RESUMO

The Amazon forest contains globally important carbon stocks, but in recent years, atmospheric measurements suggest that it has been releasing more carbon than it has absorbed because of deforestation and forest degradation. Accurately attributing the sources of carbon loss to forest degradation and natural disturbances remains a challenge because of the difficulty of classifying disturbances and simultaneously estimating carbon changes. We used a unique, randomized, repeated, very high-resolution airborne laser scanning survey to provide a direct, detailed, and high-resolution partitioning of aboveground carbon gains and losses in the Brazilian Arc of Deforestation. Our analysis revealed that disturbances directly attributed to human activity impacted 4.2% of the survey area while windthrows and other disturbances affected 2.7% and 14.7%, respectively. Extrapolating the lidar-based statistics to the study area (544,300 km2), we found that 24.1, 24.2, and 14.5 Tg C y-1 were lost through clearing, fires, and logging, respectively. The losses due to large windthrows (21.5 Tg C y-1) and other disturbances (50.3 Tg C y-1) were partially counterbalanced by forest growth (44.1 Tg C y-1). Our high-resolution estimates demonstrated a greater loss of carbon through forest degradation than through deforestation and a net loss of carbon of 90.5 ± 16.6 Tg C y-1 for the study region attributable to both anthropogenic and natural processes. This study highlights the role of forest degradation in the carbon balance for this critical region in the Earth system.

Assuntos

RESUMO

Improving our ability to monitor fragmented tropical ecosystems is a critical step in supporting the stewardship of these complex landscapes. We investigated the structural characteristics of vegetation classes in Ucayali, Peru, employing a co-production approach. The vegetation classes included three agricultural classes (mature oil palm, monocrop cacao, and agroforestry cacao plantations) and three forest regeneration classes (mature lowland forest, secondary lowland forest, and young lowland vegetation regrowth). We combined local knowledge with spaceborne lidar from NASA's Global Ecosystem Dynamics Investigation mission to classify vegetation and characterize the horizontal and vertical structure of each vegetation class. Mature lowland forest had consistently higher mean canopy height and lower canopy height variance than secondary lowland forest (µ = 29.40 m, sd = 6.89 m vs. µ = 20.82 m, sd = 9.15 m, respectively). The lower variance of mature forest could be attributed to the range of forest development ages in the secondary forest patches. However, secondary forests exhibited a similar vertical profile to mature forests, with each cumulative energy percentile increasing at similar rates. We also observed similar mean and standard deviations in relative height ratios (RH50/RH95) for mature forest, secondary forest, and oil palm even when removing the negative values from the relative height ratios and interpolating from above-ground returns only (mean RH50/RH95 of 0.58, 0.54, and 0.53 for mature forest, secondary forest, and oil palm, respectively) (p < .0001). This pattern differed from our original expectations based on local knowledge and existing tropical forest succession studies, pointing to opportunities for future work. Our findings suggest that lidar-based relative height metrics can complement local information and other remote sensing approaches that rely on optical imagery, which are limited by extensive cloud cover in the tropics. We show that characterizing ecosystem structure with a co-production approach can support addressing both the technical and social challenges of monitoring and managing fragmented tropical landscapes.

RESUMO

This work reports the development of an efficient and precise indoor positioning system utilizing two-dimensional (2D) light detection and ranging (LiDAR) technology, aiming to address the challenging sensing and positioning requirements of the beyond fifth-generation (B5G) mobile networks. The core of this work is the implementation of a 2D-LiDAR system enhanced by an artificial neural network (ANN), chosen due to its robustness against electromagnetic interference and higher accuracy over traditional radiofrequency signal-based methods. The proposed system uses 2D-LiDAR sensors for data acquisition and digital filters for signal improvement. Moreover, a camera and an image-processing algorithm are used to automate the labeling of samples that will be used to train the ANN by means of indicating the regions where the pedestrians are positioned. This accurate positioning information is essential for the optimization of B5G network operation, including the control of antenna arrays and reconfigurable intelligent surfaces (RIS). The experimental validation demonstrates the efficiency of mapping pedestrian locations with a precision of up to 98.787%, accuracy of 95.25%, recall of 98.537%, and an F1 score of 98.571%. These results show that the proposed system has the potential to solve the problem of sensing and positioning in indoor environments with high reliability and accuracy.

RESUMO

In mobile robotics, LASER scanners have a wide spectrum of indoor and outdoor applications, both in structured and unstructured environments, due to their accuracy and precision. Most works that use this sensor have their own data representation and their own case-specific modeling strategies, and no common formalism is adopted. To address this issue, this manuscript presents an analytical approach for the identification and localization of objects using 2D LiDARs. Our main contribution lies in formally defining LASER sensor measurements and their representation, the identification of objects, their main properties, and their location in a scene. We validate our proposal with experiments in generic semi-structured environments common in autonomous navigation, and we demonstrate its feasibility in multiple object detection and identification, strictly following its analytical representation. Finally, our proposal further encourages and facilitates the design, modeling, and implementation of other applications that use LASER scanners as a distance sensor.

RESUMO

Light Detection and Ranging (LiDAR) technology is positioning itself as one of the most effective non-destructive methods to collect accurate information on ground crop fields, as the analysis of the three-dimensional models that can be generated with it allows for quickly measuring several key parameters (such as yield estimations, aboveground biomass, vegetation indexes estimation, perform plant phenotyping, and automatic control of agriculture robots or machinery, among others). In this survey, we systematically analyze 53 research papers published between 2005 and 2022 that involve significant use of the LiDAR technology applied to the three-dimensional analysis of ground crops. Different dimensions are identified for classifying the surveyed papers (including application areas, crop species under study, LiDAR scanner technologies, mounting platform technologies, and the use of additional instrumentation and software tools). From our survey, we draw relevant conclusions about the use of LiDAR technologies, such as identifying a hierarchy of different scanning platforms and their frequency of use as well as establishing the trade-off between the economic costs of deploying LiDAR and the agronomically relevant information that effectively can be acquired. We also conclude that none of the approaches under analysis tackles the problem associated with working with multiple species with the same setup and configuration, which shows the need for instrument calibration and algorithmic fine tuning for an effective application of this technology.

RESUMO

Building occlusions usually decreases the accuracy of boundary regularization. Thus, it is essential that modeling methods address this problem, aiming to minimize its effects. In this context, we propose a weighted iterative changeable degree spline (WICDS) approach. The idea is to use a weight function for initial building boundary points, assigning a lower weight to the points in the occlusion region. As a contribution, the proposed method allows the minimization of errors caused by the occlusions, resulting in a more accurate contour modeling. The conducted experiments are performed using both simulated and real data. In general, the results indicate the potential of the WICDS approach to model a building boundary with occlusions, including curved boundary segments. In terms of Fscore and PoLiS, the proposed approach presents values around 99% and 0.19 m, respectively. Compared with the previous iterative changeable degree spline (ICDS), the WICDS resulted in an improvement of around 6.5% for completeness, 4% for Fscore, and 0.24 m for the PoLiS metric.

RESUMO

Currently, there are no free databases of 3D point clouds and images for seedling phenotyping. Therefore, this paper describes a platform for seedling scanning using 3D Lidar with which a database was acquired for use in plant phenotyping research. In total, 362 maize seedlings were recorded using an RGB camera and a SICK LMS4121R-13000 laser scanner with angular resolutions of 45° and 0.5° respectively. The scanned plants are diverse, with seedling captures ranging from less than 10 cm to 40 cm, and ranging from 7 to 24 days after planting in different light conditions in an indoor setting. The point clouds were processed to remove noise and imperfections with a mean absolute precision error of 0.03 cm, synchronized with the images, and time-stamped. The database includes the raw and processed data and manually assigned stem and leaf labels. As an example of a database application, a Random Forest classifier was employed to identify seedling parts based on morphological descriptors, with an accuracy of 89.41%.

RESUMO

Autonomous navigation in mining tunnels is challenging due to the lack of satellite positioning signals and visible natural landmarks that could be exploited by ranging systems. Solutions requiring stable power feeds for locating beacons and transmitters are not accepted because of accidental damage risks and safety requirements. Hence, this work presents an autonomous navigation approach based on artificial passive landmarks, whose geometry has been optimized in order to ensure drift-free localization of mobile units typically equipped with lidar scanners. The main contribution of the approach lies in the design and optimization of the landmarks that, combined with scan matching techniques, provide a reliable pose estimation in modern smoothly bored mining tunnels. A genetic algorithm is employed to optimize the landmarks' geometry and positioning, thus preventing that the localization problem becomes ill-posed. The proposed approach is validated both in simulation and throughout a series of experiments with an industrial skid-steer CAT 262C robotic excavator, showing the feasibility of the approach with inexpensive passive and low-maintenance landmarks. The results show that the optimized triangular and symmetrical landmarks improve the positioning accuracy by 87.5% per 100 m traveled compared to the accuracy without landmarks. The role of optimized artificial landmarks in the context of modern smoothly bored mining tunnels should not be understated. The results confirm that without the optimized landmarks, the localization error accumulates due to odometry drift and that, contrary to the general intuition or belief, natural tunnel features alone are not sufficient for unambiguous localization. Therefore, the proposed approach ensures grid-based SLAM techniques can be implemented to successfully navigate in smoothly bored mining tunnels.

RESUMO

Elevation mapping at ground level is challenging in forested areas like the Amazon region, which is mostly covered by dense rainforest. The most common techniques, i.e. photogrammetry and short wavelength radar, provide elevations at canopy level at best, while most applications require ground elevations. Even lidar and P-band radar, which can penetrate foliage and measure elevations at ground level, have some limitations which are analyzed in here. We address three research questions: To what extent can a terrain model be replaced by a more easily available canopy-level surface model for topography-based applications? How can the elevation be obtained at ground level through forest? Can a priori knowledge of general continental relief properties be used to compensate for the limits of measurement methods in the presence of forest?(AU)

O mapeamento da elevação ao nível do solo é um desafio em áreas florestadas como a região amazônica, coberta principalmente por floresta tropical densa. As técnicas mais comuns, i.e., a fotogrametria e o radar de comprimento de onda curto, fornecem elevações ao nível do dossel na melhor das hipóteses, enquanto a maioria das aplicações requer a elevação do solo. Mesmo o lidar e o radar de banda P, que podem penetrar a folhagem e medir elevações ao nível do solo, têm algumas limitações que são analisadas aqui. Abordamos três questões: Até que ponto um modelo de terreno pode ser substituído por um modelo de superfície ao nível do dossel, mais facilmente disponível, para aplicações baseadas na topografia? Como a elevação ao nível do solo pode ser obtida através da floresta? O conhecimento a priori das propriedades gerais do relevo continental pode ser usado para compensar os limites dos métodos de medição na presença de floresta?(AU)

Assuntos

RESUMO

Mobile robots must be capable to obtain an accurate map of their surroundings to move within it. To detect different materials that might be undetectable to one sensor but not others it is necessary to construct at least a two-sensor fusion scheme. With this, it is possible to generate a 2D occupancy map in which glass obstacles are identified. An artificial neural network is used to fuse data from a tri-sensor (RealSense Stereo camera, 2D 360° LiDAR, and Ultrasonic Sensors) setup capable of detecting glass and other materials typically found in indoor environments that may or may not be visible to traditional 2D LiDAR sensors, hence the expression improved LiDAR. A preprocessing scheme is implemented to filter all the outliers, project a 3D pointcloud to a 2D plane and adjust distance data. With a Neural Network as a data fusion algorithm, we integrate all the information into a single, more accurate distance-to-obstacle reading to finally generate a 2D Occupancy Grid Map (OGM) that considers all sensors information. The Robotis Turtlebot3 Waffle Pi robot is used as the experimental platform to conduct experiments given the different fusion strategies. Test results show that with such a fusion algorithm, it is possible to detect glass and other obstacles with an estimated root-mean-square error (RMSE) of 3 cm with multiple fusion strategies.

Assuntos