advantages and limitations of each category of the taxonomy and test whether the addition of machine learning or updates on those methods improved them significantly. Thus, we evaluated each algorithm using the TUM-Mono dataset and benchmark, and we performed an inferential statistical analysis to identify the significant differences through its metrics. The results determined that the sparse-direct methods significantly outperformed the rest of the taxonomy, and fusing them with machine learning techniques significantly enhanced the geometric-based methods’ performance from different perspectives.