![]() ![]() Here, models using different epochs show variable results and the ≥100 epochs models systematically underestimates the measured rock porosities. More problematic is the reproducibility of porosities. Preliminary results show that the derived sand volumes classification reproduce the point counting results well (80 % accuracy of predicting classes or neighbouring classes). For both cases, we observe that initially high model loss for the validation data reaches low values after 50 epochs.To further test the approach, we analyse in a second stage a holdout dataset of sandstones from the Norwegian Continental Shelf. We perform the training phase with a varying number of epochs ranging between 20 and 200. A final soft max layer is added so that the recovered output can be interpreted as probability distributions. The batches are normalized after each pooling layer and a dropout layer used to reduce overfitting before flattening. For both classifications, we trained a convolutional neural network consisting of 5 convolutional layers and 4 max pool layers. In the processing stage, we normalize and scale all the images to a reference number of 128 pixels. We split the dataset into a training (85 %) and validation dataset (15 %). The images are grouped into 8 different sand volume and 8 different porosity classes. 3500 thin section images from different sandstone types with known properties. However, one step further is combining image analysis with machine learning.In this work, we evaluate the use of a neural network learning algorithm to classify selected sandstone properties from thin section images. An example using image analysis has been discussed in Roduit, 2007. Attempts to automate and digitize this process are therefore promising. This time-consuming, repetitive, and subjective work is usually done by an experienced petrographer. The classic way to gain knowledge about these parameters is point counting on thin sections. It contains tools to quantify either manually or automatically.Features:Read images in TIFF, BMP, FlashPiX, GIF, JPEG, PNG, and PNM formats Efficient visualization system Quantify components: objects or background Object analysis (size, shape, orientation, texture …) Object classification Image processing (binary and morphology operations, filtering, segmentation…) Image rectification (geometric corrections by control points) Digital point counting Tools for data collection in one or two dimensions Image annotation and description card Profile (variation of granulometry, density, objects or background) Save all measures, data, calibration and preferences in a single project fileWhat’s New in version 1.2.Sand volume and porosity measurements on sandstones are routine work in geoscientific applications, providing useful input to flow simulation in porous media-based analyses (e.g., in CO2 storage and/or hydrocarbon migration studies). The program contains most of the common image processing operations, has a and intuitive user interface, an efficient visualization system and innovative features. JMicroVision is an image analysis toolbox for measuring and quantifying components of high-definition images. ![]()
0 Comments
Leave a Reply. |