Remote Sensing Fundamentals: Concepts & Image Interpretation
Passive vs. Active Remote Sensing
Remote sensing is the science of obtaining information about objects or areas from a distance, typically from aircraft or satellites. It is broadly classified into two types: passive and active remote sensing, based on how the sensors acquire data.
Passive Remote Sensing
Passive remote sensing relies on natural energy, usually sunlight, to detect and capture information. Sensors in this system measure the reflected sunlight or emitted thermal radiation from the Earth’s surface. These sensors cannot operate effectively at night or during cloudy conditions, as they depend on sunlight or natural thermal emissions.
Common examples of passive remote sensing instruments include:
- Photographic cameras
- Radiometers
- Satellite-based optical sensors (e.g., Landsat, MODIS)
This method is widely used in monitoring vegetation, land use, oceans, and atmospheric conditions.
Active Remote Sensing
In contrast, active remote sensing systems emit their own energy toward the target and then detect the amount of energy reflected back. These sensors are not dependent on sunlight, allowing them to function both day and night and even through clouds, rain, or smoke. They are especially useful for obtaining data in all weather conditions and for applications requiring precise elevation or structural details.
Examples include:
- RADAR (Radio Detection and Ranging)
- LiDAR (Light Detection and Ranging)
- SONAR (Sound Navigation and Ranging)
Key Differences and Applications
In summary, the key difference lies in the energy source: passive sensors use external natural energy, while active sensors generate and use their own energy. Passive systems are typically used for large-scale environmental and land-use studies, while active systems are crucial for topographical mapping, military surveillance, and disaster assessment. Understanding both types is essential for choosing the appropriate method based on environmental conditions and specific information requirements.
Causes of Distortion in Aerial Photography
Distortion in aerial photographs refers to the deviation of image features from their true position on the ground. Several factors contribute to this distortion, making it challenging to interpret aerial images accurately without proper correction. The main causes of distortion in aerial photographs are:
Camera Tilt
When the camera is not perfectly vertical during image capture, it causes a tilt distortion. This results in objects being displaced radially outward or inward from the center of the photograph. The greater the tilt, the more severe the distortion, especially at the edges of the image.
Relief Displacement
This occurs due to variations in terrain elevation. Objects at different heights are displaced outward or inward from the center of the image. For example, a tall building may appear to lean away from the center. This effect is more pronounced in areas with uneven terrain or tall structures.
Lens Distortion
The curvature of the camera lens can cause radial and tangential distortions. Radial distortion makes straight lines appear curved, especially near the edges of the photo. Modern aerial cameras are calibrated to minimize this, but some distortion may still exist.
Earth Curvature and Atmospheric Refraction
At high altitudes, the curvature of the Earth and the bending of light rays as they pass through the atmosphere can lead to minor distortions. Although relatively small, they become significant in large-scale or high-altitude aerial surveys.
Film or Sensor Deformation
In traditional film-based photography, the film may shrink or expand, altering the image dimensions. In digital systems, sensor imperfections or processing errors can also introduce distortion.
In conclusion, aerial photograph distortion is caused by geometric and physical factors such as camera angle, terrain variation, lens imperfections, and atmospheric effects. To ensure accurate interpretation, these distortions are often corrected through photogrammetric techniques, creating orthophotos that closely represent the real-world surface.
Identifying Relief Displacement in Aerial Photos
Relief displacement is a common geometric distortion in aerial photographs caused by variations in terrain elevation. It results in objects being displaced outward or inward from the center of the photograph, depending on their height and position relative to the photo’s center.
To identify relief displacement in an aerial photograph, consider the following key indicators:
Radial Displacement Pattern
Relief displacement causes objects, especially tall ones like buildings, towers, or hills, to appear tilted outward from the center (nadir point) of the photo. If you draw lines from the tops of these objects toward the center of the image, they will typically point directly to it. This radial pattern is a strong indicator of relief displacement.
Displacement Proportional to Height
The higher an object is, the greater its displacement. For example, in an area with both low-rise and high-rise buildings, the taller buildings will appear to lean away more noticeably from the photo center, while shorter structures may barely show any displacement.
Distance from the Center
Objects located farther from the center of the photograph show more noticeable relief displacement. A tall structure near the edge of the photo will appear more distorted than a similar structure near the center.
Comparison with Maps or Ground Truth
By comparing the aerial photo with a map or satellite image where true positions are known, discrepancies due to relief displacement become apparent. Objects may not align correctly unless corrected.
Shadow Direction vs. Displacement
If shadows fall in one direction while the tops of tall structures lean in a radial pattern from the center, it is likely relief displacement and not just an optical illusion or tilt.
In conclusion, relief displacement in an aerial photo can be identified by observing radial displacement patterns, especially in elevated features, with effects increasing with object height and distance from the center.
Vertical vs. Oblique Aerial Photography
Aerial photographs are images taken from an aircraft or drone looking down at the Earth’s surface. These photographs are classified into two main types based on the camera’s angle at the time of exposure: vertical and oblique aerial photographs.
Vertical Aerial Photographs
Vertical aerial photographs are taken with the camera pointed directly downward, perpendicular to the ground (nadir view). The resulting image covers a relatively uniform scale across the photo, especially near the center. These photos are ideal for mapping, land use planning, topographic surveys, and photogrammetry because they minimize distortion and provide a true overhead view.
Key characteristics of vertical photographs:
- The camera axis is nearly vertical (within 3° of the vertical).
- They provide minimal distortion near the center.
- Features retain their true shape and size (ideal for scale measurement).
- Useful for creating accurate maps and orthophotos.
Oblique Aerial Photographs
Oblique aerial photographs, on the other hand, are taken with the camera tilted at an angle, so the photo captures the side view of objects. These can be further divided into:
- Low Oblique: The horizon is not visible.
- High Oblique: The horizon is visible in the image.
Key characteristics of oblique photographs:
- The camera is angled away from the vertical position.
- Provide a more natural, three-dimensional view of the landscape.
- Buildings and structures are more easily recognizable due to visible sides and depth.
- Commonly used in visual interpretation, reconnaissance, tourism, and urban planning.
Conclusion
In summary, vertical photographs are best suited for accurate measurements and mapping, while oblique photographs offer a more realistic perspective and are helpful for visual interpretation. The choice between the two depends on the specific application and the level of detail or accuracy required.
Supervised and Unsupervised Image Classification
In remote sensing and image analysis, classification is the process of categorizing pixels in an image into different land cover or land use classes. There are two primary approaches: supervised and unsupervised classification. The key difference lies in the use of prior knowledge during the classification process.
Supervised Classification
Supervised classification requires user-defined training data. The user selects representative sample areas, called training sites, for each land cover class (e.g., water, forest, urban). The software then uses these samples to classify the entire image based on the spectral signatures of the training areas.
Key Characteristics of Supervised Classification
- Requires prior knowledge of the area.
- Uses labeled data (training samples).
- Common algorithms: Maximum Likelihood, Support Vector Machines, Random Forest.
- Produces more accurate and controlled results when training data is reliable.
- Used when the analyst knows the land cover types and can identify them in the image.
Advantages of Supervised Classification
- High accuracy if training data is good.
- Allows for detailed and specific class definitions.
Disadvantages of Supervised Classification
- Time-consuming and requires expertise.
- Results depend heavily on the quality of training data.
Unsupervised Classification
Unsupervised classification does not require prior training data. Instead, the software automatically groups pixels into clusters based on spectral similarity using algorithms like K-means or ISODATA. The user then interprets and labels each group after classification.
Key Characteristics of Unsupervised Classification
- No need for prior knowledge or training sites.
- Based purely on statistical patterns in the data.
- Often used for exploratory analysis.
Advantages of Unsupervised Classification
- Faster and requires less user input.
- Useful for unfamiliar areas or when training data is unavailable.
Disadvantages of Unsupervised Classification
- Less accurate and harder to interpret.
- Classes may not correspond directly to meaningful land cover types.
Summary of Classification Approaches
In summary, supervised classification is guided by human knowledge and is more accurate, while unsupervised classification is automated but less precise and requires interpretation after clustering.
Visual Interpretation of Land Cover Types
Visual interpretation is the process of analyzing satellite images or aerial photographs using the human eye and expert knowledge to identify and classify different land cover types. This method relies on various image interpretation elements such as tone, texture, shape, size, pattern, shadow, and association. Using these cues, we can effectively distinguish between forest, fallow land, settlement, and farmland.
Forest
- Tone: Forest areas usually appear in dark green or deep tones due to dense vegetation.
- Texture: Typically rough or coarse, especially in natural or mixed forests.
- Pattern: Irregular patterns; often continuous and dense.
- Association: Found in hilly or undisturbed regions.
- Shadow: Tall trees may cast shadows, especially in oblique images.
Fallow Land
- Tone: Appears light brown, grayish, or pale yellow, depending on the dryness and vegetation residue.
- Texture: Smooth to moderately rough; lacks crop rows or regular patterns.
- Pattern: Irregular and patchy, often near cultivated fields.
- Association: Found close to farmland; may be a part of crop rotation.
Settlement
- Tone: Varies from white, gray, to blue depending on roofing materials.
- Texture: Very fine or granular due to buildings and infrastructure.
- Shape: Geometric or blocky structures; often rectangular.
- Pattern: Regular grid-like patterns in planned areas; irregular in rural or unplanned settlements.
- Shadow: Tall buildings may cast long shadows in high-resolution or oblique imagery.
Farmland
- Tone: Ranges from green to brown, depending on the crop type and growth stage.
- Texture: Fine and regular; crop rows may be visible.
- Pattern: Often follows a regular, rectilinear layout.
- Association: Found in flat areas with irrigation structures or near water sources.
By carefully analyzing these visual elements, an interpreter can accurately differentiate between land cover types in satellite images or aerial photographs.
Identifying Floodplains vs. Hill Slopes in Imagery
Distinguishing between a floodplain and a hill slope in remote sensing imagery relies on interpreting key visual elements such as tone, texture, pattern, shape, elevation, and association. These features provide clues about the topography and landform characteristics of the area.
Floodplain
A floodplain is a flat, low-lying area adjacent to a river or stream that periodically floods.
Key Image Features of Floodplains
- Tone: Generally uniform and light to medium green during the growing season due to crops or grassland. It may appear darker when wet or waterlogged.
- Texture: Smooth or fine due to uniform vegetation cover and flat terrain.
- Pattern: May show a meandering river pattern, oxbow lakes, or sediment deposition features.
- Shape: Broad, flat expanses along rivers.
- Association: Closely associated with rivers, streams, and agricultural land. Settlements may be scattered or on raised embankments.
- Shadow: Minimal shadows due to flat topography.
Hill Slope
A hill slope is an inclined surface that rises from the surrounding terrain.
Key Image Features of Hill Slopes
- Tone: Variable, depending on slope aspect and vegetation; can range from light brown (bare slope) to dark green (forested).
- Texture: Coarse or rough, especially in forested or rocky areas.
- Pattern: Irregular, with contour-like banding or visible erosion features like gullies.
- Shape: Curved or elongated forms with visible elevation change.
- Association: Often found near mountainous or upland regions; vegetation may be dense or sparse based on steepness.
- Shadow: Prominent shadows on one side, especially in oblique or low-sun-angle images, indicating steep slopes.
Conclusion
Floodplains are identified by their flatness, smooth texture, and proximity to water bodies, while hill slopes are characterized by elevation change, coarse texture, and distinct shadows. Recognizing these features helps accurately interpret terrain in aerial or satellite imagery.