Customer success

Mining 4 mining

Mining industry can benefit greatly from a new generation of data analytics, dashboarding and AI visualization. But discussing with the professionals, it appears that a main bottleneck remains to be solved: The access to the legacy data stored in documents to enable new exploration models and advanced analytics.

Capture d’écran 2021-10-20 à 12.41.44.png

Reading a composite log as a geologist!

Whatever the subsurface industry, geologists used to synthesize in complex and very graphical reports everything they have seen, measured and interpreted while drilling 50’ hole for water or 21,000’ well for oil. These composite logs are rich, frequently very esthetic, but not saved in a format which makes the access to data easy.

Capture d’écran 2021-10-20 à 12.42.05.png

Searching for seismic acquisition designs

Marine seismic acquisition is a high tech activity which has evolved a lot over the past decades and has radically changed the quality of the subsurface imaging.

When, where, and how the changes have been made exploring the East Atlantic margin? Which places should be re-shooted because of the potential of the new acquisition technics?

Capture d’écran 2021-10-20 à 12.42.27.png

Classifying large amount of documents

Machine learning can scale-up easily, but what about classifying 500,000 vintage documents related to exploration, appraisal or development wells of the UK continental shelf?

That was the very first industrial challenge we faced at AgileDD and plenty of lessons were learned from it.

Capture d’écran 2021-10-20 à 12.42.49.png

When the layout is complex!

Capturing information in complex documents with diverse layout like the ones used for the well logs may be quite challenging. It involves mastering several techniques like the OCR or printed and handwritten text occurring in several orientations. But, even when OCR is done, the technical documents cannot be simply considered as a sequence of words.

Capture d’écran 2021-10-20 à 12.43.28.png

Training iQC on forms

In a lot of organizations and particularly in the federal, state or city administrations, most data processes start with the capture of information within defined forms.

Because the iQC data capture is based at on the detection of information and the recognition of the context where it occurs, it is easy to train our AI to capture automatically even handwritten data in any types of forms.