Positional accuracy

Learn more about positional accuracy standards.


Overview

Positional accuracy indicates the reliability of the horizontal and vertical positions of geospatial features relative to their datums. Accuracy standards allow us to quickly and consistently establish positional reliability across different imagery.

Accuracy types

  • Absolute accuracy

    It measures how closely a specific feature in an image aligns with its actual ground position on Earth.

  • Relative accuracy

    It measures how closely the positions of features align relative to each other within a given image.

Accuracy standards

Circular error (CE)

CE indicates the possible horizontal distance by which a specific feature might deviate from its actual ground position. The number after CE specifies the confidence value.

For example, an accuracy of 5 m CE90 means you can be 90% confident a specific feature in an image is within a 5 m radius circle around where the image suggests.

Linear error (LE)

LE indicates the possible vertical distance a specific feature might deviate from its actual ground position. The number after LE specifies the confidence value.

For example, an accuracy of 2 m LE95 means you can be 95% confident a specific feature is within 2 m vertically of where the image suggests.

Root mean square error (RMSE)

RMSE indicates the average distance a specific feature might deviate from its actual ground position in a given direction. The letters after RMSE specify whether it is measured in the horizontal x and y or the vertical z directions.

For example, an accuracy of 10 m RMSEx/y/z means a specific feature may vary up to 10 m on average from its actual ground position, measured in the x, y, and z directions.

Spherical error (SE)

SE indicates the possible distance a specific feature might deviate from its actual ground position in all directions. The number after SE specifies the confidence value.

For example, an accuracy of 3 m SE90 means you can be 90% confident a specific feature is within a 3 m radius sphere around where the image suggests.

Troubleshooting

What does “at nadir” mean?

When a sensor points at a target directly below the sensor, it points at nadir. Optical sensors are most accurate when pointing at nadir due to the minimal distance between the sensor and the target.