Intro
Multispectral scanning allows for the acquisition, display, and interpretation of the thermal properties of the Earth’s surface. Many multispectral systems sense radiation not only in the visible and reflected infrared but also in the thermal infrared range (3 μm – 15 μm).
Thermal remote sensing differs from optical imaging: it measures emitted energy rather than reflected sunlight. The boundary occurs near 3 μm, where shorter wavelengths correspond to reflected solar energy and longer wavelengths record emitted heat. The Earth approximates a blackbody, peaking near 9.7 μm, though every surface’s emission depends on emissivity, conductivity, heat capacity, diffusivity, and inertia.
This lab used ENVI Classic to analyze a thermal image of Delta State University, converting raw sensor data into radiance and temperature values. The objective was to derive surface temperature, visualize thermal gradients through pseudocolor mapping, and interpret what these patterns reveal about material and environmental interaction.
Creating a Color Map
Using ENVI’s Color Mapping tool, the grayscale thermal image was transformed into a pseudocolor display. Each pixel’s Digital Number (DN) was assigned a hue according to its radiance value.
The RED–BLUE color table was deliberately selected. According to NASA Earth Observatory (2021) and the U.S. Geological Survey (2019), the blue-to-red thermal ramp is optimal for representing temperature gradients because it matches human expectation:
- Blue → cooler or low-emission surfaces.
- Red → warmer or high-emission surfaces.
As Jensen (2016) explains, an effective color ramp must preserve order, enhance mid-range contrast, and evoke intuitive meaning. The RED–BLUE gradient achieves this by maintaining linear luminance progression, ensuring that equal temperature changes appear as equal visual steps.
Within the Delta State scene, blue and cyan areas corresponded to trees, grass, and shaded zones—cool through evapotranspiration—while yellow to deep red tones traced roofs, asphalt, and parking lots—materials of low albedo and high heat retention.
Thus, the RED–BLUE palette was chosen not for aesthetics but for clarity of interpretation. It transforms raw physics into perception: red where heat lingers, blue where the land still breathes.
Comparing Thermal and Satellite Views
Comparing the thermal image with Google Maps satellite imagery revealed distinct patterns.
- Red/pink regions aligned with buildings and pavement—surfaces that absorb and reradiate heat.
- Blue/green regions aligned with vegetation and water—features that remain cooler through reflection and moisture loss.
Even if air temperature is uniform, emission differs by material. The thermal band makes this invisible variability visible.
Determining Radiance Range
Digital Number (DN) values were converted to spectral radiance using R=0.0370588×DN+3.2R = 0.0370588 × DN + 3.2R=0.0370588×DN+3.2
yielding Rₘᵢₙ = 3.20 W m⁻² sr⁻¹ μm⁻¹ and Rₘₐₓ = 8.76, for a range of 5.56.
This conversion, consistent with NASA Landsat Handbook (2020) guidelines, expresses pixel intensity in physical energy units. The mean radiance (6.34) and standard deviation (0.49) indicate a moderate spread typical of mixed urban–vegetation scenes.
The dynamic range confirms low-gain acquisition, suitable for capturing both bright rooftops and shaded vegetation without saturation. In essence, the radiance range validates instrument calibration and scene exposure quality.
Pixel and Feature Analysis
Using ENVI’s Pixel Locator and ROI Tool, radiance values were sampled from representative surfaces.
| Surface Type | Thermal Behavior | Interpretation |
|---|---|---|
| Grasses | Low radiance | Moisture and chlorophyll maintain coolness. |
| Trees | Very low radiance | Dense canopy intercepts sunlight. |
| Buildings | Moderate–high | Composite materials reflect and store heat. |
| Roads | High | Asphalt absorbs solar energy efficiently. |
| Parking Lots | Very high | Expansive, exposed, low-albedo surfaces. |
Each class revealed a unique thermal signature reflecting surface composition and exposure.
Radiance-to-Temperature Conversion
Spectral radiance was converted to effective at-sensor temperature using:
where K₁ = 666.09 and K₂ = 1282.71.
This Planck-based transformation expresses emitted energy as brightness temperature (Kelvin) under unit emissivity. It bridges digital values and environmental meaning (Jensen, 2016).
Temperatures across the Delta State scene ranged from approximately 295 K (vegetation) to 310 K + (pavement and rooftops). This span of over 15 K confirms measurable micro-climatic variation within a compact urban campus, illustrating the urban heat-island effect at a local scale.
Temperature Extraction and Averages
The same ROIs were evaluated for temperature. Patterns mirrored radiance:
Cool vegetation zones (≈ 295 K–298 K), moderate buildings (≈ 303 K), and hot pavement (> 308 K). The consistency validates both calibration constants and processing accuracy.
Smoothing and Pseudocolor Display
A 3 × 3 low-pass filter was applied to smooth pixel noise and enhance regional coherence. In pseudocolor display, gradients replaced speckle, yielding a readable thermal structure, warm cores near infrastructure fading into cooler peripheries.
The resulting map portrays the thermal rhythm of the campus: arteries of warmth and lungs of shade.
Density Slicing and Color Ranges
A density-slice visualization divided the temperature spectrum into ten distinct intervals.
| Range (DN) | Color | Interpretation |
|---|---|---|
| 240–246 | Blue 3 | Coldest—shade / water |
| 252–258 | Blue 2 | Cool vegetation |
| 258–264 | Cyan | Transitional zones |
| 264–270 | Green 1 | Neutral ground |
| 270–276 | Yellow 1 | Warm surfaces |
| 276–282 | Orange 1 | Hot roads |
| 282–288 | Red 1 | Very hot asphalt |
| 288–294 | Red 2 | Extremely hot metal |
| 294–300 | Red 3 | Hottest zones |
| > 300 | White | Possible saturation |
Each range marks a perceptible thermal tier, from natural coolness to engineered heat.
Reflection – Between Science and Conscience
Remote sensing is quantitative, yet its imagery speaks morally. Every red patch is a design decision; every blue patch, a reprieve. The map silently records human influence, the geometry of heat shaped by architecture and neglect.
The Delta State University thermal scene serves as both a study and a mirror: proof that materials dictate microclimate and that sustainability begins with shade, not slogans.
Outro
This experiment began as a calibration exercise and ended as a revelation.
By translating radiance into temperature and color, we measure not just energy but intent, the way humans trade comfort for consequence.
To see the invisible is to acknowledge that even the quiet heat of a campus carries our fingerprint.
References
Jensen, J. R. (2016). Introductory Digital Image Processing: A Remote Sensing Perspective (5th ed.). Pearson.
NASA Earth Observatory. (2021). Color and Temperature in Satellite Imagery.
NASA Landsat Handbook. (2020). Thermal Infrared Sensor (TIRS) Calibration and Radiance Conversion Methods.
U.S. Geological Survey (USGS). (2019). Landsat Surface Temperature Product Guide v2.0. Reston, VA.