No abstract

Meteorological and hydrological chart records information, measurement data of rainfall, water level, humidity, temperature and other types of measured parameters. These parameters are collected from hydrometeorological measurement stations nationwide. The storage of this information is extremely important for the purpose of researching and forecasting weather and natural disasters in the future. However, at present, the storage of all types of schemas is in paper form, the reading of data depends on the expert. Therefore, it is difficult to guarantee the integrity of the data over time. In this paper, we propose a solution for schema recognition and self-recording of schema information using today's most advanced machine vision and artificial intelligence technologies to help store and digitize data, diagrams automatically. The solution integrates the page structure analysis algorithm, the grid detection algorithm and the alignment algorithm to combine the line detection algorithm and the objects in the schema to separate the line. By experiment, the solution has achieved high accuracy, more than 90% of the diagrams can be digitized, including all types of diagrams of precipitation, water level, humidity, pressure, and temperature.

No abstract