Landsat DN to Spectral Radiance Calculator

An essential tool for remote sensing professionals for calculating offset using gain Landsat data to convert raw pixel values into scientific units.

Calculation Results

Intermediate Values

Copied!

Radiance vs. Digital Number Relationship

What is Calculating Offset Using Gain for Landsat?

In remote sensing, satellite sensors like those on the Landsat series record the intensity of electromagnetic radiation reflected or emitted from the Earth’s surface. However, this raw data is stored as a simple integer value called a Digital Number (DN) for each pixel. This is done to save space and simplify data transmission. The process of calculating offset using gain Landsat data refers to the essential radiometric calibration step that converts these unitless DNs into a physically meaningful unit: at-sensor spectral radiance (Lλ).

This conversion is not just a simple scaling; it’s a linear transformation defined by two unique coefficients for each spectral band of the sensor: a multiplicative gain (G) and an additive offset or bias (B). These values are determined through rigorous pre-launch and on-orbit calibration of the sensor and are provided in the metadata file that accompanies every Landsat scene. Without this conversion, quantitative analysis of satellite imagery would be impossible. If you work with Earth observation data, you might find our Atmospheric Correction Calculator useful for subsequent processing steps.

The Landsat DN to Radiance Formula

The conversion from a Digital Number (Qcal) to Top-of-Atmosphere (TOA) Spectral Radiance (Lλ) is governed by a straightforward linear equation. This formula is the core of calculating offset using gain for any Landsat data.

Lλ = (Gain * Qcal) + Offset

This equation ensures that the raw digital counts are accurately rescaled to represent the actual energy that was recorded by the sensor at the top of the atmosphere.

Formula Variables

Variable	Meaning	Unit (Auto-Inferred)	Typical Range
Lλ	At-Sensor Spectral Radiance	Watts / (meter² * steradian * µm)	Varies by surface type (e.g., 0 to ~100)
Gain (G)	Band-specific multiplicative scaling factor	(W/(m²·sr·µm)) / DN	Typically a very small positive number (e.g., 1.2829E-04)
Qcal	Quantized Calibrated Pixel Value (Digital Number)	Unitless Integer	0-255 (Landsat 1-7), 0-65535 (Landsat 8/9)
Offset (B)	Band-specific additive scaling factor (Bias)	W / (m²·sr·µm)	A small positive or negative number (e.g., -0.2)

Practical Examples

Understanding how the inputs affect the output is key. Here are two practical examples of calculating offset using gain for Landsat 8 data.

Example 1: Healthy Vegetation in the Near-Infrared (NIR) Band

• Inputs:
  • Digital Number (DN): 35000 (a high value, typical for healthy vegetation in NIR)
  • Gain (for Band 5): 9.6675E-05
  • Offset (for Band 5): -0.15
• Calculation: Lλ = (9.6675E-05 * 35000) – 0.15 = 3.3836 – 0.15 = 3.2336
• Result: The spectral radiance is approximately 3.23 W/(m²·sr·µm).

Example 2: Deep Water in the Blue Band

• Inputs:
  • Digital Number (DN): 8000 (a low value, typical for water which absorbs NIR)
  • Gain (for Band 2): 1.2829E-04
  • Offset (for Band 2): -0.2
• Calculation: Lλ = (1.2829E-04 * 8000) – 0.2 = 1.02632 – 0.2 = 0.82632
• Result: The spectral radiance is approximately 0.83 W/(m²·sr·µm).

These examples show how different surface features result in vastly different DN and, consequently, radiance values. For further analysis, consider exploring our NDVI Calculator.

How to Use This Landsat Radiance Calculator

1. Locate Metadata: Find the metadata file (usually ending in `_MTL.txt`) that came with your Landsat data download.
2. Find Calibration Coefficients: Open the file and search for `RADIANCE_MULT_BAND_x` (this is the Gain) and `RADIANCE_ADD_BAND_x` (this is the Offset), where `x` is the band number you are analyzing.
3. Enter Values: Copy the Gain, Offset, and the Digital Number (DN) of a pixel of interest into the corresponding fields in the calculator.
4. Calculate & Interpret: Click the “Calculate” button. The primary result is the spectral radiance in scientific units, which is now comparable across different scenes and sensors. The chart visualizes the linear relationship between DN and radiance for your specific inputs.

Key Factors That Affect Landsat Radiance Values

• Sensor Type: Different Landsat missions (e.g., Landsat 5 TM, Landsat 7 ETM+, Landsat 8 OLI) have different sensors with unique gain and offset values. You must use the coefficients specific to your sensor.
• Spectral Band: Each band (e.g., Blue, Green, Red, NIR, SWIR) is designed to capture a different part of the electromagnetic spectrum and has its own unique gain and offset.
• Atmospheric Conditions: The calculated value is “at-sensor” or “Top-of-Atmosphere” (TOA) radiance. It includes the effects of atmospheric scattering and absorption. To get surface reflectance, an additional step of atmospheric correction is required.
• Sun Angle & Earth-Sun Distance: The illumination conditions at the time of image acquisition significantly impact the amount of radiance received by the sensor.
• Sensor Degradation: Over time, a sensor’s sensitivity can change. Calibration parameters are periodically updated by agencies like the USGS to account for this drift. Always use the most recent metadata files.
• Data Quantization: The bit depth (8-bit for older sensors, 16-bit for newer ones) determines the range of DNs and the precision of the raw data.

Frequently Asked Questions (FAQ)

1. Why can’t I just use the Digital Number (DN) for analysis?

DNs are not a physical unit. They can vary between scenes of the same area taken at different times due to changes in illumination, atmospheric haze, or sensor settings. Converting to radiance is the first step in creating a standardized, comparable scientific product.

2. Where do the gain and offset values come from?

They are derived from extensive calibration efforts by space agencies (like NASA and USGS). The sensor is tested against sources of known radiance levels both before launch and throughout its operational life on-orbit to establish this relationship.

3. Is radiance the same as reflectance?

No. Radiance is the measure of energy at the sensor. Reflectance is a unitless ratio (0-1) of the energy reflected by a surface to the energy incident upon it. To convert from TOA radiance to TOA reflectance, you must also account for solar irradiance and sun angle. Our Landsat Reflectance Calculator can help with this.

4. What does a negative offset mean?

A negative offset (or bias) simply means that a DN of zero does not correspond to a radiance of zero. It is part of the calibration to ensure the full dynamic range of the sensor is utilized correctly. The electronics of the sensor have a baseline reading even in total darkness, and the offset corrects for this.

5. Can the calculated radiance be negative?

Yes, it’s possible for very low DNs and certain gain/offset combinations to produce a small negative radiance value. This is typically considered noise or an artifact, and these values are often clipped to zero in subsequent analysis.

6. Does this process apply to other satellites like Sentinel?

Yes, the principle is the same. Other satellite systems like Sentinel-2 also provide raw data that needs to be scaled using coefficients found in their metadata to convert to radiance, although the specific formula or coefficient names might differ.

7. Why are there different gain/offset values for each band?

Each spectral band uses a different set of detectors with unique sensitivities. The calibration process characterizes each of these detector arrays independently, resulting in unique gain and offset values for each band.

8. What is the difference between “gain” and “offset” in calculating Landsat values?

Gain is the multiplicative factor; it determines the slope of the line relating DN to radiance. It defines how much the radiance increases for each unit increase in DN. Offset is the additive factor; it’s the y-intercept of that line, representing the base radiance when the DN is zero.