Visual SLAM loopback detection method combining semantic features and bag-of-words model
The invention belongs to the technical field of visual SLAM, and discloses a visual SLAM loopback detection method combining semantic features and a bag-of-word model. The method comprises the following steps: collecting motion video data of a factory, obtaining each frame of picture from the motion...
Saved in:
Main Authors | , , , , , , , , , |
---|---|
Format | Patent |
Language | Chinese English |
Published |
25.10.2022
|
Subjects | |
Online Access | Get full text |
Cover
Loading…
Summary: | The invention belongs to the technical field of visual SLAM, and discloses a visual SLAM loopback detection method combining semantic features and a bag-of-word model. The method comprises the following steps: collecting motion video data of a factory, obtaining each frame of picture from the motion video data, judging a similarity value between any two frames of pictures, deleting redundant pictures based on the similarity value to obtain training data, and carrying out loopback detection on the training data. Training the U-net network based on the training data to obtain a trained U-net network; the method comprises the following steps: collecting a picture of a factory, obtaining an RGB image of the picture, and carrying out semantic segmentation on the RGB image by using a training U-net network to obtain a semantic tag; the semantic tags are clustered; similarity comparison is carried out on the clustered semantic tags, and candidate key frames are obtained; and after the key frame is detected to be ins |
---|---|
Bibliography: | Application Number: CN202210896120 |