We propose a bearing health management framework leveraging large language models (BearLLM), a novel multimodal model that unifies multiple bearing-related tasks by processing user prompts and vibration signals. Specifically, we introduce a prior knowledge-enhanced unified vibration signal representation to handle various working conditions across multiple datasets. This involves adaptively sampling the vibration signals based on the sampling rate of the sensor, incorporating the frequency domain to unify input dimensions, and using a fault-free reference signal as an auxiliary input. To extract features from vibration signals, we first train a fault classification network, then convert and align the extracted features into word embedding, and finally concatenate these with text embedding as input to an LLM. To evaluate the performance of the proposed method, we constructed the first large-scale multimodal bearing health management (MBHM) dataset, including paired vibration signals and textual descriptions. With our unified vibration signal representation, BearLLM using one set of pre-trained weights achieves state-of-the-art performance on nine publicly available fault diagnosis benchmarks, outperforming specific methods designed for individual datasets. We provide a dataset, our model, and code to inspire future research on building more capable industrial multimodal models.

The application of multiple sensors significantly enhances the accuracy of industrial fault diagnosis, but existing algorithms are structurally complex and rely heavily on extensive training data. To optimize the efficiency of diagnosis, this article proposes a lightweight time-frequency-statistical domain fusion network. The model comprises three data streams that analyze the time-domain, frequency-domain, and statistical features of vibration signals, employing an improved channel attention mechanism for weighted fusion. In addition, two model-agnostic few-shot enhancement strategies are introduced, aiming to improve accuracy where training samples are scarce by reducing signal sample variations and optimizing the distribution of signals in the feature space. By combining these techniques, the proposed method exhibits superior performance in few-shot learning on two datasets compared to other multisensor fusion methods, while also achieving higher computational speed. The results of this research are of significant importance in enhancing the fault diagnostic capabilities of multisensor systems in practical industrial applications.

This paper focuses on synthesizing high-quality and complete textures directly on the surface of 3D models within 3D space. 2D diffusion-based methods face challenges in generating 2D texture maps due to the infinite possibilities of UV mapping for a given 3D mesh. Utilizing point clouds helps circumvent variations arising from diverse mesh topologies and UV mappings. Nevertheless, achieving dense point clouds to accurately represent texture details poses a challenge due to limited computational resources. To address these challenges, we propose an efficient octree-based diffusion pipeline called TexOct. Our method starts by sampling a point cloud from the surface of a given 3D model, with each point containing texture noise values. We utilize an octree structure to efficiently represent this point cloud. Additionally, we introduce an innovative octree-based diffusion model that leverages the denoising capabilities of the Denoising Diffusion Probabilistic Model (DDPM). This model gradually reduces the texture noise on the octree nodes, resulting in the restoration of fine texture. Experimental results on ShapeNet demonstrate that TexOct effectively generates high-quality 3D textures in both unconditional and text / image-conditional scenarios.

Marchenko imaging has the particular ability to generate the subsurface image free of spurious artifacts related to internal multiples. However, conventional Marchenko imaging (C-MI) is performed based on the assumption that the sources and receivers of the recorded seismic data are placed on a flat surface, which is restrictive and hard to satisfy in mountainous areas with rugged topography. The elevation-static correction (or time shift) is a usual solution to ensure this assumption holds. But it only works well if the surface consistency is satisfied. To alleviate the limitations of C-MI in processing seismic data acquired from mountainous areas with rugged topography, we present a technique for conducting Marchenko imaging from a floating datum. The proposed Topography-Marchenko imaging (T-MI) is achieved by estimating an initial down-going focusing function between a floating datum and a focal point in the subsurface. In this work, we use seismic data corrected to a floating datum rather than a final datum as input to the iterative Marchenko scheme to retrieve Green’s functions. The retrieved Marchenko Green’s functions are further used to generate the subsurface image. The T-MI method can effectively avoid imaging distortions caused by the elevation-static correction. The proposed T-MI method is validated through applications to a synthetic model with rugged topography and a land dataset acquired from a mountainous area in Northwest Sichuan, China.

Abstract We present InstanceFusion, a robust real‐time system to detect, segment, and reconstruct instance‐level 3D objects of indoor scenes with a hand‐held RGBD camera. It combines the strengths of deep learning and traditional SLAM techniques to produce visually compelling 3D semantic models. The key success comes from our novel segmentation scheme and the efficient instance‐level data fusion, which are both implemented on GPU. Specifically, for each incoming RGBD frame, we take the advantages of the RGBD features, the 3D point cloud, and the reconstructed model to perform instance‐level segmentation. The corresponding RGBD data along with the instance ID are then fused to the surfel‐based models. In order to sufficiently store and update these data, we design and implement a new data structure using the OpenGL Shading Language. Experimental results show that our method advances the state‐of‐the‐art (SOTA) methods in instance segmentation and data fusion by a big margin. In addition, our instance segmentation improves the precision of 3D reconstruction, especially in the loop closure. InstanceFusion system runs 20.5Hz on a consumer‐level GPU, which supports a number of augmented reality (AR) applications (e.g., 3D model registration, virtual interaction, AR map) and robot applications (e.g., navigation, manipulation, grasping). To facilitate future research and reproduce our system more easily, the source code, data, and the trained model are released on Github: https://github.com/Fancomi2017/InstanceFusion .