Filtern
Dokumenttyp
- Video (57)
- Konferenzveröffentlichung (8)
Schlagworte
- Robotik (30)
- Flugkörper (21)
- UAV (21)
- Rettungsrobotik (8)
- 3D Modell (7)
- Rasenmäher (4)
- DRZ (3)
- Deutsches Rettungsrobotik-Zentrum (3)
- SLAM (3)
- Kartierung (2)
Institut
- Informatik und Kommunikation (65) (entfernen)
Wie können mit Luftbildaufnahmen 3D Modelle generiert werden?
- Planen von kreisförmigen und einen rasterförmigen Flug Trajektorien.
- Autonomes Abfliegen und Aufnahme der Bilder
- Verortung der Bilder mittels GPS und Structure from Motion Algorithmen.
- Generierung von 3D Modellen mithilfe von Multi-View Stereo Algorithmen.
Venice 2018: Tradr Review
(2018)
The video shows an orthopoto and a textured 3D model of the location. 300 images were recorded in two short flights with a Mavic Pro in 50 meter height. The first one was a single grid while the camera facing down and the second one was a double grid facing the camera at an 60 degree angle. The 3D model is computed with OpenDroneMap.
In this paper, we present a method for detecting objects of interest, including cars, humans, and fire, in aerial images captured by unmanned aerial vehicles (UAVs) usually during vegetation fires. To achieve this, we use artificial neural networks and create a dataset for supervised learning. We accomplish the assisted labeling of the dataset through the implementation of an object detection pipeline that combines classic image processing techniques with pretrained neural networks. In addition, we develop a data augmentation pipeline to augment the dataset with utomatically labeled images. Finally, we evaluate the performance of different neural networks.
Sperical UAV: Crash Test with 1/2 liter bottle from 2 meters
The video shows the first test of a small spherical UAV (35 cm) with 4 rotors for missions in complex environments such as buildings, caves or tunnels. The spherical design protects the vehicle's internal components and allows the UAV to roll over the ground when the environment allows. The drone can land and take off in any position and come into contact with objects without endangering the propellers and can restart even after crashes.
This technical report is about the architecture and integration of very small commercial UAVs (< 40 cm diagonal) in indoor Search and Rescue missions. One UAV is manually controlled by only one single human operator delivering live video streams and image series for later 3D scene modelling and inspection. In order to assist the operator who has to simultaneously observe the environment and navigate through it we use multiple deep neural networks to provide guided autonomy, automatic object detection and classification and local 3D scene modelling. Our methods help to reduce the cognitive load of the operator. We describe a framework for quick integration of new methods from the field of Deep Learning, enabling for rapid evaluation in real scenarios, including the interaction of methods.
In the realm of digital situational awareness during disaster situations, accurate digital representations,
like 3D models, play an indispensable role. To ensure the
safety of rescue teams, robotic platforms are often deployed
to generate these models. In this paper, we introduce an
innovative approach that synergizes the capabilities of compact Unmaned Arial Vehicles (UAVs), smaller than 30 cm, equipped with 360° cameras and the advances of Neural Radiance Fields (NeRFs). A NeRF, a specialized neural network, can deduce a 3D representation of any scene using 2D images and then synthesize it from various angles upon request. This method is especially tailored for urban environments which have experienced significant destruction, where the structural integrity of buildings is compromised to the point of barring entry—commonly observed post-earthquakes and after severe fires. We have tested our approach through recent post-fire scenario, underlining the efficacy of NeRFs even in challenging outdoor environments characterized by water, snow, varying light conditions, and reflective surfaces.
9 Panoramen, das erste ist aus größerer Höhe aufgenommen und enthält im Himmel eine Karte mit den Positionen der aufgenommenen Punkte (gelb). Das aktuelle Bild ist im Fadenkreuz (rot). Zusätzlich noch ein paar Details zu dem aktuellen Punkt. Jedes Panorama ist 10 Sekunden lang.
Zum Betrachten die höchste Auflösungsstufe wählen und die Pausetaste verwenden. Mit dem gedrückten linken Button kann man sich im Bild bewegen.
The video showcases a 3D model of a chemical company following a tank explosion that occurred on August 17, 2023, in Kempen computed with the AI algorithm Neural Radiance Field (NeRF). Captured by a compact mini drone measuring 18cm x 18cm and equipped with a 360° camera, these images offer an intricate perspective of the aftermath. After a comprehensive aerial survey and inspection of the 360° images taken within the facility, authorities confirmed that it was safe for the evacuated residents to return to their homes. See also:
https://www1.wdr.de/fernsehen/aktuelle-stunde/alle-videos/video-grosser-chemieunfall-in-kempen-100.html
Nerf(acto) for the 3D modeling of the Computer Science building of Westfälische Hochschule GE
(2023)
The video shows a very high resolution 3D point cloud !!! of the computer science building of the University of Applied Science Gelsenkirchen. For the recording a 3 minute flight with a M30T was performed. The 105 images taken by the wide-angle camera during this flight were localized within 3 minutes using colmap and processed using Neural Radiance Fields (NeRF). The nerfacto model of Nerfstudio was trained on an Nvidia RTX 4090 for 8 minutes. Thus, a top 3D model is available after about 15 minutes.
The video shown here shows a free camera path rendered at 60 hz (Full HD).
The video shows a very high resolution 3D point cloud !!! of the outdoor area of the German Rescue Robotics Center. For the recording, a 25-second POI flight was performed with a Mavic 3. From the 4K video footage captured during this flight, 77 images were cropped and localized within 4 minutes using colmap and processed using Neural Radiance Fields (NeRF). The nerfacto model of Nerfstudio was trained on an Nvidia RTX 4090 for 8 minutes. In summary, a top 3D model is available to task forces after about 13 minutes. The calculation is performed locally on site by the RobLW of the DRZ. The video shown here shows a free camera path rendered at 60 hz (Full HD).
Durch Panoramen in Kombination mit dem ORB-SLAM ist ein schnelles Tracking möglich, liefert jedoch ausschließlich spärliche Daten. Durch die Kombination mit einem neuronalen Netz soll der SLAM Algorithmus zu einem RGBD-SLAM erweitert werden, um ein besseres Tracking und eine dichtere Punktwolke zu gewährleisten.
This technical report is about the architecture and integration of commercial UAVs in Search and Rescue missions. We describe a framework that consists of heterogeneous UAVs, a UAV task planner, a bridge to the UAVs, an intelligent image hub, and a 3D point cloud generator. A first version of the framework was developed and tested in several training missions in the EU project TRADR.
At the integration sprint of the E-DRZ consortium in march 2023 we improve the information captured by the human spotter (of the fire brigade) by extending him through a 360° drone. The UAV needs 3 minutes to capture the outdoor scenario and the hall from inside and outside. The hall ist about 70 x 20 meters. When the drone is landed we have all information in 360° degree at 5.7k as you can see it in the video. Furthermore it is a perfect documentation of the deployment scenario. In the next video we will show how to spatial localize the 360° video and how to generate a 3D point cloud from it.
At the integration sprint of the E-DRZ consortium in march 2023 we improve the information captured by the human spotter (of the fire brigade) by extending him through a 360° drone i.e. the DJI Avata with an Insta360 on top of it. The UAV needs 3 minutes to capture the outdoor scenario and the hall from inside and outside. The hall ist about 70 x 20 meters. When the drone is landed we have all information in 360° degree at 5.7k as you can see it in the video. Furthermore it is a perfect documentation of the deployment scenario. In the next video we will show how to spatial localize the 360° video and how to generate a 3D point cloud from it.
Gaussian Splatting: 3D Reconstruction of a Chemical Company After a Tank Explosion in Kempen 8/2023
(2023)
The video showcases a 3D model of a chemical company following a tank explosion that occurred on August 17, 2023, in Kempen computed with the gaussian splatting algorithm. Captured by a compact mini drone measuring 18cm x 18cm and equipped with a 360° camera, these images offer an intricate perspective of the aftermath. The computation need 29 minutes and uses 2770 images (~350 equirectangular images). After a comprehensive aerial survey and inspection of the 360° images taken within the facility, authorities confirmed that it was safe for the evacuated residents to return to their homes. See also:
https://www1.wdr.de/fernsehen/aktuelle-stunde/alle-videos/video-grosser-chemieunfall-in-kempen-100.html
This video features a flight test conducted in our robotics lab, showcasing a custom-built thermal camera drone. We've enhanced a DJI Avata with a specialized thermal camera system. With its compact dimensions measuring 18 x 18 x 17 cm, this drone is designed to navigate and provide critical thermal information within post-fire or post-explosion environments. For more insights, be sure to check out our previous videos on this channel.
This technical report is about the mission and the experience gained during the reconnaissance of an industrial hall with hazardous substances after a major fire in Berlin. During this operation, only UAVs and cameras were used to obtain information about the site and the building. First, a geo-referenced 3D model of the building was created in order to plan the entry into the hall. Subsequently, the UAVs were used to fly in the heavily damaged interior and take pictures from inside of the hall. A 360° camera mounted under the UAV was used to collect images of the surrounding area especially from sections that were difficult to fly into. Since the collected data set contained similar images as well as blurred images, it was cleaned from non-optimal images using visual SLAM, bundle adjustment and blur detection so that a 3D model and overviews could be calculated. It was shown that the emergency services were not able to extract the necessary information from the 3D model. Therefore, an interactive panorama viewer with links to other 360° images was implemented where the links to the other images depends on the semi dense point cloud and located camera positions of the visual SLAM algorithm so that the emergency forces could view the surroundings.
Challenging visual localization of an UAV while flying out of a room into a snowy environment (~ 4:50). The UAV is equipped with a 360° camera. The localization is done with OpenVSLAM.
The video was recorded in Jan. 2019 at the Fire Brigade training center in Dortmund
To achieve nearly real time conditions the original resolution of 5k (30 fps) was reduced to 2k (ffmpeg -i video.mp4 -vf scale=1920:-1 -crf 25 vido-small.mp4) with high compression (-crf 25). This reduce the original size from 3.2 GB to 93MB (~ 4 MBit/s which could be transmitted online via a radio link). The localization shown did not use frameskip. With a frameskip above 1 the localization fails while the UAV is flying through the window. Indoor localization can be done with a frameskip of 3 in real time.
The dataset is used for 3D environment modeling, i.e. for the generation of dense 3D point clouds and 3D models with PatchMatch algorithm and neural networks. Difficult for the modeling algorithm are the reflections of rain, water and snow, as well as windows and vehicle surface. In addition, lighting conditions are constantly changing.
At the beginning of the pandemic in Feb. 2020 I had a little time and wanted to do something new i.e. bring my 3D printer, AI and computer science together somehow. The result is a printed portrait with a lot of computer science. Using style transfer I transferred the etching style of a Göthe portrait to a young girl I call Carolin. By means of image processing I made a black and white picture out of it. Then, using the problem of the traveling salesman, each black point in the picture is interpreted as a city and the whole picture is drawn by only one line. Since this line is very long, it is optimized and shortened by a so-called simulated annealing algorithm. The result is printed in 5 layers on a 3D printer.
ARGUS is a tool for the systematic acquisition, documentation and evaluation of drone flights in rescue operations. In addition to the very fast generation of RGB and IR orthophotos, a trained AI can automatically detect fire, people and cars in the images captured by the drones. The video gives a short introduction to the Aerial Rescue and Geospatial Utility System -- ARGUS
Check out our Github repository under
https://github.com/RoblabWh/argus/
You can find the dataset on kaggle under
https://www.kaggle.com/datasets/julienmeine/rescue-object-detection
The two churches, San Francesco and Sant'Agostino in Amatrice, Italy was hit by an earthquake on August 24 2016. Both churches are in a state of partial collapse, in need of shoring to prevent potential further destruction and to preserve the national heritage. The video show the mission at 1.Sept.2016 in clips of 10 seconds.
The TRADR project was asked by the Italian firebrigade Vigili del Fuoco to provide 3D textured models of two churches.
The team entered San Francesco with two UGVs (ground robots) and one UAV (drone, flown by Prof. Surmann), teleoperating them entirely out of line of sight and partially in collaboration. We entered Sant'Agostino with one UAV (also flown by Prof. Surmann) while two other UAVs were providing a view from different angles to facilitate maneuvering them entirely out of line of sight.
The video shows a snapshot of a 16 minute flight of a DJI Phantom 3 professional over the Schloss Birlinghoven at Sankt Augustin, Germany. The castle is located at the Fraunhofer Campus at Sankt Augustin. The 3D model is generated out of 400 key frames of the 4k video which are cut out with ffmpeg. The work is part of an evaluation in the Tradr Project (www.tradr-project.eu)
Global registration of heterogeneous ground and aerial mapping data is a challenging task. This is especially difficult in disaster response scenarios when we have no prior information on the environment and cannot assume the regular order of man-made environments or meaningful semantic cues. In this work we extensively evaluate different approaches to globally register UGV generated 3D point-cloud data from LiDAR sensors with UAV generated point-cloud maps from vision sensors. The approaches are realizations of different selections for: a) local features: key-points or segments; b) descriptors: FPFH, SHOT, or ESF; and c) transformation estimations: RANSAC or FGR. Additionally, we compare the results against standard approaches like applying ICP after a good prior transformation has been given. The evaluation criteria include the distance which a UGV needs to travel to successfully localize, the registration error, and the computational cost. In this context, we report our findings on effectively performing the task on two new Search and Rescue datasets. Our results have the potential to help the community take informed decisions when registering point-cloud maps from ground robots to those from aerial robots.
From the 360° images of the former video (
• German rescue robotic center captured... ) we now generate the 3D point cloud. The UAV needs 3 minutes to capture the outdoor scenario and the hall from inside and outside. The 3D point cloud generation is 5x slower than the video. It uses a VSLAM algorithm to localize the k-frames (green) and with 3 k-frames it use a 360° PatchMatch algorithm implemented at a NVIDIA graphic card (CUDA) to calculated the dense point clouds.The hall ist about 70 x 20 meters.
This video shows a model computed from 320 images taken at the Tjex 2015 of the trade project (www.tradr-project.eu). The images were acquired with a falcan 8 drone (AscTec) and reconstruct the structure with VisualSfm software. The flight was in 150 m. The Tower is about 95 meter high.
This video shows a model computed from 124 images taken at the Tjex 2015 of the trade project (www.tradr-project.eu). The images were acquired by walking around the object and reconstruct the structure with VisualSfm software.
This paper presents a novel approach to build consistent 3D maps for multi robot cooperation in USAR environments. The sensor streams from unmanned aerial vehicles (UAVs) and ground robots (UGV) are fused in one consistent map. The UAV camera data are used to generate 3D point clouds that are fused with the 3D point clouds generated by a rolling 2D laser scanner at the UGV. The registration method is based on the matching of corresponding planar segments that are extracted from the point clouds. Based on the registration, an approach for a globally optimized localization is presented. Apart from the structural information of the point clouds, it is important to mention that no further information is required for the localization. Two examples show the performance of the overall registration.
360° UAV Flight in a collapse test setup at the German Resuce Robotik Center
360° Camera at a small UAV
(2021)
360° and IR- Camera Drone Flight Test: Superimposition of two data sources for Post-Fire Inspection
(2023)
This video highlights a recent flight test carried out in our cutting-edge robotics lab, unveiling the capabilities of our meticulously crafted thermal and 360° camera drone! We've ingeniously upgraded a DJI Avata with a bespoke thermal and 360° camera system. Compact yet powerful, measuring just 18 x 18 x 17 cm, this drone is strategically engineered to effortlessly navigate and deliver crucial thermal and 360° insights concurrently in post-fire or post-explosion environments.
The integration of a specialized thermal and 360° camera system enables the simultaneous capture of both data sources during a single flight. This groundbreaking approach not only reduces inspection time by half but also facilitates the seamless superimposition of thermal and 360° videos for comprehensive analysis and interpretation.