A method for measuring flow velocity based on computer vision

Abstract

The invention discloses a method for measuring flow velocity based on computer vision, which comprises the following steps: selecting a flow measurement section on a river to be measured, fixing and erecting a camera, and adjusting the angle of the camera so that the selected velocity measurement section completely enters the field of view of the camera; Set up a calibration plate on the water surface within the field of view of the camera, calibrate the pixel coordinates of the camera and the physical coordinates of the water flow plane, and fix the camera angle and focal length of the camera; the images of the flow measurement section are obtained through the camera, according to the two consecutive frames of images The key point changes, calculate the displacement vector of the key point based on the pixel coordinates, convert the calibration relationship between the pixel coordinates and the physical coordinates in step S2 to obtain the displacement vector of the key point based on the physical coordinates, combine the frame rate of the camera, and calculate the flow velocity of the key point . The invention has the advantages that the calibration method is simple and controllable, can reduce the amount of calculation, has low requirements on hardware resources, and the like.

Description

technical field [0001] The invention relates to the technical field of flow velocity measurement, in particular to a flow velocity measurement method based on computer vision. Background technique [0002] After years of development and accumulation, artificial intelligence (AI) technology has been widely used in all walks of life, especially in 2016 after the AlphaGo artificial intelligence developed by Google defeated the world champion Lee Sedol 4:1, AI technology once again attracted the attention of the world , has shined in various application scenarios in recent years. As an important part of artificial intelligence, computer vision (Computer Vision) technology and deep learning (Deep learning) methods have made great progress. Face recognition, industrial detection, video structuring, pedestrian detection, behavior analysis, etc. have been continuously integrated and used in various fields, such as smart security, smart communities, smart construction sites, smart c...

AI Technical Summary

Problems solved by technology

[0003] At present, when computer vision is used to measure water flow velocity, existing methods often have the disadvantages of high hardware resource requirements and large computational loads.

Images