Rapid visual orientation and remote monitoring system and method based on embedded mobile robot

Abstract

The invention relates to rapid visual orientation and remote monitoring system and method based on an embedded mobile robot. The system mainly comprises an embedded base plate, an embedded core plate, an image data collecting module, a digital signal processing module, a network communication module, a storage module, a USB interface and a terminal device. The embedded core plate collects image information by the image data collecting module, forwards the image information to the digital signal processing module, rapidly extracts key characteristic points of the images according to improved SIFT (Scale Invariant Feature Transform) algorithm and compares the key characteristic points with the image characteristic point information of the special marker stored in the storage module so as to identify the images, further determine the position of the mobile robot according to coordinates of the special marker and complete visual orientation of the mobile robot. After determining the position of the marker, the digital signal processing module sends the positioning information through a communication module to the terminal device. The system is compact and reliable in structure, and can rapidly and effectively identify the special marker during remote orientation of the mobile robot which is oblique, deformed, dirty, partially sheltered and has a complicated background and changing light.

Description

technical field [0001] The invention relates to a mobile robot navigation technology, in particular to a mobile robot remote fast visual positioning system and method combining embedded computer technology and digital image recognition technology based on improved SIFT algorithm. Background technique [0002] As one of the core and basic technologies of the visual positioning system of mobile robots, image recognition technology is the frontier and hotspot in the field of visual positioning of mobile robots. It is widely used, and its image matching algorithm, SIFT algorithm, is gradually adopted by people. [0003] The SIFT algorithm is an algorithm for extracting local features, looking for extreme points in the scale space, and extracting position, scale, and rotation invariants. The SIFT feature point vector extracted by using the SIFT algorithm has the following advantages: SIFT feature is a local feature of the image, which remains invariant to rotation, scaling, and ...

AI Technical Summary

Problems solved by technology

[0004] At present, in order to improve the real-time performance of visual recognition, the visual positioning of mobile robots mostly uses powerful industrial computers for visual recognition calculations, which results in a large robot system structure, high redundancy, high power consumption, weak adaptability to the environment, and mobile The performance is not good, the price is expensive, and the original SIFT algorithm requires a large amount of calculation, which reduces the real-time performance of mobile robot positioning

Images