Integrated data intelligent labeling method and system

Abstract

The invention discloses an integrated data intelligent labeling method and system, and the method comprises the steps: receiving a labeling task request of a user, and starting a corresponding labeling task module to execute a labeling task; receiving a first to-be-labeled sample set, labeling template setting data and parameter setting data; calling a matched preprocessing method in a preset preprocessing tool library to preprocess the to-be-labeled sample set; according to the labeling template setting data and the parameter setting data, sequentially performing judgment according to a priority sequence in a plurality of labeling methods in the started labeling task module to obtain a labeling method which is optimally matched with the labeling task. According to the method, the integrated data labeling of the to-be-labeled sample set of electric power voice, text, video and image types is realized, the unification of a sample data preprocessing method and a sample labeling method is further realized, and the problems of different labeling modes of electric power data companies, repeated data labeling operation and low efficiency are avoided.

Description

technical field [0001] The invention relates to the technical field of data processing, in particular to an integrated data intelligent labeling method and system. Background technique [0002] At present, State Grid Corporation has a variety of artificial intelligence application requirements, such as equipment image recognition, voiceprint recognition and other business fields, and then has the following technical problems in the labeling of power images, voiceprint data and other types of data: [0003] (1) The amount of labeling of multi-type data is large, the cost is high, there is no unified labeling standard, and the labeling method needs to be improved and enriched: the State Grid has a large amount of power-related data, and the data is comprehensive. However, the current labeling method of the State Grid relies on manual labor. , marked as overpriced. Moreover, each company of the State Grid has different labeling methods, which leads to the need for each company...

AI Technical Summary

Problems solved by technology

[0003] (1) The amount of multi-type data labeling is large, the cost is high, there is no unified labeling standard, and the labeling method needs to be improved and enriched: the State Grid has a large amount of power-related data, and the data is comprehensive. However, the current labeling method of the State Grid relies on manual labor. , mark the price too high

Moreover, the labeling methods of each company of the State Grid are different, resulting in the need for each company to reprocess when using other companies' data, resulting in low team efficiency, repeated data operations, and a waste of a lot of resources

[0004] (2) The lack of a unified management platform will lead to increased data security risks and high management costs. It is urgent to develop a multi-type data integrated labeling platform for management; each company of the State Grid uses different data management platforms, and due to limited storage conditions, data loss, resulting in data security risks

Images