Discriminative bit selection hashing in RGB-D based object recognition for robot vision

Abstract

Purpose The purpose of this paper is to promote the efficiency of RGB-depth (RGB-D)-based object recognition in robot vision and find discriminative binary representations for RGB-D based objects. Design/methodology/approach To promote the efficiency of RGB-D-based object recognition in robot vision, this paper applies hashing methods to RGB-D-based object recognition by utilizing the approximate nearest neighbors (ANN) to vote for the final result. To improve the object recognition accuracy in robot vision, an “Encoding+Selection” binary representation generation pattern is proposed. “Encoding+Selection” pattern can generate more discriminative binary representations for RGB-D-based objects. Moreover, label information is utilized to enhance the discrimination of each bit, which guarantees that the most discriminative bits can be selected. Findings The experiment results validate that the ANN-based voting recognition method is more efficient and effective compared to traditional recognition method in RGB-D-based object recognition for robot vision. Moreover, the effectiveness of the proposed bit selection method is also validated to be effective. Originality/value Hashing learning is applied to RGB-D-based object recognition, which significantly promotes the recognition efficiency for robot vision while maintaining high recognition accuracy. Besides, the “Encoding+Selection” pattern is utilized in the process of binary encoding, which effectively enhances the discrimination of binary representations for objects.

Rethinking of Construction Robot in the Whole Project Life Cycle

“…Important precondition of the practical application of robot is to have the technology of perceiving the external environment. With the approaches like machine learning(MI) (Lawaniya 2020), artificial intelligence(AI), approximate nearest neighbors (ANN) (Wahrmann 2019), pre-assumption, computer vision system could assist robot to achieve navigation (Feng 2019), identification and modeling in completely unknown environments. Yet, there are some challenges such as object recognition.…”

Performance enhancement of two-camera robotic system using adaptive gain approach

“…Robotic systems need feedback from its environment for the generation of control signals for characterisation and decision making in industrial applications. Non-touching sensors like camera (Qu et al, 2017), RGB-D sensors (Feng et al, 2019) and laser technology (Lindner et al, 2016) are revelations in this area. The use of these technologies are numerous and varied in robotic industries and the former (Sun et al, 2018) has more reported applications compared with the latter (Lindner et al, 2017;Rodriguez-Quinonez et al, 2014).…”