Water level monitoring method based on image identification

Abstract

The invention discloses a water level monitoring method based on image identification. The water level monitoring method based on the image identification comprises that an image of a staff gauge used for marking a water level of a water level area to be detected is collected through a camera arranged in the water level area to be detected; the image of the staff gauge is uploaded to the server; the server performs normalization processing on the size of the image of the staff gauge and performs calibration on a measuring range of the staff gauge and key scales of the staff gauge; the server adopts a machine study SVM model to perform training categorization according to the type of the staff gauge images; the current staff gauge image to be identified is obtained through the camera and is uploaded to the server; an optimal algorithm is chosen through a machine study SVM model to calculate a Y coordinate of a current water level to be identified in the current image to be identified of the staff gauge, and the Y coordinate is converted to the current water level value. The invention brings forward a water monitoring method based on image identification, adopts a non-contact water level measuring and calculation method, can accurately and directly identify the current water level height and has characteristics of small reconstruction work, strong stability and a wind application range.

Description

technical field [0001] The invention relates to the technical field of computer image recognition, in particular to a water level monitoring method based on image recognition. Background technique [0002] Existing water level monitoring systems are complex to install, inflexible in application scenarios, cumbersome in configuration, difficult to adapt to severe floods, and high in construction costs. Moreover, most of the existing water level monitoring methods use reference objects whose accuracy depends greatly on the recognition results, such as photographing objects and knowing the height of the reference object, slope angle, horizontal distance and height of the camera, resulting in inability to accurately monitor the water level. . At the same time, the existing water level monitoring methods do not consider and solve the effects of reflection, refraction and water gauge dirt on the water level, which further causes problems such as poor applicability of the monitori...

AI Technical Summary

Problems solved by technology

[0002] The existing water level monitoring system is complicated to install, the application scenarios are not flexible, the configuration is cumbersome, some are difficult to adapt to the severe flood process, and the construction cost is high

Moreover, most of the existing water level monitoring methods use reference objects whose accuracy depends greatly on the recognition results, such as photographing objects and knowing the height of the reference object, slope angle, horizontal distance and height of the camera, resulting in inability to accurately monitor the water level.

At the same time, the existing water level monitoring methods do not consider and solve the impact of reflection, refraction and water gauge dirt on the water level, which further causes problems such as poor applicability of the monitoring method

Images