US 12,406,504 B2
Edge computing-based additive manufacturing real-time monitoring method and system
Hui Li, Hubei (CN); Shengnan Shen, Hubei (CN); Zhixin Bao, Hubei (CN); Wenkang Zhu, Hubei (CN); and Yu Zhang, Hubei (CN)
Assigned to WUHAN UNIVERSITY, Hubei (CN)
Filed by WUHAN UNIVERSITY, Hubei (CN)
Filed on Oct. 28, 2024, as Appl. No. 18/929,566.
Claims priority of application No. 202410010843.3 (CN), filed on Jan. 3, 2024.
Prior Publication US 2025/0218181 A1, Jul. 3, 2025
Int. Cl. G06V 20/52 (2022.01); G06T 3/4053 (2024.01); G06V 10/77 (2022.01); G06V 10/774 (2022.01); G06V 10/82 (2022.01)
CPC G06V 20/52 (2022.01) [G06T 3/4053 (2013.01); G06V 10/7715 (2022.01); G06V 10/774 (2022.01); G06V 10/82 (2022.01)] 10 Claims
OG exemplary drawing
 
1. An edge computing-based real-time monitoring method for additive manufacturing, comprising: establishing a dataset for additive manufacturing, wherein the dataset comprises a video super-resolution dataset and feature segmentation dataset; constructing an real-time monitoring model for additive manufacturing, wherein the real-time monitoring model comprises a video super-resolution model and a feature segmentation model, wherein the video super-resolution model takes video sequences as input, the feature segmentation model takes the output of the video super-resolution model as input, and the feature segmentation model outputs feature segmentation maps, wherein the video super-resolution model comprises a spatiotemporal encoder, a residual feature extraction module, a query matrix generation module, a key matrix generation module, and a value matrix generation module, wherein pixel resolution of any frame image in the input video sequence is (h0/r, w0/r, 1), where h0 and w0 are number of pixels in the height and width directions of the image, respectively, and r is magnification factor of super-resolution reconstruction, the video sequence, after being encoded by the spatiotemporal encoder, is input to the residual feature extraction module to obtain a feature map matrix, the feature map matrix is then input to the query matrix generation module, the key matrix generation module, and the value matrix generation module to obtain a query matrix, a key matrix, and a value matrix, respectively, wherein, by the video super-resolution model, obtaining a total feature map based on the query matrix, the key matrix, and the value matrix, wherein, by the video super-resolution model, rearranging the total feature map by pixels to obtain a super-resolution reconstruction result with a pixel resolution of (h0, w0, 1); training the real-time monitoring model using the dataset and obtaining a trained real-time monitoring model; and deploying the trained real-time monitoring model on edge computing equipment, and use the edge computing equipment to obtain real-time monitoring information of additive manufacturing.