Search

The adoption of 3D survey techniques is essential to promote efficient and timely information acquisition on constructed buildings. This article addresses terrestrial LiDAR (TLS) and close-range photogrammetric data fusion for the 3D modeling of a building in Level of Detail (LoD) 3. The selected building presents challenging elements for modeling, such as extended curved slabs, external glass walls, recessed facades and diverse roof pitches. It is located on the campus of the Federal University of Paraná (UFPR) in Curitiba, Brazil. The accuracy of the data integration was obtained through the analysis of deviations between the clouds of primary points. The accuracy of the point cloud model was verified by comparing its dimensions with the real dimensions of the building, obtained by means of a handheld laser distance meter (EDM). The results demonstrate that there was a correspondence between the EDM measures and the model, with a satisfactory statistical agreement between the estimated and reference values and a general maximum absolute error of 4.5 cm. The article focuses on the accuracy of point cloud models for the cadastral updating of buildings, providing information for decision making in projects documentation and interventions. [ABSTRACT FROM AUTHOR]

The classification of tree species can significantly benefit from high spatial and spectral information acquired by unmanned aerial vehicles (UAVs) associated with advanced feature extraction and classification methods. Different from the traditional feature extraction methods, that highly depend on user's knowledge, the convolutional neural network (CNN)-based method can automatically learn and extract the spatial-related features layer by layer. However, in order to capture significant features of the data, the CNN classifier requires a large number of training samples, which are hardly available when dealing with tree species in tropical forests. This study investigated the following topics concerning the classification of 14 tree species in a subtropical forest area of Southern Brazil: i) the performance of the CNN method associated with a previous step to increase and balance the sample set (data augmentation) for tree species classification as compared to the conventional machine learning methods support vector machine (SVM) and random forest (RF) using the original training data; ii) the performance of the SVM and RF classifiers when associated with a data augmentation step and spatial features extracted from a CNN. Results showed that the CNN classifier outperformed the conventional SVM and RF classifiers, reaching an overall accuracy (OA) of 84.37% and Kappa of 0.82. The SVM and RF had a poor accuracy with the original spectral bands (OA 62.67% and 59.24%) but presented an increase between 14% and 21% in OA when associated with a data augmentation and spatial features extracted from a CNN. [ABSTRACT FROM AUTHOR]

Synthetic aperture radar (SAR) image processing and analysis rely on statistical modeling and parameter estimation of the probability density functions that characterize data. The method of log-cumulants (MoLC) is a reliable alternative for parameter estimation of SAR data models and image processing. However, numerical methods are usually applied to estimate parameters using MoLC, and it may lead to a high computational cost. Thus, MoLC may be unsuitable for real-time SAR imagery applications such as change detection and marine search and rescue, for example. Our paper introduces a fast approach to overcome this limitation of MoLC, focusing on parameter estimation of single-channel SAR data modeled by the G0I distribution. Experiments with simulated and real SAR data demonstrate that our approach performs faster than MoLC, while the precision of the estimation is comparable with that of the original MoLC. We tested the fast approach with multitemporal data and applied the arithmetic-geometric distance to real SAR images for change detection on the ocean. The experiments showed that the fast MoLC outperformed the original estimation method with regard to the computational time. [ABSTRACT FROM AUTHOR]

Knowing the spatial distribution of endangered tree species in a forest ecosystem or forest remnants is a valuable information to support environmental conservation practices. The use of Unmanned Aerial Vehicles (UAVs) offers a suitable alternative for this task, providing very high-resolution images at low costs. In parallel, recent advances in the computer vision field have led to the development of effective deep learning techniques for end-to-end semantic image segmentation. In this scenario, the DeepLabv3+ is well established as the state-of-the-art deep learning method for semantic segmentation tasks. The present paper proposes and assesses the use of DeepLabv3+ for mapping the threatened Dipteryx alata Vogel tree, popularly also known as cumbaru. We also compare two backbone networks for feature extraction in the DeepLabv3+ architecture: the Xception and MobileNetv2. Experiments carried out on a dataset consisting of 225 UAV/RGB images of an urban area in Midwest Brazil demonstrated that DeepLabv3+ was able to achieve in mean overall accuracy and F1-score above 90%, and IoU above 80%. The experimental analysis also pointed out that the MobileNetv2 backbone overcame its counterpart by a wide margin due to its comparatively simpler architecture in view of the available training data. [ABSTRACT FROM AUTHOR]

Google Earth Engine (GEE) platform is an online tool, which generates fast solutions in terms of image classification and does not require high performance computers locally. We investigate several data input scenarios for mapping native-vegetation and non-native-vegetation in the Atlantic Forest region encompassed in a Landsat scene (224/076) acquired on November 28, 2019. The data input scenarios were: I- spectral bands (blue to shortwave infrared); II- NDVI (Normalized Difference Vegetation Index); III- mNDWI (modified Normalized Difference Water Index); IV- scenarios I and II; and V- scenarios I to III. Our results showed that the use of spectral bands added NDVI and mNDWI (scenario V) provided the best performance for the native-vegetation mapping, with accuracy of 96.64% and kappa index of 0.91. [ABSTRACT FROM AUTHOR]

This work explores the integration of airborne Light Detection and Ranging (LiDAR) data and WorldView-2 (WV2) images to classify the land cover of a subtropical forest area in Southern Brazil. Different deep and machine learning methods were used: one based on convolutional neural network (CNN) and three ensemble methods. We adopted both pixel- (in the case of CNN) and object-based approaches. The results demonstrated that the integration of LiDAR and WV2 data led to a significant increase (7% to 16%) in accuracies for all classifiers, with kappa coefficient (κ) ranging from 0.74 for the random forest (RF) classifier associated with the WV2 dataset, to 0.92 for the forest by penalizing attributes (FPA) with the full (LiDAR + WV2) dataset. Using the WV2 dataset solely, the best κ was 0.81 with CNN classifier, while for the LiDAR dataset, the best κ was 0.8 with the rotation forest (RotF) algorithm. The use of LiDAR data was especially useful for the discrimination of vegetation classes because of the different height properties among them. In its turn, the WV2 data provided better performance for classes with less structure variation, such as field and bare soil. All the classification algorithms had a nearly similar performance: the results vary slightly according to the dataset used and none of the methods achieved the best accuracy for all classes. It was noticed that both datasets (WV2 and LiDAR) even when applied alone achieved good results with deep and machine learning methods. However, the advantages of integrating active and passive sensors were evident. All these methods provided promising results for land cover classification experiments of the study area in this work. [ABSTRACT FROM AUTHOR]