Simultaneous Detection of Loop-Closures and Changed Objects

Loop-closure detection (LCD) in large non-stationary environments remains an important challenge robotic visual simultaneous localization and mapping (vSLAM). To reduce computational perceptual complexity, it is helpful if a vSLAM system has the ability to perform image change (ICD). Unlike previous applications of ICD, time-critical cannot assume offline background modeling stage, or rely on maintenance-intensive models. address this issue, we introduce novel maintenance-free ICD framework that requires no modeling. We demonstrate LCD can be reused as main process for with minimal extra cost. Based these concepts, develop component enables ICD. experiments based challenging cross-season scenarios validate efficacy proposed method.