Power grid image intelligent annotation crowdsourcing platform and working method

Abstract

The invention relates to a power grid image intelligent annotation crowdsourcing platform and a working method, and belongs to the technical field of power grid image data processing. The working method comprises the following steps: collecting to-be-annotated picture collection, performing initial annotation, performing manual adjustment annotation, performing difference re-annotation and data storage. The power grid image intelligent annotation crowdsourcing platform comprises a to-be-annotated image collection module, an initial annotation module, a manual adjustment annotation module, a difference re-annotation module and a data storage module, and is used for executing the power grid image intelligent annotation crowdsourcing platform working method. According to the invention, the preset model is used to carry out initial annotation on the data; meanwhile, platform crowdsourcing is used for manually adjusting annotation; multi-person cooperation is achieved, the annotation efficiency is improved, the unqualified annotation result is modified according to IOU parameters, the annotation quality is effectively improved, meanwhile, the data classification and arrangement functionmeets the requirement for specific model training for a certain hidden danger, and powerful data support is provided for model training and model precision improvement in the future.

Description

technical field [0001] The invention relates to a power grid image intelligent tagging crowdsourcing platform and a working method, belonging to the technical field of power grid image data processing. Background technique [0002] With the rapid development of big data and artificial intelligence, the application of artificial intelligence in various fields of the power grid has taken root, such as: detection of hidden dangers in transmission channels, detection of wear of substation construction personnel, temperature detection of distribution tower heads, etc. The realization of the above functions is far from Without high-precision recognition model support, a good model not only requires excellent algorithm support, but also requires a large amount of labeled data for training. [0003] At present, manual labeling is still the mainstream of data labeling, and its labeling process is roughly as follows: through issuing labeling tasks, manual labeling, and finally submitt...

AI Technical Summary

Problems solved by technology

The entire labeling process takes a long time, and the labeling effect is often different due to different understandings of each person. Mislabeling and missing labels also occur from time to time. It is difficult to guarantee the quality of labeling, and manual inspection is time-consuming and laborious.

At the same time, the problem of calling a large number of personnel in a short period of time in the process of data labeling is also an urgent problem to be solved

The above problems not only affect the progress of project development, but also bury hidden dangers for model training

[0004] Chinese patent 201910218449.8 is a transmission line defect identification method, which is used to solve the problems of slow detection time and low detection efficiency in traditional manual detection. Whether there is an image of the target to be detected; the method of determining whether there is a target to be detected is generally to identify various targets in the screen through machine learning algorithms, or mark the target area to be recognized through expert experience to form an image of the target to be detected

Realize the identification of transmission line defects, but this method obtains the output results of the preset network model and directly marks them, the quality of the output results cannot be judged, and how the preset network model is trained is not disclosed

Images