Landsat Missions
Landsat 7 images are delivered to customers in one of five levels, depending on the degree of processing performed on the data: Level 0 Reformatted (L0R) , Level 1 Radiometrically Corrected (L1R), Level 1 Geometrically Corrected (L1G), Level 1 Systematically Terrain Corrected (L1Gt), and Level 1 Terrain Corrected (L1T). A brief explanation of each is given below with small image subsets to illustrate the differences.
Level 0 Reformatted (L0R)
The Level 0R data product is reformatted raw data. Reformatting involves shifting pixels by integer amounts to correct for three effects: 1) the alternating forward-reverse scanning pattern of the Landsat ETM+ sensor; 2) the odd-even detector arrangement within each band; and 3) the detector offsets inherent in engineering design of the focal plane array. Pixels in L0R images are not resampled, nor are they geometrically corrected or registered, which means that the pixels are NOT aligned per scan line.
Figure 1 shows a small city in an arid region, with the crossing runways of an airport in the center of the image. The scans in this image have been aligned to a common starting point at the extreme left of the image (not shown). Visible scan misalignment is still seen in both runways and in the roads to their left. Although radiometric artifacts are not readily apparent in this sample, any L0R image will still contain any artifacts such as impulse noise, coherent noise, memory effects, etc.
Figure 1. L0R Example
Level 1 Radiometrically Corrected (L1R)
The Level 1R data product is a radiometrically corrected L0R product. L1R products have three improvements: 1) detector artifacts, such as coherent noise, are corrected; 2) cosmetic artifacts, such as banding, striping, and dropped lines or pixels are also corrected; and 3) the L1R image is calibrated to radiance units (or color corrected) as integer values. Radiometric corrections are not reversible. As with L0R, pixels are not resampled nor are they geometrically corrected or registered, which means that the pixels are NOT aligned per scan line.
Figure 2 shows the Level 1R image. The same "stretch" was applied to this image as the L0R sample image. The radiometric corrections mentioned above are typically subtle and not readily apparent to casual visual inspection. Scan misalignments are not corrected in L1R processing and are still visible in the image. L1R imagery also has slightly less color depth than L0R or L1G imagery because L1R pixels are scaled to 16-bit color values and L0R and L1G imagery are scaled to 24-bit color values. A visual comparison of the two pixel value scales is illustrated here.
Figure 2. L1R Example
Level 1 Systematically Corrected (L1G)
The Level 1G (L1G) data product provides systematic radiometric and geometric accuracy, which is derived from data collected by the sensor and spacecraft.
Figure 3 shows a small subset of the L1G image. Scans have been aligned, straightening the runways and roads. Although these data have had radiometric corrections applied, they were reduced from 16-bit to 8-bit data.
Figure 3. L1G Example
The distribution of Landsat 7 ETM+ Terrain correction data is restricted to the U. S. Government and its affiliated users. Please contact USGS EROS Customer Services if you have questions regarding this policy.
Level 1 Systematically Terrain Corrected (L1Gt)
The Level 1Gt (L1Gt) data product provides systematic radiometric and geometric accuracy, while also employing a Digital Elevation Model (DEM) for topographic accuracy. Topographic accuracy of the product depends on the resolution of the DEM used.
Level 1 Terrain Corrected (L1T)
The Level 1T (L1T) data product provides systematic radiometric accuracy, geometric accuracy by incorporating ground control points, while also employing a Digital Elevation Model (DEM) for topographic accuracy. Geodetic accuracy of the product depends on the accuracy of the ground control points and the resolution of the DEM used.
Level 1T SLC-off Segment-based Gap-fill
This Level 1T (L1T) product (includes systematic radiometric accuracy, geometric accuracy by incorporating ground control points, while also employing a Digital Elevation Model for topographic accuracy) incorporates a multi-scale segmentation model to fill the gaps in the SLC-off image. The segmentation model is based on terrain-corrected ETM+ SLC-on data from circa 2000. The gaps are filled with the same-day radiometric values using a nearest neighbor algorithm, while maintaining the landscape patterns derived from the segmentation model.