Automated Kerogen Classification in Microscope Images of Dispersed Kerogen Preparation
- PDF / 541,143 Bytes
- 14 Pages / 439.37 x 666.142 pts Page_size
- 41 Downloads / 171 Views
Automated Kerogen Classification in Microscope Images of Dispersed Kerogen Preparation L.I. Kuncheva · J.J. Charles · N. Miles · A. Collins · B. Wells · I.S. Lim
Received: 2 September 2007 / Accepted: 24 March 2008 / Published online: 30 April 2008 © International Association for Mathematical Geology 2008
Abstract We develop the classification part of a system that analyses transmitted light microscope images of dispersed kerogen preparation. The system automatically extracts kerogen pieces from the image and labels each piece as either inertinite or vitrinite. The image pre-processing analysis consists of background removal, identification of kerogen material, object segmentation, object extraction (individual images of pieces of kerogen) and feature calculation for each object. An expert palynologist was asked to label the objects into categories inertinite and vitrinite, which provided the ground truth for the classification experiment. Ten state-of-the-art classifiers and classifier ensembles were compared: Naïve Bayes, decision tree, nearest neighbour, the logistic classifier, multilayered perceptron (MLP), support vector machines (SVM), AdaBoost, Bagging, LogitBoost and Random Forest. The logistic classifier was singled out as the most accurate classifier, with an accuracy greater than 90. Using a 10 times 10-fold cross-validation provided within the Weka software, we found that the logistic classifier was significantly better than five classifiers (p < 0.05) and indistinguishable from the other four classifiers. The initial set of 32 features was subsequently reduced to 6 features without compromising the classification accuracy. A further evaluation of the system alerted us to the possible sensitivity of the classification to the ground truth that might vary from one human expert to another. The analysis also revealed that the logistic classifier made most of the correct classifications with a high certainty. L.I. Kuncheva () · J.J. Charles · I.S. Lim School of Computer Science, Bangor University, Bangor LL57 1UT, UK e-mail: [email protected] N. Miles PetroStrat Limited, Llandudno, UK A. Collins · B. Wells Conwy Valley Systems Ltd, Deganwy, UK
640
Math Geosci (2008) 40: 639–652
Keywords Kerogen recognition · Machine learning · Image processing · Transmitted light microscopy
1 Introduction Kerogen is organic matter which can yield hydrocarbons on heating. Anglicizing the kerogen content of sedimentary rocks, using images of material seen under the microscope, is therefore a vital component of oil and gas exploration (Tyson 1990). The kerogen pieces found in images of palynofacies can be broadly classified as inertinite, vitrinite or other. Vitrinite and inertinite macerals are derived from mainly higher plant material and include cell walls and gelified tissues. Figure 1 shows a typical image of a slide prepared for transmitted light examination containing vitrinite, inertinite, microfossils and other debris. Distinguishing between inertinite and vitrinite, based on their appearance in the imag
Data Loading...
