Predictive Geological Drawing With Machine-Learned Sketch Harmonization

The patent introduces a machine learning model as an intermediary between the geologist's sketch and the final digital representation. The ML model receives the sketch as input and automatically generates standardized geological features, acting as a mediator that translates varied human drawings into consistent digital formats without requiring the geologist to manually adjust each sketch.

The system creates a digital copy of the geologist's sketch and processes this copy through the machine learning model. The original sketch remains unchanged (preserving intuitiveness), while the digital copy is transformed into a standardized representation, allowing harmonization without affecting the ease of the original drawing process.

The machine learning model is pre-trained on extensive geological data and sketch examples, so it already contains the knowledge of what geological features should look like. When a geologist makes a sketch, the model immediately applies this pre-acquired knowledge to predict and standardize the features, eliminating the need for the geologist to manually select characteristics during the drawing process.

The system enables the sketch itself to 'select' its characteristics automatically. Instead of the geologist having to manually specify depth intervals, grain sizes, and other parameters, the machine learning model analyzes the sketch and automatically determines these characteristics, allowing the system to serve itself rather than requiring human intervention for each parameter.