CornerNet and CenterNet-based water meter character wheel area positioning method

Abstract

The invention relates to water meter image recognition, in particular to a CornerNet and CenterNet-based water meter character wheel area positioning method, which comprises the following steps of detecting the type of a water meter in a collected image by using a water meter type recognition model, calling a corresponding CornerNet deep neural network based on the type of the water meter to perform diagonal positioning on a character wheel area in the collected image, calling a corresponding Center Net deep neural network based on the type of the water meter to carry out center positioning on a character wheel area in the collected image, and carrying out image correction on the collected image by using a diagonal point and a central point; according to the technical scheme provided by the invention, the defect that the water meter character wheel region part in the acquired image cannot be accurately positioned in the prior art can be effectively overcome.

Description

technical field [0001] The invention relates to water meter image recognition, in particular to a water meter character wheel area positioning method based on CornerNet and CenterNet. Background technique [0002] At present, most of the water companies in various places use manual meter reading, but this method has always had problems such as wrong reading, missing reading, and inaccurate valuation. [0003] In order to solve various problems arising from manual meter reading, related solutions in the prior art can generally be divided into two types: bus-type remote meter reading and wireless remote meter reading. The bus-type remote meter reading scheme is to set up electronic water meters in the community, and transmit the data of each electronic water meter to the server through the bus. The bus-type remote meter reading solution has the advantages of stable transmission and low price. The disadvantage is that it requires a large amount of construction and the need to ...

AI Technical Summary

Problems solved by technology

[0008] Aiming at the above-mentioned shortcomings existing in the prior art, the present invention provides a method for locating the character wheel area of ​​a water meter based on CornerNet and CenterNet, which can effectively overcome the problem existing in the prior art that it is impossible to accurately locate the part of the character wheel area of ​​the water meter in the collected image. defect

Images