Raster Image Digitization With ML Denoising and Object Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Legacy raster images of seismic and well-log data lack digital data representation, requiring manual review and are time-consuming to process, hindering efficient data acquisition and analysis.
Innovation Solution
A method utilizing machine learning models for denoising and object segmentation to extract and digitize target objects from raster images, including training models to identify and discretize curves, and storing the results in a data file for visualization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If manual review methods are used to acquire data from legacy raster images, then data acquisition can be performed, but the process is time-consuming and requires significant human effort
Solution Approach 1:
The system enables self-service automation where machine learning models automatically perform data extraction from raster images without requiring human intervention. The denoising model and segmentation model work together to autonomously identify and extract curve data, eliminating the need for manual review while maintaining high accuracy in data acquisition.
2Measurement precision
If traditional image processing methods are used, then simple noise reduction is possible, but accurate identification and extraction of target objects remains difficult
Solution Approach 1:
The system segments the image processing task into distinct functional components: a denoising model that removes noise from the raster image, and a separate segmentation model that identifies target objects in the cleaned image. This segmentation of functionality allows each model to specialize in one aspect of the problem, achieving high identification accuracy while maintaining manageable system complexity through modular architecture.
Data Source
AI summary
A method for digitizing image-based data includes receiving an image file including one or more target objects, generating an intermediate image by removing noise from the image file using a denoising machine learning model, identifying the one or more target objects included in the intermediate image using an object segmentation machine learning model, discretizing the one or more target objects that were identified using the trained object segmentation machine learning model, and storing the one or more target objects that were discretized in a data file, visualizing the one or more target objects, or both.