Loopback detection method and device based on semantic point cloud

Abstract

The embodiment of the invention discloses a loopback detection method and device based on semantic point cloud. The method comprises the steps of obtaining semantic point cloud corresponding to a current video frame; determining a first geometric center of the to-be-processed earliest acquired semantic point cloud, and calculating a second geometric center of the semantic point cloud correspondingto the current video frame and a first distance between the first geometric center and the second geometric center; when the first distance is greater than a first preset threshold, constructing a current semantic point cloud grid sub-graph; when the current semantic point cloud lattice sub-graph and the adjacent semantic point cloud lattice sub-graph do not contain the same loopback semantic information, constructing a current semantic point cloud sub-graph; calculating a second distance between the target geometric center of the point cloud corresponding to each piece of loopback semantic information contained in the current semantic point cloud sub-graph and the geometric center of the point cloud corresponding to the same type of loopback semantic information contained in each constructed semantic point cloud sub-graph; and determining that a loopback triggering condition is met when the second distancesmaller than a second preset threshold value exists. According to the embodiment of the invention, the loopback detection accuracy can be improved.

Description

technical field [0001] The present invention relates to the technical field of intelligent driving, in particular to a method and device for loop closure detection based on semantic point cloud. Background technique [0002] In the field of autonomous driving, in the process of building a high-precision map using a surround-view vision solution, unmanned vehicles cannot obtain global and accurate observations, so cumulative errors in map building are inevitable. In order to eliminate the cumulative error and build a globally consistent map, when the vehicle travels for a period of time and re-drives to the area where the map has been constructed, loop closure detection is required to further optimize the loop closure. [0003] The known loop-closing detection method mainly uses an image bag of words (Bag of Words) to trigger loop-closing. Specifically, the currently observed picture information is compared with the previously observed picture information, and when two pictu...

AI Technical Summary

Problems solved by technology

[0004] However, in practical applications, when vehicles pass through the same area, their driving routes are not necessarily exactly the same, and the similarity of the observed pictures will also be reduced.

Therefore, in the above method, only relying on the similarity of the two pictures to trigger the loopback is less accurate, which means that the loopback cannot be effectively triggered in many cases.

Therefore, in order to solve the problem of low accuracy of loopback triggering in the prior art, there is an urgent need for a loopback detection method

Images