Spectral mineralogy of a porphyry-Cu pebble
This is a great exercise to learn more about drill-core scanning with a Shortwave Infrared (SWIR) hyperspectral imager. The exercise shows how lab-based, high-resolution hyperspectral imaging provides very useful and subtle mineralogical information about mineralized rocks. Exploratory analysis of a hyperspectral image of a rock acquired at 0.2 mm pixel size shows i) subtle compositional variation in illite-muscovite minerals that are related to copper- mineralization, and ii) occurrence of other relevant minerals. Methods in this exercise include wavelength mapping, interpretation of pixel spectra and the use of shadow images.
The exercise was developed in 2012 as part of the Advanced Remote Sensing course in the Applied Earth Sciences MSc program at the faculty of ITC of the University of Twente. The version here was used at the Hyperspectral Remote Sensing workshop of the Colloquium of African Geology in Addis Abeba in 2013.
A document with example answers to questions in the instructions and example screenshots of some of the results are available, as well as a review on my YouTube channel.
• Instructions: Ex - Porphyry Cu Pebble-Dec2021.pdf and Ex - Porphyry Cu Pebble-Dec2021 – table 2.doc
• Data: Hyperspectral image: Pebble_60_HS; Folder with USGS library spectra
• Example answers to questions: ex - porphyry Cu pebble-Dec2021- ANSWERS
• Screenshots of selected results: Results.pdf
Hyperspectral classification of a porphyry-Cu pebble using a decision tree
In this exercise you will learn about mineral mapping of a rock sample with SWIR hyperspectral imagery. The exercise shows how lab-based, high-resolution hyperspectral imaging provides very useful and subtle mineralogical information about mineralized rocks. This tutorial follows up on the previous exercise “Spectral mineralogy of a porphyry-Cu pebble.” In that exercise we did an exploratory analysis of a hyperspectral image of the pebble, and we came up with a number of spectral endmembers. In this exercise “Hyperspectral classification of a porphyry-Cu pebble using a decision tree,” a mineral map will be created using the spectral characteristics of the spectral endmembers by classifying a wavelength image using a decision tree. The decision tree classification is relatively simple, applies expert knowledge of spectral minerals and uses the wavelength position of absorption features, which are representative of molecular bonds in mineral crystal lattices.
Currently, we have no access to user-friendly open-source software to build a decision tree. Instead, the commercial ENVI package will be used to create and run a decision tree. The exercise was developed as a follow-up on the “Spectral mineralogy of a porphyry-Cu pebble” in 2018 as part of the MSc program in Applied Remote Sensing at ITC.
A document with example answers to questions in the instructions and example screenshots of some of the results are available.
• Instructions: ex - hyperspectral classification of porphyry Cu pebble using a decision tree.pdf
• Data: Hyperspectral image: Pebble_60_HS and images derived from the reflectance data, such as the wavelength image and map; Folder with image endmember spectra
• Example answer to question: ANSWER to question in exercise.pdf
• Selected results: Decision tree classification - example result.pdf