Positioning method and system supporting online learning

• Main Authors: YING, RENDONG, JI XINGWU, NIU HAOCHEN, ZHANG LANTAO, LIU PEILIN, WEN FEI
• Format: Patent
• Language: Chinese; English
• Published: 26.04.2024
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; IMAGE DATA PROCESSING OR GENERATION, IN GENERAL; PHYSICS

The invention discloses a positioning method and system supporting online learning, and relates to the field of computer vision, a depth estimation module estimates a dense depth map of an image by using a depth estimation algorithm based on a low-rank adapter, and realizes online learning of depth estimation of a target domain by using a self-supervision mechanism; the online learning module calculates a sorting loss function and an uncertainty loss function by using the dense depth map and the pose information, and supervises the training of a depth estimation network; the feature selection module extracts ORB features and descriptors according to the image, and calculates pose information by using estimated uncertainty and a dense depth graph; the pose estimation module calculates the scale information of the pose of the image by using the alignment of the dense depth map and the sparse depth map, and ensures the scale consistency of the system in the positioning process. By designing the low-rank adapter,