Summary of Valuable Data Extraction
Picturing logging, it has become a popular means of useful logging since it can visually represent the lithological and structural characteristics of strata. The manual interpretation of imaging logging is hindered by the limitations of the naked eye and experiential factors. Consequently, manual interpretation accuracy is low. Therefore, developing useful automatic imaging logging interpretation by machine learning is very useful. The paper proposes an automatic extraction procedure for the geological features in resistivity imaging logging images since it is a dominant use of imaging logging. This technique is grounded on machine learning and achieves brilliant results in a real-world application. Admitting the existence of valueless data significantly touches the recognition effect, the three strategies are recommended for the identification of worthless facts based on binary classification. Merging method is proven to be the best amongst the three policies both on an experimental dataset and in a production environment. It’s efficient in identifying of the valueless data in the excellent logging images, which has significantly improve the automatic recognition effect of geological features in resistivity logging images.