Slam closed-loop detection method based on particle swarm optimization algorithm

Abstract

The invention discloses a SLAM closed loop detection method based on a particle swarm optimization algorithm, and mainly solves the problems that a Bag-of-Word method in the existing closed loop detection technology is complex in an offline training process; the detection method comprises the following steps: 1, determining whether an obtained current frame image is the key frame image or not; 2,calculating descriptors of the current key frame image; 3, determining whether the current key frame image is the first frame key frame image or not; 4, creating a frame image description sub-library;5, expanding the frame image description sub-library; 6, determining whether the key frame image number in the frame image description sub-library equals to 50 or not; 7, obtaining the optimal frameimage; 8, matching the current key frame image with the optimal frame image; 9, determining whether the violence matching pair number equals to 25 or not; 10, outputting the optimal frame image satisfying closed loop conditions.

Description

technical field [0001] The invention belongs to the field of image technology, and further relates to a real-time positioning and composition SLAM (simultaneous localization and mapping) closed-loop detection method based on particle swarm optimization algorithm in the field of robot vision technology. The present invention obtains the picture to be detected through the camera, and searches for a frame of picture most similar to the picture to be detected through the particle swarm optimization algorithm, which can be used to realize the closed-loop detection method. Background technique [0002] At present, robot technology is known as one of the ten fields with great development potential in the 21st century. Real-time positioning and composition SLAM means that the robot determines its own spatial position through sensor information in an unknown environment, and establishes an environmental model of the space in which it is located. In recent years, with the emergence o...

AI Technical Summary

Problems solved by technology

However, the shortcomings of this method are: offline training of feature points takes up a lot of memory and takes a lot of time

The shortcomings of this method are: because the visual word book generated by training quantifies the local feature point descriptors, it does not consider the data association of the local scale invariant FAST feature points in the scene image, causing the robot to use different locations Highly similar scenes are mistaken for the same scene, which leads to errors in loop closure detection

The disadvantage of the method proposed in this patent application is that the closed-loop detection effect of the bag-of-words BOW method is highly dependent on the number of words, and the high-accuracy closed-loop detection needs to maintain a word book with increasing size, and the detection process is relatively complicated.

Images