Studying growth anddevelopment of plants is of central importance in botany. Current quantitative are either limited to tedious and sparse manual measurements, or coarse image-based 2d measurements. Availability of c...
详细信息
Studying growth anddevelopment of plants is of central importance in botany. Current quantitative are either limited to tedious and sparse manual measurements, or coarse image-based 2d measurements. Availability of cheap and portable 3d acquisition devices has the potential to automate this process and easily provide scientists with volumes of accurate data, at a scale much beyond the realms of existing methods. However, during their development, plants grow new parts (e.g., vegetative buds) and bifurcate to different components - violating the central incompressibility assumption made by existing acquisition algorithms, which makes these algorithms unsuited for analyzing growth. We introduce a framework to study plant growth, particularly focusing on accurate localization and tracking topological events like budding and bifurcation. This is achieved by a novel forward-backward analysis, wherein we track robustly detected plant components back in time to ensure correct spatio-temporal event detection using a locally adapting threshold. We evaluate our approach on several groups of time lapse scans, often ranging from days to weeks, on a diverse set of plant species and use the results to animate static virtual plants or directly attach them to physical simulators.
LidAR-based 3d scene perception is a fundamental and important task for autonomous driving. Most state-of-the-art methods on LidAR-based 3d recognition tasks focus on single-frame 3dpointclouddata, ignoring tempora...
详细信息
LidAR-based 3d scene perception is a fundamental and important task for autonomous driving. Most state-of-the-art methods on LidAR-based 3d recognition tasks focus on single-frame 3dpointclouddata, ignoring temporal information. We argue that the temporal information across the frames provides crucial knowledge for 3d scene perceptions, especially in the driving scenario. In this paper, we focus on spatial and temporal variations to better explore temporal information across 3d frames. We design a temporal variation-aware interpolation module and a temporal voxel-point refinement module to capture the temporal variation in the 4d point cloud. The temporal variation-aware interpolation generates local features from the previous and current frames by capturing spatial coherence and temporal variation information. The temporal voxel-point refinement module builds a temporal graph on the 3dpointcloud sequences and captures the temporal variation with a graph convolution module, transforming coarse voxel-level predictions into fine point-level predictions. With our proposed modules, we achieve superior performances on SemanticKITTI, SemantiPOSS and NuScenes.
We address a timely and relevant problem in signal processing: The recognition of patterns from spatial data in motion through a zero-shot learning scenario. We introduce a neural network architecture based on Siamese...
详细信息
ISBN:
(纸本)9781728163383
We address a timely and relevant problem in signal processing: The recognition of patterns from spatial data in motion through a zero-shot learning scenario. We introduce a neural network architecture based on Siamese networks to recognize unseen classes of motion patterns. The approach uses a graph-based technique to achieve permutation invariance and also encodes moving pointclouds into a representation space in a computationally efficient way. We evaluated the model on an open dataset with twenty-one gestures. The model outperformes state-of-the-art architectures with a considerable margin in four different settings in terms of accuracy while reducing the computational complexity up to 60 times.
In recent years, exponential growth has been detected in research efforts focused on automated construction progress monitoring. despite various data acquisition methods and approaches, the success is limited. This pa...
详细信息
In recent years, exponential growth has been detected in research efforts focused on automated construction progress monitoring. despite various data acquisition methods and approaches, the success is limited. This paper proposes a new method, where changes are constantly perceived and as-built model continuously updatedduring the construction process, instead of periodical scanning of the whole building under construction. It turned out that low precision 3d scanning devices, which are closely observing active workplaces, are sufficient for correct identification of the built elements. Such scanning devices are small enough to fit onto workers' protective helmets and on the applied machinery. In this way, workers capture all workplaces inside and outside of the building in real time and record partial pointclouds, their locations, and time stamps. The partial pointclouds are then registered and merged into a complete 4d as-built pointcloud of a building under construction. Identification of as-designed BIM elements within the 4d as-built pointcloud then results in the 4d as-built BIM. Finally, the comparison of the 4d as-built BIM and the 4d as-designed BIM enables identification of the differences between both models and thus the deviations from the time schedule. The differences are reported in virtual real-time, which enables more efficient project management.
With improved sensor resolution and advanced multi-pass interferometric techniques such as SAR tomographic inversion (TomoSAR), it is now possible to reconstruct both shape and motion of urban infrastructures. These s...
详细信息
With improved sensor resolution and advanced multi-pass interferometric techniques such as SAR tomographic inversion (TomoSAR), it is now possible to reconstruct both shape and motion of urban infrastructures. These sophisticated techniques not only opens up new possibilities to monitor and visualize the dynamics of urban infrastructure in very high level of details but also allows us to take a step further towards generation of 4d (space-time) or even higher dimensional dynamic city models that can potentially incorporate temporal (motion) behaviour along with the 3d information. Motivated by these chances, this paper presents a post processing approach that systematically allows automatic reconstruction of building facades from 4d point cloud generated from tomographic SAR processing and put the particular focus on robust reconstruction of large areas. The approach is modular and consists of extracting facade points via pointdensity estimation procedure based on directional window approach. Segmentation of facades into individual segments is then carried out using an unsupervised clustering procedure combining both the density-based clustering and the mean-shift algorithm. Subsequently, points of individual facade segments are identified as belonging to flat or curved surface and general 1st and 2nd order polynomials are used to model the facade geometry. Finally, intersection points of the adjacent facades describing the vertex points are determined to complete the reconstruction process. The proposed approach is illustrated and validated by examples using TomoSAR pointclouds over the city of Las Vegas generated from a stack of TerraSAR-X high resolution spotlight images.
We propose a new biometric approach where the tissue thickness of a person's forehead is used as a biometric feature. Given that the spatial registration of two 3d laser scans of the same human face usually produc...
详细信息
ISBN:
(数字)9781510607163
ISBN:
(纸本)9781510607156;9781510607163
We propose a new biometric approach where the tissue thickness of a person's forehead is used as a biometric feature. Given that the spatial registration of two 3d laser scans of the same human face usually produces a low error value, the principle of pointcloud registration and its error metric can be applied to human classification techniques. However, by only considering the spatial error, it is not possible to reliably verify a person's identity. We propose to use a novel near-infrared laser-based head tracking system to determine an additional feature, the tissue thickness, and include this in the error metric. Using MRI as a ground truth, data from the foreheads of 30 subjects was collected from which a 4d reference pointcloud was created for each subject. The measurements from the near-infrared system were registered with all reference pointclouds using the ICP algorithm. Afterwards, the spatial and tissue thickness errors were extracted, forming a 2d feature space. For all subjects, the lowest feature distance resulted from the registration of a measurement and the reference pointcloud of the same person. The combined registration error features yielded two clusters in the feature space, one from the same subject and another from the other subjects. When only the tissue thickness error was considered, these clusters were less distinct but still present. These findings could help to raise safety standards for head and neck cancer patients and lays the foundation for a future human identification technique.
暂无评论