What does satellite image resolution mean?
Resolution in satellite imagery does not mean only the number of meters represented by a pixel. Spatial resolution concerns sampled ground detail, spectral resolution separates wavelengths, temporal resolution concerns observation frequency, and radiometric resolution concerns numerical sensitivity to signal levels. Smaller pixels can be useful for a particular purpose, but they do not automatically produce a more accurate conclusion.
Four different dimensions of resolution
1. What is spatial resolution? The ground area represented by a sensor sample limits visible spatial detail. A 30-meter Landsat pixel records combined light from that area rather than every small tree individually. A road being wider than a pixel does not always guarantee easy detection; contrast and neighboring surfaces matter too. Enlarging an image on a screen creates more display squares, not new observations. NASA's Landsat mission descriptions distinguish sampling sizes among panchromatic, visible and thermal bands. Read the properties of the actual band and product rather than attaching one number to every wavelength or assuming that the size of a web tile is the size of a measured footprint. 2. What changes with spectral resolution? A sensor may observe visible, near infrared, shortwave infrared and thermal wavelengths in separate bands. Narrower or more numerous bands may help distinguish materials, with tradeoffs in signal, noise and data volume. The three color channels shown on a monitor are not the entire sensor output. Vegetation and snow require different band choices. A band called red on two missions may not cover precisely identical wavelengths, so cross-sensor comparisons need specifications and calibration. Spectral detail helps identify patterns but is not automatic material identification without field or independent reference data. 3. What does temporal resolution mean? How often a place can be observed depends on orbital design, swath width and the number of satellites. A nominal repeat cycle is different from the availability of cloud-free usable imagery. In a cloudy region, a sixteen-day repeat does not guarantee one clear ground observation every sixteen days. A long, consistent archive may suit annual forest change, whereas a flood evolving within hours may require more frequent sensing. Check actual acquisition dates in the catalog instead of treating theoretical recurrence as a promise for one location.
Choosing a sensor and data product
4. What is radiometric resolution? Radiometric resolution describes how many numerical levels can represent the energy detected by a sensor. More levels may preserve subtler differences, but cannot by themselves correct noise or poor calibration. A scientific raster exported as an eight-bit picture for display may compress its original range. Do not confuse a quantitative surface reflectance dataset with a contrast-stretched JPEG. Product documentation should identify data type, scale factors, valid ranges and missing-data values. A visually smoother image can be less suitable for measurement if its numerical values have been altered for presentation. 5. Why is a pixel not necessarily one object? A small plot, road, building and shadow can contribute to one pixel's signal. The mixed-pixel problem depends on object size and placement as well as the stated resolution. Very high detail can show more geometry, while frequent consistent coverage of a large region may be harder to obtain. Resampling two datasets to a common pixel grid does not restore information that the coarser sensor never measured. Report both the ground sampling and the scale at which results are interpreted; a regional classification cannot validate individual properties from pixels too coarse for them. 6. Which product should be chosen? An annual vegetation study across a continent may favor a repeatedly calibrated multiyear archive, while examining a single building requires finer geometric detail. A thermal band can have a coarser native footprint than visible bands; delivering it on a smaller output grid does not create finer thermal observations. NASA mission pages list spatial, spectral and temporal characteristics separately. Select a product by the target size, required wavelengths, cloud-free date window, data license and independent validation options. The most impressive single resolution figure is rarely sufficient to make a sensor suitable for a research question.