Pointer type instrument indication value reading method based on CNN neural network

Abstract

The invention discloses a pointer type instrument indication value reading method based on a CNN neural network. The method can detect an instrument in an image and the type of the instrument and calculate an indication value of the instrument through instrument detection and type judgment, instrument feature point detection, instrument indication value calculation and the like. According to the invention, a pointer instrument in an image can be identified more accurately, and the indication value of the pointer instrument can be read; the method can be applied to occasions requiring automaticreading of the pointer instrument, can replace manual work to intelligently monitor the instrument, and can give an early warning when a preset value is exceeded.

Description

technical field [0001] The invention relates to the technical field of image recognition, in particular to a method for reading indication values ​​of pointer instruments based on a CNN neural network. Background technique [0002] Pointer instruments have fast response speed, strong anti-shock and anti-interference ability, clear display, and low price, and mechanical pointer instruments basically do not have the hysteresis of digital instrument response, and pointer instruments also have the ability to display overload ratios; some occasions Digital meters are not applicable, such as measuring the polarity of voltage and current transformers, digital meters cannot respond at all, and only pointer meters can be used. [0003] Although electronic technology is very developed today, pointer instruments are still widely used in some industrial production due to their irreplaceable advantages of digital instruments. With the vigorous development of modern industry 4.0, the tra...

AI Technical Summary

Problems solved by technology

[0004] However, at present, pointer instruments mostly adopt the principle of automatic recognition based on machine vision. The technology is mainly traditional image processing, which has its defects. It usually requires a stable light source to ensure that the system acquisition environment will not have large fluctuations. It is usually difficult to apply to some old equipment and equipment that is not convenient to modify, and it will be affected by unfavorable factors such as illumination changes, background interference, shadows, rotation, and fast movement in the natural environment. The adaptability is poor, and the analysis effect will be very good. Large volatility, high requirements for monitoring equipment installation, poor anti-shake ability, etc.

Images