5.5 Quiz: Hands-On training

In the EnMAP-Box, you can visualize the following data types (multiple-choice)

Raster data
Vector data
Spectral libraries
Exclusively EnMAP data but in all levels (L1B, L1C, L2A)

The EnMAP-Box is a great tool for hyperspectral data visualization and analyses … (multiple-choice)

… as you can easily select which bands to combine in a RGB view of an image
… as image and vector data can be displayed together in the same “map view” panel
… as several “map view” panels can be spatially linked in side by side representations
… as spectral pixel information can be displayed as spectral profiles in a separate “spectral library” window

In the exercises, we are working with scenes acquired by which sensors? (multiple choice)

HyMAP
AVIRIS
EnMAP
HySpex
AVIRIS-NG

In the EnSoMAP tutorial, we applied a mask to exclude all pixels that do not represent (mostly) bare soil. The masks were derived from a combination of indices, namely … (multiple choice)

Normalized Difference Vegetation Index – NDVI (to identify green vegetation)
Normalized Difference Vegetation Index – NDVI (to identify dry vegetation)
Normalized Cellulose Absorption Index – NCAI (to identify dry vegetation)
Normalized Difference Red Blue Index – NDRBI (to identify water)
Normalized Difference Soil Index – NDSI (to identify pure soil pixels)

In the regression tutorial, you selected – per default – a 10-fold cross validation. What does that mean? (multiple-choice)

It means that the model established on my training data set is validated 10 times against the test dataset
It means that my training dataset is split into 10 parts, as equally sized as possible, and that the model is run until each of these parts has been used for validation once while the others 9 parts were used for training. The output is then an average of all results
It means my result is 10-fold and that I need to devide it by ten to be able to compare it to other model results

What is meant by the term “semi-quantitative map”? (Single choice)

A map that you can trust to 50% to be valid
A map that can be generated with only half of the reference samples
A map that can be generated without any reference samples
A map that can be generated without applying a bare soil mask

Which model metrics are displayed in the accuracy report of the regression tutorial? (Single choice)

e.g. R², RPD, RMSE
e.g. R², RMSE, RPIQ
e.g. R², RMSE, variable importance value

What do you need to convert a semi-quantitative map into a fully quantitative map? (single choice)

Expert knowledge of the study site
Reference data from the study site
Laboratory spectra of the samples collected at the study site
Information from the local farmer on the type of field crop growing on the study site

What do spectral feature analyses include? (multiple choice)

Band ratios are applied using spectral bands specific for the desired soil component
Slopes between spectral bands specific for the desired soil component are used
PLSR/RF models are applied to specific spectral bands only
The values of the spectral feature analyses can be linked to reference data with linear models to receive fully-quantitative maps

What are sources of uncertainty in soil spectroscopy? (multiple choice)

Spatial inaccuracy of the reference data
Spatial inaccuracy of the hyperspectral image data
Model inaccuracy
Inaccuracy in wet-chemistry measurements of reference data

Beyond the Visible – Imaging Spectroscopy for Soil Applications

5.5 Quiz: Hands-On training

HYPERedu

Quiz Summary

Information

Results

Results

Categories

1. Question

2. Question

3. Question

4. Question

5. Question

6. Question

7. Question

8. Question

9. Question

10. Question

More