Non-Lambertian surfaces and their challenges for visual SLAM

Abstract

Non-Lambertian surfaces are special surfaces that can cause specific type of reflectances called specularities. These specularities pose a potential issue in industrial SLAM, where non-Lambertian surfaces are aplenty naturally, as specularities can violate two common and basic assumptions of computer vision, namely the brightness and colour constancies, which assume that the brightness and colour of same real-world points stays constant across images. The performance of industrial navigation systems can as such be hindered by specularities. This paper reviews fundamental surface reflectance models, modern state-of-the-art computer vision algorithms and two public datasets, KITTI and DiLiGenT, related to non-Lambertian surfaces' research. A new dataset, SPINS, is presented for the purpose of studying non-Lambertian surfaces in navigation and an empirical performance evaluation with ORB SLAM 3 is performed on the data. The paper concludes with discussion about the results of empirical evaluation and the findings of the survey.