Method and related device for automatic focusing of fiber end face based on neural network

Abstract

The invention discloses an optical fiber end face automatic focusing method based on a neural network. An optical fiber end face image of an initial position and an optical fiber end face image movingto a second position in a preset stepping manner are obtained; the data is input into a trained neural network for judgment; the neural network generates corresponding moving direction labels according to the definition change states of the two input pictures; according to moving direction labels, the next moving direction and moving step of the optical fiber clamp are determined; the neural network is continuously and circularly called; continuous judgment of definition change states is carried out on the optical fiber end face images at different axial positions; according to the method andthe device, the optimal focusing position can be found quickly and accurately by combining the focusing mode and the focusing mode until the focusing clearness degree is achieved, efficient and high-precision automatic focusing can be achieved without manual focusing, and a focusing window has pertinence and can better resist interference caused by the background.

Description

technical field [0001] The invention relates to the field of computer software, in particular to a neural network-based method, device, equipment and storage medium for automatic focusing of an optical fiber end face. Background technique [0002] As the carrier of information, optical fiber is an important part of optical fiber communication system. Optical fiber is widely used in communication and other fields. In optical fiber communication, the active connection of optical fiber is realized by optical fiber connector, and the cleanliness of optical fiber end face It has a decisive impact on the performance of the connector, such as permanent damage to the fiber end face during the polishing process or online business operations such as fiber insertion and removal, such as scratches, cracks, etc. The fiber end face will also be subject to various Temporary pollution, such as dirt, oil stains, water or cleaning agent residues, will affect its transmission performance, whic...

AI Technical Summary

Problems solved by technology

[0003] Traditional optical fiber end-face defect detection usually adopts manual detection and traditional image processing algorithm auto-focus method. First, a clear image of the optical fiber end face is obtained by manual focusing, and then manually judged by naked eye observation. This detection method is inefficient and the detection results are subjective. It is very sensitive, and manual focusing can neither guarantee the accuracy of focusing, nor is it conducive to the automation of the focusing process. In addition, the traditional image processing algorithm is used for automatic focusing, and its focusing effect is greatly limited by the equipment environment in specific application scenarios. parameters, the focusing speed is slow, the focusing accuracy is low, the focusing window is fixed, and it is easy to be interfered by the focusing background, and the traditional focusing method has poor versatility, so it is necessary to select a suitable focusing evaluation function for the specific imaging environment

Images