Instrument reading neural network identification method

Abstract

The invention discloses an instrument reading neural network identification method, which belongs to the field of visual identification and automation, and comprises the following steps of: (1) acquiring a pointer instrument image: shooting and recording the pointer instrument image; (2) acquiring a correction reference; (3) training a neural network of instrument panel images at different positions and a digital display reading result, and learning parameters by applying a feedforward algorithm; and (4) performing instrument reading identification detection: inputting a pointer instrument camera image, and giving a corresponding instrument reading identification result and a confidence interval by a neural network model. The defects of an existing instrument reading extraction method areovercome, the parameter learning capacity of the artificial neural network model, the learning capacity based on the artificial neural network model and the physical significance of error feedforwardare fully utilized, the pointer instrument reading is accurately detected by processing the shot image, and the invention has the advantages of being high in real-time performance, high in accuracy and the like. Good practical values are realized.

Description

technical field [0001] The invention belongs to the field of visual recognition and automation, and in particular relates to a neural network recognition method for instrument readings. Background technique [0002] Automating instrument reading is a widespread approach in measurement system applications. For example, water meter reading meter billing and so on. The timing and periodic reading of the instrument is also used in the monitoring system. Currently, meter readings have the following methods: [0003] 1) Manual reading, such as manual meter reading of water, electricity and gas data, still uses manual methods. This method cannot be automated and is time-consuming and labor-intensive. [0004] 2) Digitization of measuring instruments, direct acquisition of digital readings, digital replacement of original instruments, reduction of labor costs, and improvement of reading efficiency. However, the initial investment cost is relatively high. For example, dismantlin...

AI Technical Summary

Problems solved by technology

[0017] Aiming at the defects of the prior art, the present invention provides an instrument calibration method based on a neural network, thereby solving the problems of the existing pointer instrument reading recognition being uncertain at the fixed position of the camera, manual reading instability, and the pointer instrument itself being inaccurate.

Images