Loop Closure Detection Based on Generative Adversarial Networks for Simultaneous Localization and Mapping Systems

摘要

Loop closure detection is important in simultaneous localization and mapping (SLAM) systems. In this paper, Generative Adversarial Networks (GAN), an unsupervised deep architecture is employed to detect the loop closure for vision-based SLAM systems. Instead of extracting handcrafted features like SIFT, SURF or ORB. Generative Adversarial Networks are based on image features. Similar to the task about judging whether a loop closure is real or fake, the discriminator in DCGAN is used to judge whether the input image is real or fake. Therefore, the detection model based on DCGAN is suitable for the task. Experimental results show that compared to traditional methods like BoW (Bag of Words), DCGAN is more effective to detect the loop closure in outdoor environments.