Optical splitter tail fiber label detection method and system based on semantic segmentation algorithm

Abstract

The invention discloses an optical splitter tail fiber label detection method and a system based on a semantic segmentation algorithm. The method comprises the steps of performing labeling based on animage labeling tool; collecting a marked picture Json file, and analyzing the marked picture Json file based on any picture Json file; classifying the processed label image pixel points, segmenting apixel comprising a label and a correspondingly connected port, and constructing a semantic segmentation model; performing label detection based on the semantic segmentation model, and a pixel picturecomprises a label and a label connection port and a picture for port number positioning through a tail fiber label are output; and detecting all port area positions according to the red pixel point positions in the label graph detected by the test picture and a port detection algorithm, and calculating and screening the port with the highest overlapping rate of the target pixel point position andthe pixel point of any port detection area. The construction quality inspection efficiency of the optical splitter is improved, the labor cost is reduced, the construction process is normalized and standardized, and the construction quality is improved.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a method and system for detecting a label of an optical splitter pigtail based on a semantic segmentation algorithm. Background technique [0002] As home broadband enters thousands of households, the quality of installation and maintenance work has become one of the important factors that affect users' network experience. The core issue in the broadband installation and maintenance scenario is how to conduct reasonable and efficient quality inspections on the installation and maintenance process, so that the installation and maintenance personnel can meet the construction specifications, ensure construction accuracy, reduce labor costs in the installation and maintenance quality inspection, and improve quality inspection. efficiency. [0003] At present, the traditional installation and maintenance quality inspection relies on manual review of the photos uploaded by th...

AI Technical Summary

Problems solved by technology

However, the number of installation work orders is very large every day, and it is unrealistic to do full manual quality inspection. The quality of installation can only be inspected through random inspection. The scope of random inspection is small, and the coverage rate is very low, and it consumes a lot of manual repetitive labor.

Images