Search

As the world population continues to grow, the demand for food is also increasing. At the same time, a series of global and local challenges are threatening ‘food security’. In this context, various technologies and techniques have been proposed and considered in recent decades to secure the efficient usage of the planet’s agricultural resources, i.e. ‘smart agriculture’. This requires the accurate and advanced acquisition, modelling and management of relevant data. This article presents a brief discussion of how unmanned aerial vehicles (UAVs or ‘drones’) can play a critical role in smart agriculture, including a focus on their capabilities and applications.

The use of drones to explore indoor spaces has gained attention and popularity for disaster management and indoor navigation applications. In this paper we present the operations and mapping techniques of two drones that are different in terms of size, the sensors deployed, and the positioning and mapping techniques used. The first drone is a low-cost commercial quadcopter microdrone, a Crazyflie, while the second drone is a relatively expensive research quadcopter macrodrone, called MAX. We investigated their feasibility in mapping areas where satellite positioning is not available, such as indoor spaces. We compared the point clouds obtained by a multi-ranger deck, a multi-layer LIDAR scanner and a stereo camera, and assessed each against ground truth obtained with a terrestrial laser scanner. Results showed that both drones are capable of mapping relatively cluttered indoor environments and can provide point clouds that are sufficient for a quick exploration. Furthermore, the LIDAR scanner-based system can handle a relatively large office environment with an accumulated drift less than 0.02% (1 cm) on the Z-axis and 0.77% (50 cm) on the X and Y axes over a length trajectory of about 65 m. Despite the limited features of the sensor configuration of the Crazyflie, its performance is promising for mapping indoor spaces, given the relatively low deviation from the ground truth: cloud-to-cloud distances measured were generally less than 20 cm.

Self-supervised monocular depth estimation that does not require ground truth for training has attracted attention in recent years. It is of high interest to design lightweight but effective models so that they can be deployed on edge devices. Many existing architectures benefit from using heavier backbones at the expense of model sizes. This paper achieves comparable results with a lightweight architecture. Specifically, the efficient combination of CNNs and Transformers is investigated, and a hybrid architecture called Lite-Mono is presented. A Consecutive Dilated Convolutions (CDC) module and a Local-Global Features Interaction (LGFI) module are proposed. The former is used to extract rich multi-scale local features, and the latter takes advantage of the self-attention mechanism to encode long-range global information into the features. Experiments demonstrate that Lite-Mono outperforms Monodepth2 by a large margin in accuracy, with about 80% fewer trainable parameters., Comment: Accepted to CVPR2023

Semantic segmentation models are often affected by illumination changes, and fail to predict correct labels. Although there has been a lot of research on indoor semantic segmentation, it has not been studied in low-light environments. In this paper we propose a new framework, LISU, for Low-light Indoor Scene Understanding. We first decompose the low-light images into reflectance and illumination components, and then jointly learn reflectance restoration and semantic segmentation. To train and evaluate the proposed framework, we propose a new data set, namely LLRGBD, which consists of a large synthetic low-light indoor data set (LLRGBD-synthetic) and a small real data set (LLRGBD-real). The experimental results show that the illumination-invariant features effectively improve the performance of semantic segmentation. Compared with the baseline model, the mIoU of the proposed LISU framework has increased by 11.5%. In addition, pre-training on our synthetic data set increases the mIoU by 7.2%. Our data sets and models are available on our project website.

The increasing development and production of unmanned aerial vehicles (UAVs or ‘drones’) with a higher level of autonomy can reduce human error and assist in safer, faster and more accurate inspections. Thanks to their ability to perform repeated inspections, even in challenging and dynamic environments, they make it easier to monitor infrastructural changes over time. This will benefit several potential applications, such as civil engineering inspections. To ensure sufficient inspection outputs, a conventional UAV mission needs to be controlled by a highly skilled pilot or operator. This professional should be aware of what kind of information the inspectors need and adapt to any on-site difficulties such as obstacles, time of day and weather conditions. After the UAV flight mission is finished, a large number of acquired images are processed. Whether the images are used in visual inspection or 3D model generation, both cases require significant computing power and experience. The recent development towards autonomous inspections using conventional UAVs equipped with either image-based or Lidar-based tools can certainly reduce human error and assist in a safer, faster, more accurate inspection generating data that is accessible online. Furthermore, UAVs can perform repeated inspections even in challenging and dynamic environments, making it easier to monitor infrastructural changes over time.

Unmanned aerial vehicles (UAVs) are considered an innovative tool for land administration. However, despite the prospects and market opportunities in the domain, there is a gap between experimentation and widespread technology diffusion. In this work, the Framework for Effective Land Administration (FELA) and the Hype Cycle concept are integrated to understand the dynamics of the innovation process of UAVs for the land administration sector. Empirical data stems from literature and interviews of UAV and land administration experts worldwide. The majority of experts estimate UAV technology to be in a phase in which the innovation needs to overcome initial unmet expectations to foster market development and increased adoption. The assessment indicates the changing importance of different FELA pathways during this process. Enabling laws and policies and supporting governance, accountability and institutions are crucial to create such a UAV-friendly national ecosystem early on and allay exaggerated expectations. Once this ecosystem has been made, market demand is expected to surge driven by partnerships, adapted standards, tech advocacy and awareness-raising campaigns, highlighting the superiority of high-resolution data amongst other benefits of UAV technology. These insights can be used as a baseline to direct national strategic decisions towards the increased adoption of UAVs in land administration.

Urban search and rescue (USaR) teams require a fast and thorough building damage assessment, to focus their rescue efforts accordingly. Unmanned aerial vehicles (UAV) are able to capture relevant data in a short time frame and survey otherwise inaccessible areas after a disaster, and have thus been identified as useful when coupled with RGB cameras for façade damage detection. Existing literature focuses on the extraction of 3D and/or image features as cues for damage. However, little attention has been given to the efficiency of the proposed methods which hinders its use in an urban search and rescue context. The framework proposed in this paper aims at a more efficient façade damage detection using UAV multi-view imagery. This was achieved directing all damage classification computations only to the image regions containing the façades, hence discarding the irrelevant areas of the acquired images and consequently reducing the time needed for such task. To accomplish this, a three-step approach is proposed: i) building extraction from the sparse point cloud computed from the nadir images collected in an initial flight; ii) use of the latter as proxy for façade location in the oblique images captured in subsequent flights, and iii) selection of the façade image regions to be fed to a damage classification routine. The results show that the proposed framework successfully reduces the extracted façade image regions to be assessed for damage 6 fold, hence increasing the efficiency of subsequent damage detection routines. The framework was tested on a set of UAV multi-view images over a neighborhood of the city of L’Aquila, Italy, affected in 2009 by an earthquake.

Unmanned Aerial Vehicles (UAVs) have gained popularity in acquiring geotagged, low cost and high resolution images. However, the images acquired by UAV-borne cameras often have poor georeferencing information, because of the low quality on-board Global Navigation Satellite System (GNSS) receiver. In addition, lightweight UAVs have a limited payload capacity to host a high quality on-board Inertial Measurement Unit (IMU). Thus, orientation parameters of images acquired by UAV-borne cameras may not be very accurate. Poorly georeferenced UAV images can be correctly oriented using accurately oriented airborne images capturing a similar scene by finding correspondences between the images. This is not a trivial task considering the image pairs have huge variations in scale, perspective and illumination conditions. This paper presents a procedure to successfully register UAV and aerial oblique imagery. The proposed procedure implements the use of the AKAZE interest operator for feature extraction in both images. Brute force is implemented to find putative correspondences and later on Lowe’s ratio test (Lowe, 2004) is used to discard a significant number of wrong matches. In order to filter out the remaining mismatches, the putative correspondences are used in the computation of multiple homographies, which aid in the reduction of outliers significantly. In order to increase the number and improve the quality of correspondences, the impact of pre-processing the images using the Wallis filter (Wallis, 1974) is investigated. This paper presents the test results of different scenarios and the respective accuracies compared to a manual registration of the finally computed fundamental and essential matrices that encode the orientation parameters of the UAV images with respect to the aerial images.

In the last years we have witnessed a rapid development of UAVs (Unmanned Aerial Vehicles), especially for image collection. One of the advantages is the possibility to perform high resolution and repeated flights in a cheap way to detect changes over time. Thus, dynamic scenes can be monitored acquiring image blocks in different epochs in a flexible way. Anyway, most of UAVs are not able to provide accurate direct geo-referencing information, so image blocks from different epochs still need to be co-registered to efficiently detect changes. This task is mostly completed using GCPs (Ground Control Points), although this approach is time consuming as manual intervention is needed. This paper aims at investigating new techniques to automate the co-registration of image blocks without the use of GCPs, just relying on an image based co-registration (IBCR) approach. The image alignment is initially performed on a reference (anchor) epoch and the registration of the following (slave) epochs is performed including some (anchor) images from the reference epoch with fixed external orientation parameters. This allows constraining the Bundle Block Adjustment of the slave epoch to be consistent with the reference one. The study involved the use of 10 multi-temporal image block over a large building construction site, and spanning a time frame of 2 years. Different tests have been performed for the reference image choice with a manual approach and then evaluating the reached accuracy. The performed tests on the chosen test site have shown that the accuracy of the proposed methodology provides results comparable to the common GCPs registration approach.

For many applications within urban environments the combined use of images taken from the ground and from unmanned aerial platforms seems interesting: while from the airborne perspective the upper parts of objects including roofs can be observed, the ground images can complement the data from lateral views to retrieve a complete visualisation or 3D reconstruction of interesting areas. The automatic co-registration of air- and ground-based images is still a challenge and cannot be considered solved. The main obstacle is originating from the fact that objects are photographed from quite different angles, and hence state-of-the-art tie point measurement approaches cannot cope with the induced perspective transformation. One first important step towards a solution is to use airborne images taken under slant directions. Those oblique views not only help to connect vertical images and horizontal views but also provide image information from 3D-structures not visible from the other two directions. According to our experience, however, still a good planning and many images taken under different viewing angles are needed to support an automatic matching across all images and complete bundle block adjustment. Nevertheless, the entire process is still quite sensible – the removal of a single image might lead to a completely different or wrong solution, or separation of image blocks. In this paper we analyse the impact different parameters and strategies have on the solution. Those are a) the used tie point matcher, b) the used software for bundle adjustment. Using the data provided in the context of the ISPRS benchmark on multi-platform photogrammetry, we systematically address the mentioned influences. Concerning the tie-point matching we test the standard SIFT point extractor and descriptor, but also the SURF and ASIFT-approaches, the ORB technique, as well as (A)KAZE, which are based on a nonlinear scale space. In terms of pre-processing we analyse the Wallis-filter. Results show that in more challenging situations, in this case for data captured from different platforms at different days most approaches do not perform well. Wallis-filtering emerged to be most helpful especially for the SIFT approach. The commercial software pix4dmapper succeeds in overall bundle adjustment only for some configurations, and especially not for the entire image block provided.

Since the past two years, unmanned aerial vehicles (UAVs) are evolving as a mapping tool in African countries. However, conditions for flying, controlling and referencing respective data are more complex than in Europe and are often underestimated. Especially geometric accuracy and radiometry are negatively influenced by densely populated areas and adverse meteorological conditions. More than 40 datasets from 8 UAV field campaigns and demonstrations in Zanzibar, Rwanda and Kenya are used to conclude findings on opportunities and challenges to achieve high levels of data quality. The research is associated with the EU H2020 project “its4land”

The number of scientific studies that consider possible applications of remotely piloted aircraft systems (RPASs) for the management of natural hazards effects and the identification of occurred damages strongly increased in the last decade. Nowadays, in the scientific community, the use of these systems is not a novelty, but a deeper analysis of the literature shows a lack of codified complex methodologies that can be used not only for scientific experiments but also for normal codified emergency operations. RPASs can acquire on-demand ultra-high-resolution images that can be used for the identification of active processes such as landslides or volcanic activities but can also define the effects of earthquakes, wildfires and floods. In this paper, we present a review of published literature that describes experimental methodologies developed for the study and monitoring of natural hazards.

During the last decade the use of airborne multi camera systems increased significantly. The development in digital camera technology allows mounting several mid- or small-format cameras efficiently onto one platform and thus enables image capture under different angles. Those oblique images turn out to be interesting for a number of applications since lateral parts of elevated objects, like buildings or trees, are visible. However, occlusion or illumination differences might challenge image processing. From an image orientation point of view those multi-camera systems bring the advantage of a better ray intersection geometry compared to nadir-only image blocks. On the other hand, varying scale, occlusion and atmospheric influences which are difficult to model impose problems to the image matching and bundle adjustment tasks. In order to understand current limitations of image orientation approaches and the influence of different parameters such as image overlap or GCP distribution, a commonly available dataset was released. The originally captured data comprises of a state-of-the-art image block with very high overlap, but in the first stage of the so-called ISPRS/EUROSDR benchmark on multi-platform photogrammetry only a reduced set of images was released. In this paper some first results obtained with this dataset are presented. They refer to different aspects like tie point matching across the viewing directions, influence of the oblique images onto the bundle adjustment, the role of image overlap and GCP distribution. As far as the tie point matching is concerned we observed that matching of overlapping images pointing to the same cardinal direction, or between nadir and oblique views in general is quite successful. Due to the quite different perspective between images of different viewing directions the standard tie point matching, for instance based on interest points does not work well. How to address occlusion and ambiguities due to different views onto objects is clearly a non-solved research problem so far. In our experiments we also confirm that the obtainable height accuracy is better when all images are used in bundle block adjustment. This was also shown in other research before and is confirmed here. Not surprisingly, the large overlap of 80/80% provides much better object space accuracy – random errors seem to be about 2-3fold smaller compared to the 60/60% overlap. A comparison of different software approaches shows that newly emerged commercial packages, initially intended to work with small frame image blocks, do perform very well.

The extraction of information from image and range data is one of the main research topics. In literature, several papers dealing with this topic has been already presented. In particular, several authors have suggested an integrated use of both range and image information in order to increase the reliability and the completeness of the results exploiting their complementary nature. In this paper, an integration between range and image data for the geometric reconstruction of man-made object is presented. In particular, the focus of this paper is on the edge extraction procedure performed in an integrated way exploiting the information provided by both range and image data. Both terrestrial and aerial applications have been analysed for the façade extraction in terrestrial acquisitions and the roof outline extraction from aerial data. The algorithm and the achieved results are described and discussed in detail.

The realm of land administration is currently being challenged: conventional western-oriented land administration systems have mostly failed to supply their expected results in developing countries. Amongst others, unmanned aerial vehicles (UAVs) are evolving as a tool for alternative land tenure data acquisition approaches. However, major bottlenecks such as regulatory frameworks and time-consuming ground truthing are issues currently hindering large-scale implementation. This study sheds light into the design process of UAV-based data acquisition workflows and reveals results of initial field tests in Rwanda and Germany. Insights of operational challenges and data quality measures will be presented. The research is associated with the EU H2020 project “its4land”.

Automated damage assessment based on satellite imagery is crucial for initiating fast response actions. Several methods based on supervised learning approaches have been reported as effective for automated mapping of damages using remote sensing images. However, adopting these methods for practical use is still challenging, as they typically demand large amounts of training samples to build a supervised classifier, which are usually not readily available. With the advancement in technologies local and detailed damage assessment for individual buildings is being made available, for example through analysis of images captured by unmanned aerial vehicles, monitoring systems installed in buildings, and through crowdsourcing. Often such assessments are being done in parallel, with results becoming available progressively. In this paper, an online classification strategy is adopted where a classifier is built incrementally using the streaming damage labels from various sources as training samples, i.e. without retraining it from the scratch when new samples stream in. The Passive-Aggressive online classifier is used for the classification process. Apart from the classifier, the choice of image features plays a crucial role in the performance of the classification. The features extracted using recently reported deep learning approaches such as Convolutional Neural Networks (CNN), which learns features directly from images, have been reported to be more effective than conventional handcrafted features such as gray level co-occurrence matrix and Gabor wavelets. Thus in this study, the potential of CNN features is explored for online classification of satellite image to detect structural damage, and is compared against handcrafted features. The feature extraction and classification process is carried out at an object level, where the objects are obtained by over-segmentation of the satellite image. The proposed online framework for damage classification achieves a maximum overall accuracy of about 73%, which is comparable to that of batch classifier accuracy (74%) obtained for the same training and testing samples, however at a significantly lesser time and memory requirements. Moreover, the CNN features always significantly outperform handcrafted features.

Many supervised and unsupervised algorithms for the automated and reliable classification of large regions using high resolution data have been presented in the remote sensing community in the last decades. Most of these approaches exploit a single input data: high resolution orthophotos or 3D point clouds. An increasing number of contributions has more recently exploited the combined use of orthophotos and LiDAR DSM taking advantage from the complementarity of these inputs. Nevertheless, very few applications have considered the use of overlapping multi-view images on the same area for classification. In this paper the first tests to investigate this classification architecture are presented. Different typologies of DSM and orthophoto as well as a variable number of images on the same area have been considered in the experiments. The preliminary results on the two available test areas will be shown, in order to draw the first conclusions on this approach and discuss the further developments of this research.

Augmented Reality (AR) allows the direct and intuitive access of digital objects visualized in real time on computer screens. AR applications are commonly used in many different areas, such as entertainment, sports and tourism, while their use in archaeology is still limited. When employed, current archaeological AR applications use expensive devices or flat targets, which are insufficient for visualizing complex artefacts. In this paper, we present a low‐cost, automated method called the ARCube System, which allows the expansion and enrichment of AR applications focused on archaeological objects. In this paper, the newly developed ARCube System is described by presenting several archaeological examples to show the system's ability to visualize and investigate archaeological finds. The reported tests demonstrate the method's reliability on a variety of objects characterized by different shapes and sizes. The ARCube System allows users to interact with digital 3D models, rotate them through 360° and explore details in high resolution (without any risk of damaging the find). The system renders 1:1 scale between digital and actual object and offers a low‐cost flexible system for the interactive visualization of archaeological finds that can be beneficial to public research (e.g., museum exhibits) and education (e.g., the classroom) and may possibly lead to new avenues of archaeological research.

International audience; 3D Optical Metrology (3DOM) unit, Bruno Kessler Foundation (FBK), Trento, Italy @fbk.eu, http://3dom.fbk.eu Commission III-WG4 ABS TRACT: The use of oblique imagery has become a standard for many civil and mapping applications, thanks to the development of airborne digital multi-camera systems, as proposed by many companies (Blomoblique, IGI, Leica, M idas, Pictometry, Vexcel/M icrosoft, VisionM ap, etc.). The indisputable virtue of oblique photography lies in its simplicity of interpretation and understanding for inexperienced users allowing their use of oblique images in very different applications, such as building detection and reconstruction, building structural damage classification, road land updating and administration services, etc. The paper reports an overview of the actual oblique commercial systems and presents a workflow for the automated orientation and dense matching of large image blocks. Perspectives, potentialities, pitfalls and suggestions for achieving satisfactory results are given. Tests performed on two datasets acquired with two multi-camera systems over urban areas are also reported. Figure 1: Large urban area pictured with an oblique multi-camera system. Once advanced image triangulation methods have retrieved interior and exterior parameters of the cameras, dense point clouds can be deriv ed for 3D city modelling, feature extraction and mapping purposes.

Nowadays, multi-camera platforms combining nadir and oblique cameras are experiencing a revival. Due to their advantages such as ease of interpretation, completeness through mitigation of occluding areas, as well as system accessibility, they have found their place in numerous civil applications. However, automatic post-processing of such imagery still remains a topic of research. Configuration of cameras poses a challenge on the traditional photogrammetric pipeline used in commercial software and manual measurements are inevitable. For large image blocks it is certainly an impediment. Within theoretical part of the work we review three common least square adjustment methods and recap on possible ways for a multi-camera system orientation. In the practical part we present an approach that successfully oriented a block of 550 images acquired with an imaging system composed of 5 cameras (Canon Eos 1D Mark III) with different focal lengths. Oblique cameras are rotated in the four looking directions (forward, backward, left and right) by 45° with respect to the nadir camera. The workflow relies only upon open-source software: a developed tool to analyse image connectivity and Apero to orient the image block. The benefits of the connectivity tool are twofold: in terms of computational time and success of Bundle Block Adjustment. It exploits the georeferenced information provided by the Applanix system in constraining feature point extraction to relevant images only, and guides the concatenation of images during the relative orientation. Ultimately an absolute transformation is performed resulting in mean re-projection residuals equal to 0.6 pix.

The paper presents the research carried out in the on-going 3DSolarWeb project to test and implement a complete pipeline for the generation of a solar cadastre of building roofs located in alpine areas. The project aims at providing reliable results in a costeffective way, using (low resolution) available data and new aerial imagery acquisitions as input. The environmental context is digitally represented using already existing low resolution LiDAR data (1–2 m resolution), while the urban area is modelled using high resolution aerial images (10–20 cm GSD) and photogrammetric DSM. Reliable models and algorithms for the estimation of the incoming sun radiance are then adopted and a WebGIS is set up for the interactive calculation of the photovoltaic (PV) potential in a raster-based form. The paper summarizes the entire pipeline and the results (Figure 1) achieved on the test areas to show the potentialities of the method and the web-based service.

The paper reviews the state-of-the-art in 3D city models and building block generation, with a description of the most common solutions and approaches. Then the digital reconstruction and comparison of LoD1 and LoD2 building models obtained with commercial packages and using different input data are presented. As input data, a DSM at 1m resolution derived from a GeoEye-1 stereo-pair, a DSM from an aerial block at 50 cm GSD and a LiDAR-based DSM at 1 m resolution are used. The geometric buildings produced with each dataset are evaluated with respect to some ground-truth measurements but also compared between them. Problems such as the quality of the input DSM, the accuracy of the necessary vector datasets containing the building footprints, the flexibility of the approaches and the potentialities of each dataset will be discussed. As reconstruction of the building models is largely influenced by the quality of the building footprints, which may be out-of-date or slightly shifted with respect to the employed DSMs/DTMs, an in-house method is being developed to derive them starting from the produced DSMs.