GISCI 241 Chap.3 Aerial Photos and Image Interpretation
Aerial Photos and Image Interpretation
Aerial photographs offer fine spatial detail and rich visual context, but they become map evidence only after their geometry and reference are understood. This chapter compares vertical and oblique perspectives, explains georeferencing with stable and distributed control points, and shows how digitised features inherit source uncertainty.
It also develops systematic image interpretation through tone, colour, shape, shadow, texture, pattern, size, site and association. Strong interpretations combine several cues and state plausible alternatives.
What this chapter covers
- 01
Vertical and oblique aerial perspectives
- 02
Image coordinates and real-world map coordinates
- 03
Stable, visible and distributed ground control points
- 04
Residuals, validation and orthorectification boundaries
- 05
Digitising points, lines and polygons from imagery
- 06
Converging visual evidence and alternative explanations
Worked example · free
Reject an unstable control point
- +1Confirm that the pier corners are the same unchanged features in the historic image and reference map.
- +1Retain broad spatial distribution so the transformation is constrained across the useful image.
- +1Reject the vegetation edge if field or image evidence confirms it is unstable rather than merely inconvenient.
- +1Refit and validate against independent stable features, reporting residual pattern and remaining uncertainty.
Key terms
- Georeferencing
- Assigning real-world coordinates to positions in an image or scanned map.
- Control point
- A recognisable stable feature paired between image coordinates and a trusted geographic reference.
- Residual
- The remaining positional disagreement between a transformed control point and its reference location.
- Orthorectification
- Geometric correction that accounts for sensor geometry and terrain to improve planimetric consistency.
- Digitising
- Creating digital point, line or polygon features from imagery, maps or field observations.
- Image texture
- The local arrangement and variability of tone or colour, described as smooth, intermediate or rough.
- Association
- Interpretive evidence based on features or activities that are commonly found near one another.
Aerial Photos and Image Interpretation FAQ
What makes a good control point?
A good point is stable through time, sharply recognisable in both image and reference, accurately located and distributed with other points around the area of interest. Avoid shadows, moving shorelines, vegetation edges and ambiguous corners. Reserve independent features or surveyed points to test the fitted transformation.
Can I measure directly from an oblique photograph?
Not safely with one uniform scale because foreground and background scale differ. Oblique images provide excellent context, depth and vertical form, but direct plan measurement requires rectification or orthorectification with suitable geometry and control. Otherwise keep the claim qualitative and explain the perspective limitation.
How should I identify an unknown feature?
Combine several cues rather than trusting one. Consider tone or colour, shape, shadow, size, texture, pattern, site and association. State the leading interpretation, name a plausible alternative and identify an additional observation that would separate them. This makes uncertainty visible and the reasoning testable.
How should I report georeferencing quality?
Report the control features, their distribution, transformation and residual pattern, then evaluate independent checkpoints away from the fitted points. A single average residual can hide edge drift or directional bias. Also state whether relief displacement or camera geometry remains, because georeferencing alone does not make an unrectified photograph uniformly measurable.
Exam move
Practise georeferencing as a decision sequence: identify stable features, spread control around the scene, fit a suitable transformation, inspect point residuals, remove only points with a documented geographic reason, and validate elsewhere. For interpretation, take one aerial image and write a cue table for tone, colour, shape, shadow, size, texture, pattern, site and association.
Build conclusions from three or more converging cues and always include an alternative explanation. When comparing dates, list camera geometry, relief, shadow and registration differences before calling a feature movement real. Keep visibility and measurability separate; an object can be obvious in an oblique view while unsafe for plan measurement.
Draw two control layouts, one clustered in a corner and one distributed around the frame, and predict where residual error is likely to grow. Add a separate set of checkpoints so you can explain why fitting and validation need different points. Compare a simple and a flexible transformation: the flexible model may lower control residuals but can distort regions with weak support.
For vertical and oblique photographs, make a table of what each view reveals and what it permits you to measure. Practise refusing a uniform scale on a high-oblique image unless rectification or camera geometry supports it. For interpretation, select an ambiguous feature and write two plausible identities.
List the cues that favour each, then choose one extra observation, such as another band, date, elevation layer or ground visit, that would separate them. When digitising, define the boundary rule before tracing and keep display scale consistent.
End every date-comparison exercise with a false-movement audit covering camera position, relief displacement, shadows, tides or season, transformation quality and independent landmarks. This keeps visual change from being confused with geometric change.