Research on the effect of image size on real-time performance of robot vision positioning

Abstract

In order to improve the real-time performance of visual positioning of the indoor mobile robot, the researchers found that the shape and size of the positioned image have a great influence on the real-time performance of the positioning calculation. In order to verify the conclusion and find the appropriate image shape and size to meet the robot's visual positioning requirements, this paper adopts four different shapes, such as quadrilateral and circular, and uses SURF algorithm to extract and recognize the features of the image. The effect of image shape and size on real-time localization is studied from two aspects: the localization of different shape models under the same size by the visual robot and the localization of the different shape models by the visual robot. It is found that the accuracy and real time of positioning squares and circles are higher than the accuracy and real time of positioning triangles and hexagons under the same size. And when the image size ratio is between 40 and 60% of the original image, the change of the number of feature points is relatively stable and the number of feature points is moderate. It can improve the real-time performance of mobile robot vision localization under the premise of a certain positioning accuracy.

A velocity program using the Kanade–Lucas–Tomasi feature‐tracking algorithm with demonstration for pressure and electroosmosis conditions

“…The impact of image size on computational speeds in computer vision is demonstrated by Lyu et al. [55]. In this study, feature extraction time was tested in a real‐time visual robot which used the SURF (Speeded‐Up Robust Features) algorithm [55].…”

“…[55]. In this study, feature extraction time was tested in a real‐time visual robot which used the SURF (Speeded‐Up Robust Features) algorithm [55]. It was shown that the computation time and image size have a linearly positive relationship.…”

“…It was shown that the computation time and image size have a linearly positive relationship. An image was scaled from 10 to 100% of the original image, where computation time to detect all features in the image ranged from 3000 to over 9000 ms [55] in revised manuscript.…”

Enhanced Positioning Algorithm Using a Single Image in an LCD-Camera System by Mesh Elements’ Recalculation and Angle Error Orientation

“…Vision positioning systems are increasingly common in process automation [6][7][8][9][10], autonomous driving [11][12][13][14][15], or augmented reality assistants [16][17][18][19][20]. Indeed, this is one of the most promising elements in the Industry 4.0 revolution.…”

Analysis and implementation of an artificial vision method for micropositioning based on a camera-screen system

“…Los sistemas de posicionamiento por visión son cada vez más comunes en la automatización de procesos [104][105][106][107], la conducción autónoma [108][109][110][111][112], o los asistentes de realidad aumentada [113][114][115][116][117]. De hecho, este es uno de los elementos más prometedores en la revolución de la Industria 4.0.…”