Search

It is hard to create consistent ground truth data for interest points in natural images, since interest points are hard to define clearly and consistently for a human annotator. This makes interest point detectors non-trivial to build. In this work, we introduce an unsupervised deep learning-based interest point detector and descriptor. Using a self-supervised approach, we utilize a siamese network and a novel loss function that enables interest point scores and positions to be learned automatically. The resulting interest point detector and descriptor is UnsuperPoint. We use regression of point positions to 1) make UnsuperPoint end-to-end trainable and 2) to incorporate non-maximum suppression in the model. Unlike most trainable detectors, it requires no generation of pseudo ground truth points, no structure-from-motion-generated representations and the model is learned from only one round of training. Furthermore, we introduce a novel loss function to regularize network predictions to be uniformly distributed. UnsuperPoint runs in real-time with 323 frames per second (fps) at a resolution of 224x320 and 90 fps at 480x640. It is comparable or better than state-of-the-art performance when measured for speed, repeatability, localization, matching score and homography estimation on the HPatch dataset.

Today, agricultural vehicles are available that can drive autonomously and follow exact route plans more precisely than human operators. Combined with advancements in precision agriculture, autonomous agricultural robots can reduce manual labor, improve workflow, and optimize yield. However, as of today, human operators are still required for monitoring the environment and acting upon potential obstacles in front of the vehicle. To eliminate this need, safety must be ensured by accurate and reliable obstacle detection and avoidance systems.In this thesis, lidar-based obstacle detection and recognition in agricultural environments has been investigated. A rotating multi-beam lidar generating 3D point clouds was used for point-wise classification of agricultural scenes, while multi-modal fusion with cameras and radar was used to increase performance and robustness. Two research perception platforms were presented and used for data acquisition. The proposed methods were all evaluated on recorded datasets that represented a wide range of realistic agricultural environments and included both static and dynamic obstacles.For 3D point cloud classification, two methods were proposed for handling density variations during feature extraction. One method outperformed a frequently used generic 3D feature descriptor, whereas the other method showed promising preliminary results using deep learning on 2D range images. For multi-modal fusion, four methods were proposed for combining lidar with color camera, thermal camera, and radar. Gradual improvements in classification accuracy were seen, as spatial, temporal, and multi-modal relationships were introduced in the models. Finally, occupancy grid mapping was used to fuse and map detections globally, and runtime obstacle detection was applied on mapped detections along the vehicle path, thus simulating an actual traversal.The proposed methods serve as a first step towards full autonomy for agricultural vehicles. The study has thus shown that recent advancements in autonomous driving can be transferred to the agricultural domain, when accurate distinctions are made between obstacles and processable vegetation. Future research in the domain has further been facilitated with the release of the multi-modal obstacle dataset, FieldSAFE.

Today, agricultural vehicles are available that can automatically perform tasks such as weed detection and spraying, mowing, and sowing while being steered automatically. However, for such systems to be fully autonomous and self-driven, not only their specific agricultural tasks must be automated. An accurate and robust perception system automatically detecting and avoiding all obstacles must also be realized to ensure safety of humans, animals, and other surroundings. In this paper, we present a multi-modal obstacle and environment detection and recognition approach for process evaluation in agricultural fields. The proposed pipeline detects and maps static and dynamic obstacles globally, while providing process-relevant information along the traversed trajectory. Detection algorithms are introduced for a variety of sensor technologies, including range sensors (lidar and radar) and cameras (stereo and thermal). Detection information is mapped globally into semantical occupancy grid maps and fused across all sensors with late fusion, resulting in accurate traversability assessment and semantical mapping of process-relevant categories (e.g., crop, ground, and obstacles). Finally, a decoding step uses a Hidden Markov model to extract relevant process-specific parameters along the trajectory of the vehicle, thus informing a potential control system of unexpected structures in the planned path. The method is evaluated on a public dataset for multi-modal obstacle detection in agricultural fields. Results show that a combination of multiple sensor modalities increases detection performance and that different fusion strategies must be applied between algorithms detecting similar and dissimilar classes.

Udgivelsesdato: Januar 2010 This paper examines the historical background, current context and future challenges for housing energy consumption in Denmark. There has been a radical transformation in housing energy consumption over the last 30 years, with an absolute reduction in heat consumption and a rapid growth in electricity consumption, reflecting wider technological and social transformations in the movement from an industrial to a knowledge based society. In new housing it is shown that electricity consumption now dominates the total primary energy consumption, and that as a consequence traditional heat saving paradigms are relatively less effective, and can result in overheating and rising electricity consumption. At the same time, climate change concerns show that rising temperatures will in the future result in a falling heat demand and increasing cooling demand in housing. With this background, a theoretical framework is proposed for defining low-energy paradigms, based on which components of energy consumption are regulated. It is shown that there has been a historical movement from older, narrow paradigms to newer, broader paradigms in Denmark, best exemplified by the movement towards zero-energy housing.

In recent years, mapping and automation has been increasingly investigated and applied in precision agriculture. The ultimate goal of this development is to apply autonomous vehicles operating efficiently without any human intervention. Such autonomous operation imposes severe safety hazards, demanding accurate and robust risk detection, and avoidance systems. It is unlikely that one sensor can single­handedly guarantee this, and therefore multiple sensing modalities are often combined in order to increase detection performance and introduce redundancy. In this paper, we present a global mapping approach utilizing diverse sensor technologies to achieve a uniform obstacle interpretation of the environment. Using occupancy grid maps, we fuse information from a monocular color camera, a RADAR, and a LIDAR in combination with IMU­assisted GPS­positioning. For each sensor, we present detection algorithms, mapping from raw sensor data to a 2D grid­based obstacle interpretation of the environment. These are then fused temporally with the occupancy grid algorithm, and afterwards spatially in a competitive and complementary way to produce a combined global obstacle map. The method is evaluated on an extensive dataset recorded at Research Centre Foulum, Denmark, in June 2015. The dataset comprises sensor data from a tractor­mounted recording system in a grass mowing scenario with various obstacles. A ground truth map has been obtained with a mapping drone. Results show promising obstacle detection capabilities and an increase in performance when fusing information across sensor modalities and layers. The proposed mapping framework is able to fuse a vast amount of information across a diverse sensor set, using an efficient and novel approach for obstacle detection in agriculture.

In recent years, the drive of the Industry 4.0 initiative has enriched industrial and scientific approaches to build self-driving cars or smart factories. Agricultural applications benefit from both advances, as they are in reality mobile driving factories which process the environment. Therefore, acurate perception of the surrounding is a crucial task as it involves the goods to be processed, in contrast to standard indoor production lines. Environmental processing requires accurate and robust quantification in order to correctly adjust processing parameters and detect hazardous risks during the processing. While today approaches still implement functional elements based on a single particular set of sensors, it may become apparent that a unified representation of the environment compiled from all available information sources would be more versatile, sufficient, and cost effective. The key to this approach is the means of developing a common information language from the data provided. In this paper, we introduce and discuss techniques to build so called inverse sensor models that create a common information language among different, but typically agricultural, information providers. These can be current live sensor data, farm management systems, or long term information generated from previous processing, drones, or satellites. In the context of Industry 4.0, this enables the interoperability of different agricultural systems and allows information transparency., IROS 2017 Workshop on Agri-Food Robotics, 6 pages, 12 figures

The research is structured around complex optical effects of undulated glass and coatings that exhibit high reflectivity, especially at higher incidence angles. This optical behavior allows great directional and intensity variation even with minimal curvatures. This is especially visible in undulated glass facades, where the ratio of reflection and transmission changes dramatically with every change of viewing angle, producing dynamic effects. Furthermore, caustic and dispersion light effects may additionally enrich the visual architectural language. However, high-frequency and high-amplitude curvatures also create optical distortion and lens effects, thus producing undesirable glare and potential overheating. Therefore, these undulated glass geometries require special consideration and analyses during the design phase to maximize their optical benefits and reduce unwanted side-effects.The research explores self-shading potential and optical benefits of undulated glass geometries through modulating surface curvature. Sinusoidal functions are used to define coherent procedural glass surface curvature with variable amplitude values. Due to the high directional sensitivity, virtual prototyping of undulated geometries was performed with a high temporal and spatial resolution and assuming the most common vertical setting.Optical benefits are presented through g-value and daylighting metrics for a set of undulated glass facade states, varying curvature amplitude, and directionality. The results indicate moderate self-shading potential and daylighting benefit of undulated glass facades. The research findings aimed to provide a valuable resource for articulating design and performance aspects of undulated glass facades, potentially reducing the need for additional shading and daylight-redirect elements.