Rune' Inversion (April 2019) - Artificial Intelligence

The very first Seismic Post-Stack Inversion algorithm that allows estimating Velocity and Density separately; with no well logs data required!

Artificial kinematic constrains, simulated annealing and global optimizations are used as the basis for the algorithm.

-  There is no coupling between Vp and Dn beyond frequency of the initial models.

-   Only Seismic data and Seismic Velocity are used.

-   The quality of result depends on the quality of the input data.

-    It does not use Zoepprits, Aki-Richards and Fatti equations or theirs improved approximations and modifications.


Rune’ Inversion appeared to be a good link between geology and geophysics without a human interaction.

What are the advantages compare to the conventional algorithms?

  • Post-Stack Inversion of deterministic approach, allows to predict only Impedance (P-imp=Vp*Dn).

  • Pre-Stack Inversion of deterministic approach, allows to predict Impedances (P-imp=Vp*Dn, S-imp=Vs*Dn) and Density that is normally very poor resolved from AVO inversion.

  • Stochastic Inversion gives better visual resolution. Nevertheless, the results of mathematical variations in the long run are the average of the deterministic approach.

All known algorithms required skilled specialists and software. The result is very dependent on the hand of geophysics.


The proposed algorithm can be used by a geologist that have Seismic Stack and Stacking Velocities. He/she can invert a whole seismic section or a cube to P-impedance, Vp and Density.

Contact for more information here

Example 1

Only public data is used: Seismic post-stack data and seismic velocity.

Well 35/3-2. QC: well logs and inverted result by deterministic approach

Agat field
Example 2

Only public data is used: Seismic post-stack data and seismic velocity calibrated to check shots.

Vp, Inversion, post-stack, AI, Seismic velocity.