A Object-augmented Semantic Mapping System for Indoor Mobile Robots

With the rapid development of artificial intelligence technology, indoor mobile robots are widely applied in human's daily life. Traditional SLAM-based methods mostly rely on low-level geometric features, such as points, lines, etc., which cannot achieve human-robot interaction and intelligent decision-making. In this paper, we propose an object-augmented semantic mapping method, which exploits the technique of object detection and Joint calibration to construct the object semantics of indoor environments. In order to improve the mapping accuracy for robots, we perform the data association for maintaining the temporal consistency of semantic representations. Experimental results indicate that the proposed object-augmented semantic mapping system exhibits great performance in the accuracy of semantic mapping.