Nevertheless, the highly sophisticated modern, automated 
image analyzing systems cannot yet replace the human brain. 
Thus, the lack of these systems should not deter any geoscientist 
from using ERTS imagery as an aid to his studies. 
The process of interpretation could be carried out at 
various levels of expertise, such as rapid recognition, skilled 
interpretation and automated analysis, as Gregory (1973) 
describes. 
1. Rapid recognition of obvious or self-evident information may be 
carried out by anyone who has appreciation of scale , maps and 
regional geography. Only very generalized calibration or "ground 
truth" is required. The type of information that is readily 
derived by such interpretation of ERTS images relates 3 for example 3 
to rivers and lakes , outcrop and soil patterns 3 extent of urban 
development , forest and field patterns 3 smoke plumes 3 turbidity in 
water and snow and ice cover. Some of this information can be 
mapped by automated methods 3 such as density slicing and colour 
enhancement. Minimal formal training is required to undertake 
such interpretation 3 excluding the automated aspects. 
2. Skilled interpretation by a discipline-oriented specialist or team 
of specialists. The interpretation is still relatively simple 3 often 
comprising only visual analysis 3 but it may include machine enhance 
ment of mayor contrasts in spectral 3 spatial or temporal aspects of 
the data. Skilled interpretation depends upon local expertise and 
specialized knowledge about resources and the environment. It 
need not require geometric correction nor special equipment nor 
training other than that normally received by specialists in their 
disciplines and in the methodologies of air photo interpretation. 
Calibration data from selected areas are essential to validate the 
interpretation 3 aerial surveys may help in extrapolating beyond 
the calibration sites and by providing greater detail. At this 
level of expertise 3 interpretation of ERTS data provides such 
information as geological structure 3 extent of erosion and sedimen- 
tation 3 distribution and character of surficial deposits 3 and 
bedrock recognition. 
3. Automated analysis 3 or spectral 3 spatial and temporal discrimination. 
This level of interpretation is machine-assisted 3 complete and 
relatively expensive. In many cases 3 computers 3 digital and optical 
processing 3 extensive field data and concurrent aerial surveys 3 may 
be required. The products are tables 3 print-outs 3 classifications 
and other data formats that must be validated by human knowledge 
and experience in order to provide useful information. 
It must be kept in mind that ERTS imagery is basically an 
extension of aerial photography in a more sophisticated form, 
to view and study large areas. Some types of information