Indoor scene and object simultaneous recognition and modeling method

Abstract

The invention discloses an indoor scene and object simultaneous recognition and modeling method, and the method comprises the steps: inputting an RGB-D image; carrying out the segmentation; extracting SIFT features and FPFH features; carrying out the fusion of the SIFT features and FPFH features; carrying out the object recognition; and carrying out the object modeling. The object modeling comprises the steps: calculating the posture relation among object parts, setting a threshold value, carrying out the fusion of two object parts and taking the two object parts after fusion as a node in a view image if the posture change is less than a threshold value, or else enabling the two object parts to remain in the view image, i.e., taking the two object parts as two nodes. Compared with the prior art, the method can achieve the online recognition and modeling of the object, proposes an improved view image method, reduces the redundant data, reduces the data storage burden, and improves the recognition efficiency.

Description

technical field [0001] The invention relates to the field of information processing, in particular to a method for simultaneously recognizing and modeling indoor scene objects. Background technique [0002] At present, in the field of robot object recognition and scene understanding research, a large number of algorithm researches focus on matching test pictures with training data sets or training classifiers to achieve recognition, and generally do not consider the pose of objects. For example, Liang Mingjie et al. proposed an object model representation method based on the view graph, and provided a probabilistic observation model of the object on the basis of the object representation. Finally, the identification and modeling were attributed to probabilistic reasoning, and realized by maximum likelihood estimation. Simultaneous optimization of recognition and modeling; Alvaro Collet Romea proposed an indoor object recognition and modeling method based on a constraint fram...

AI Technical Summary

Problems solved by technology

[0003] However, precise attitude information is used in many object recognition applications, such as robot operation, path planning, augmented reality, etc., and the algorithm needs training data, so it cannot achieve online recognition effect, which seriously affects the robot's performance in indoor scenes. performance

In terms of modeling, if the robot walks in the indoor environment for a long time, the observation part of the indoor object will be obtained frequently, so that the established model will generate redundant data. The longer the robot walks, the more observation data of the same object will be obtained. More, the larger the object model is, the more redundant data will increase, which is not conducive to object modeling and data storage, but also has a negative impact on the efficiency of later object recognition

Images