Final Quiz: Lesson ‘Hands-On training: Data and software’ Copy

From which sensors you currently can get data? (multiple-choice)

PRISMA
AVIRIS NG
EnMAP
Hyperion
SBG

The EnMAP campaign portal contains datasets from target sites in which of the following countries? Please guess or check the portal. (multiple-choice)

Germany
Spain
USA
Namibia

Which preprocessing steps should be applied to hyperspectral data? (multiple-choice)

Atmospheric correction
Radiometric correction
Cosmetic correction
Geometric correction

Fill in the blank (Options: gain, offset)

Every sensor has its specific gains and offsets applied to the recorded signals. The is determined by measuring the closed shutter and represents the thermally induced dark current. The is determined by measuring the sensor’s response to a known radiance target in the laboratory.

What is a radiometric correction and what does it do? (multiple-choice)

It transforms digital numbers to radiance
It transforms digital numbers to reflectance
It involves data resampling using nearest neighbor, bilinear interpolation or cubic convolution methods
It involves a linear transformation in which correction coefficients (gain and offset) are applied to every image pixel

Fill in the blanks, choose from: spectral non-uniformity, spatial non-uniformity, radiometric non-uniformity, temporal non-uniformity, radiometric correction, geometric correction, atmospheric correction

The smile effect is a that needs to be considered during .

What is considered in atmospheric correction? (multiple-choice)

sun position
atmospheric state
keystone compensation
terrain

Why can it be useful to apply dimensionality reduction before or as part of actual image analysis? (multiple-choice)

Because it can speed-up processing and reduce computational costs
Because hyperspectral data are often highly redundant and carry highly inter-correlated information
Because redundant data introduce noise and may lead to suboptimal model performances
Because the hundreds of contiguous bands are so confusing

In contrast to classification methods, quantification methods allow … (multiple-choice)

a direct identification of materials
the determination of material abundances in terms of percent fractional cover
the estimation of material contents and properties
assignment of unique classes to each pixel

Let’s assume you want to determine the relative abundance of known materials within each image pixel based on the materials’ spectral characteristics. Which method would you choose? (single-choice)

Spectral Mixture Analysis (SMA) or Linear Spectral Unmixing
Spectral Angle Mapper (SAM)
Support Vector Machine (SVM)
Spectral Feature Fitting (SFF)

Beyond the Visible – Einführung in die Hyperspektrale Fernerkundung

Final Quiz: Lesson ‘Hands-On training: Data and software’ Copy

Henryk Hodam

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