Robot vision navigation technology based on feature recognition

Abstract

The invention provides a robot vision navigation technology. The technical scheme for solving the technical problems is as follows: a simple pattern with a direction sign is projected to an indoor top wall through strong light with a special color or special light such as infrared invisible light, an upward camera is arranged on a robot and shoots indoor top images continuously, calculation is performed on the shot images, a pixel with the same color as a sign pattern is found out, then two-dimensional coordinates of a core sign point in the sign pattern are determined according to the direction sign of the pattern and are taken as coordinates of the robot relative to the core sign point at the moment, and later, relative two-dimensional coordinates of walking paths are stored, so that accurate path planning and navigation can be realized through matching with an obstacle avoidance sensor; the pattern with the direction sign is simple and has a bright color, so that the coordinates of the core sign point in the sign pattern can be efficiently processed without need of very high hardware cost and very complex algorithm, and the problem about accurate navigation of the robot can be solved with lower cost and higher efficiency.

Description

technical field [0001] The invention relates to a technical scheme for a robot to recognize indoor top images through a camera and plan a walking route, so that the robot can accurately know where it is in the area, plan which direction to go next, and can plan the walking route. Memory, to avoid blind walking and repeated walking, can be applied to sweeping (mopping) robots, food delivery robots and other fields. Background technique [0002] At present, the route planning of sweeping (mopping) robots mainly uses sensors such as collision, infrared, ultrasonic, and laser to assist in decision-making. High-end sweeping robots use horizontal graphic image recognition algorithms, which are complicated, expensive in hardware, and low in recognition efficiency. The image does not have a very accurate sense of distance, and the object changes, and the image recognition will lose the blindness. At present, the walking routes of most sweeping (mopping) robots on the market are sev...

AI Technical Summary

Problems solved by technology

[0006] In order to overcome the problem that the existing sweeping (mopping) robots cannot accurately locate the coordinates of their own positions, the present invention provides a robot visual navigation technology, which can measure the precise coordinates of the robot's position, and the path traveled Carry out storage, plan where to go next, so that the robot does not walk randomly, does not repeat, and finally effectively cleans the entire room, and the realization of this technology has the advantages of simple algorithm and low hardware cost

Images