Research on feature extraction method based on Simultaneous localization and mapping

In view of the problems of low precision of localization and detection method based on traditional navigation with non-Gauss and non-linear model in underwater, a feature map generation algorithm based on the simultaneous localization and mapping (SLAM) was proposed. Image processing and feature extraction method based on SLAM were presented in this paper, and characteristic of shipwreck were matched and extracted respectively. The experimental results proved that the new algorithm could achieve higher correlation degree of feature matching and complete the feature map reconstruction. Finally, by building the system state model and observation model based on SLAM, and using Square-Root Cubature Kalman filter based on Spherical Simplex (SS-SRCKF) algorithm to estimate and analysis the state of carrier and the characteristics, the experimental results showed that the new filtering algorithm could realize high precision navigation with small error.