In practical industrial scenarios， the significant divergence in bearing state evolution caused by complex operating conditions imposes dual constraints on bearing life prediction： incomplete mechanism understanding and distribution shift in monitoring data. Although the existing data-driven methods perform well under stable conditions， they suffer from strong reliance on annotated data and limited generalization capability under rappidly changing conditions. To address these issues，a method for Bearing Life Prediction based on Dynamic Knowledge embedding （DK-BLP） was proposed in this paper. The method dynamically integrated data-driven features and domain prior knowledge. Firstly， a dynamic knowledge graph was constructed， the prior knowledge of bearing degradation was encoded into computable triplets， and a sliding window confidence mechanism was introduced， so as to realize adaptive updating of domain knowledge. Second， relational graph convolutional networks were used to extract physically meaningful embedding vectors， which were then fused cross-modally with time-frequency features of vibration signals extracted by hierarchical convolutional networks. Finally， dynamic interactions between features and knowledge were modeled by using a multi-head self-attention Transformer， thereby allowing the model to adaptively balance the contributions of data features and mechanistic knowledge. Experimental results on the PHM2012 and XJTU datasets demonstrate that the proposed method significantly improves the accuracy of cross-condition bearing life prediction.

Real-time semantic segmentation methods often use dual-branch structures to store shallow spatial information and deep semantic information of images respectively. However， current real-time semantic segmentation methods based on dual-branch structure focus on mining semantic features and ignore the maintenance of spatial features， which make the network unable to accurately capture detailed features such as boundaries and textures of objects in the image， and the final segmentation effect not good. To solve the above problems， a Dual-Branch real-time semantic segmentation Network based on Detail Enhancement （DEDBNet） was proposed to enhance spatial detail information in multiple stages. First， a Detail-Enhanced Bidirectional Interaction Module （DEBIM） was proposed. In the interaction stage between branches， a lightweight spatial attention mechanism was used to enhance the ability of high-resolution feature maps to express detailed information， and promote the flow of spatial detail features on the high and low branches， improving the network’s ability to learn detailed information. Second， a Local Detail Attention Feature Fusion （LDAFF） module was designed to model the global semantic information and local spatial information at the same time in the process of feature fusion at the ends of the two branches， so as to solve the problem of discontinuity of details between feature maps at different levels. In addition， boundary loss was introduced to guide the learning of object boundary information by the network shallow layers without affecting the speed of the model. The proposed network achieved a mean Intersection over Union （mIoU） of 78.2% on the Cityscapes validation set at a speed of 92.3 frame/s， and an mIoU of 79.2% on the CamVid test set at a speed of 202.8 frame/s； compared with Deep Dual Resolution Network （DDRNet-23-slim）， the mIoU of the proposed network increased by 1.1 and 4.5 percentage points respectively. The experimental results show that DEDBNet can accurately segment scene images and meet real-time requirements.

In order to catch the edge information of high resolution remote sensing image more effectively, a new method to get image edge was proposed. Firstly, the main information was collected by Principal Component Analysis (PCA) transform. Secondly, the information was divided with symletsA wavelet while the image in each scale was processed with morphological operators. Finally, the edge of image was enhanced by implementing correlation filtering to the image in the same scale with filtering algorithm of wavelet phase, and the edge information was caught by partitioning the image with OTSU algorithm. The results show that, compared with the existing algorithms, the edge of image is located more accurately and the edge detection effect is more evident with this method.