It is believed that getting into the oil industry, where salaries are not inferior to the IT industry, is quite difficult from the outside. There is some truth in this, because the sphere is quite specific and does not favor people “off topic”. This event is aimed at simplifying entry into this underground world for young and talented teams working with image recognition and machine learning.
I place this topic in the “I PR” section because: a) I want to help the Ufa fellow countrymen; b) I believe in the high qualification of hawkers. And it will be great if some meet others. At the same time I’ll be a little bit translator from technological to human.
So what is the challenge?
The task is: “Isolation of seismic horizons in the cube of amplitudes - data segmentation using image recognition”. The championship is hosted on the Boosters.pro platform. The organizer is the corporate institute of LLC BashNIPIneft, one of the leaders (oddly enough) in the field of oil and gas software development. An indicative example of their successful work is the development and implementation of RN-GRID - its own industrial software for mathematical modeling and analysis of the process of creating cracks during hydraulic fracturing.
I translate the task into Russian
Despite the terrible name, the task boils down to the analysis of images by machine learning. But, as usual, there are many nuances.
Seismic exploration is the main method for detecting oil and gas. The method is based on the excitation of elastic vibrations and the subsequent recording of the response from rocks. These vibrations propagate through the thickness of the earth, refracting and reflecting on the boundaries of geological layers with different properties. The reflected waves return to the surface and are recorded. The output is the so-called seismic cube, which is cut into layers vertically and horizontally. We get such sections (crosslines and interlines), on which rocks with different properties are visible.
The task of the participants is to accurately determine and mark these layers-horizons throughout the seismic cube based on preliminary training at 10% of the cube. It’s easy, right?
And now in common terms:
“Correlation in seismic exploration refers to the process of identifying and tracking reflecting horizons, various seismic facies complexes (reefs, etc.) in time, depth and space, on seismograms and on the total time and depth seismic data.
In the process of tracking reflecting horizons, a combination of kinematic and dynamic seismic attributes is used. When they are comprehensively analyzed, the reflecting boundaries of the wave field in space are correlated by tracking the most pronounced extrema (or transition through 0) of the wave field, while mainly taking into account the similarity sign of neighboring seismic traces.
At the same time, the property of smoothness of change in the time of registration of wave arrival is taken into account. A line connecting the characteristic features (extremes) of the same wave on different paths is commonly called the in-phase axis. Reflected waves are usually correlated with the most distinct extrema (phases). In this case, interpreters usually adhere to the principle - from more reliable to less reliable.
First, we will trace the horizons that are confidently traced over a significant area in the study area and have the proper geological reference. Such reflecting horizons are commonly referred to as reference or reference. They are regional markers. Their tracking and interpretation can significantly increase understanding of all seismic material, tectonic history, sedimentation conditions. "
Kirilov A.S., Zakrevsky K.E., Workshop on seismic interpretation in PETREL. M .: PUBLISHING HOUSE MAI-PRINT, 2014 .-- 288 p.
Need more information?
On this issue, there is a large amount of reference information in Russian in almost any of the formats. Including on Youtube. For example, you can cite an excellent visual video about the automatic recognition of seismic horizons laid out for free access by the Kazan Center for Further Education IGiNGT KFU.
It seems to me after him the task inherent in the challenge should become more understandable.
Ok, what needs to be done?
Based on the first 10% of the seismic cube already marked by the professional interpreter, you need to mark the remaining slices in the test dataset along the boundaries of the specified classes with the maximum metric value.
What to work with?
The source dataset is a three-dimensional array of seismic data (total time cube of the seismic attribute). As mentioned above, a cube can be represented as 2D vertical slices: cross-lines and inlines.
Each slice consists of one-dimensional vectors - traces (trace) 2562 milliseconds long with a step of 2 ms. Number of crosslines: 1896. Number of inlines: 2812.
Total number of runs> 5 million
Number of segmentation classes (i.e. breed divisions): 8.
Who is waiting for the Seismic Challenge?
The organizers are waiting for the participation of specialists from the field of data analysis. The deadlines are limited and the challenge is for tech who "already knows how". Both individuals and teams of up to five people can participate in the competitive selection.
How to take part?
Participants register themselves through the RN.DIGITAL website. at Boosters.pro. According to statistics, on November 4, 402 teams registered to participate in the competition.
Dates:
10/15/19 - 12/15/19 - holding a contest
11.24.19 - the end of the opportunity to combine teams
10/15/19 - 12/01/19 - the first round of the contest
12/02/19 - 12/15/19 - the second round of the contest for the top 30 teams from the first round
12/21/19 - full-time debriefing and rewarding of 10 teams from the second round in the city of Moscow.
The organization of the final is interesting: the expert council evaluates the final work, but does not affect the choice of winners. The distribution of the finalists is determined by the results of the correspondence part of the competition based on the best segmentation quality metric (Dice Metrics). At the same time, participants can get an extra “bun” for the best presentation of their decision in the amount of 50,000 rubles.
PS
I am not the organizer of this challenge, so I can hardly answer the questions in the comments in detail. If habragiteli have questions / interest, then I can invite a representative of the organizers and guys from boosters in the comments.