Skip to main content
Conference Proceedings

Orebody Modelling and Strategic Mine Planning SMP 2014

Conference Proceedings

Orebody Modelling and Strategic Mine Planning SMP 2014

PDF Add to cart

Comparison between Gaussian Processes and Dynamic Time Warping for Classification of Marker Shales in Iron Ore Deposits

The banded iron formation-hosted iron ore deposits located in the Hamersley Ranges of Western Australia contain marker shales that are used to distinguish lithology for geological and orebody mapping. These marker shales produce distinctive signatures in the natural gamma downhole logs, which are conventionally manually interpreted. This paper proposes an alternative method of automatically identifying these shales using dynamic time warping (DTW). A Gaussian processes (GPs)-based method for automatic identification is also considered. Both methods are tested on a deposit that contains the Marra Mamba sequence of the Hamersley Province.Both methods were trained to predict the location of the desired shales, and the resulting outputs were compared with a manual interpretation of the results to determine the accuracy. The GP method can identify signatures with an accuracy between 87.0 per cent and 92.0 per cent, which increases to 92.9 per cent and 95.7 per cent where the GP is certain. The DTW method has an accuracy between 90.0 per cent and 91.6 per cent, which increases to 94.0 per cent and 95.4 per cent where the DTW is certain.The GP method has several advantages, including a more automated library building process, a probabilistic output and a shorter run time. However, the DTW method was capable of accepting more distortion in the signature and therefore identified more signatures. This is important in a deposit where the signals are frequently distorted by events such as hydration and folding. Both methods provide a viable alternative to the current manual detection, but their differences make them applicable to different cases.CITATION:Silversides, K and Melkumyan, A, 2014. Comparison between Gaussian processes and dynamic time warping for classification of marker shales in iron ore deposits, in Proceedings Orebody Modelling and Strategic Mine Planning Symposium 2014 , pp 177-184 (The Australasian Institute of Mining and Metallurgy: Melbourne).
Return to parent product
  • Comparison between Gaussian Processes and Dynamic Time Warping for Classification of Marker Shales in Iron Ore Deposits
    PDF
    This product is exclusive to Digital library subscription
  • Comparison between Gaussian Processes and Dynamic Time Warping for Classification of Marker Shales in Iron Ore Deposits
    PDF
    Normal price $22.00
    Member price from $0.00
    Add to cart

    Fees above are GST inclusive

PD Hours
Approved activity
  • Published: 2014
  • PDF Size: 0.684 Mb.
  • Unique ID: P201413019

Our site uses cookies

We use these to improve your browser experience. By continuing to use the website you agree to the use of cookies.