Interior Orientation in Photogrammetry: Making Cameras “Ideal”

In photogrammetry, Interior Orientation (IO) is a fundamental step that models how a real camera captures an image. The goal is to mathematically describe the geometry of the camera so that we can relate the image coordinates to the real-world coordinates accurately.

What Is Interior Orientation?

Interior orientation refers to the process of transforming a distorted, real camera image into the coordinate system of an ideal camera—often modeled as a pinhole camera. An ideal camera has:

• A single, infinitesimally small aperture
• No lens distortions
• A perfectly flat image plane
• Unit focal length

By defining this ideal model, we can ensure that every image point has a precise relationship to its corresponding ground point.

Image Coordinate Systems

Real images have distortions caused by lenses, misalignments, and other imperfections. To handle this, photogrammetrists define several coordinate systems:

1. Pixel Coordinate System – based on the image array; origin usually at the top-left corner.
2. Distorted Principal Point System – moves the origin to the principal point, the intersection of the optical axis with the image plane.
3. Undistorted Ideal Camera System – corrects for lens distortions and scales coordinates relative to a unit focal length.

Mathematical Model

To model interior orientation, we consider both radial and tangential distortions using parameters such as $k_1, k_2, k_3, p_1, p_2$k1​,k2​,k3​,p1​,p2​, along with scale factors $\lambda$λ and offsets $\delta$δ. The corrected image coordinates $(x’, y’)$(x′,y′) are computed as:$$x’ = x_n + x_n (k_1 r^2 + k_2 r^4 + k_3 r^6) + 2 p_1 x_n y_n + p_2 (r^2 + 2 x_n^2) + \lambda x_n + \delta y_n$$x′=xn​+xn​(k1​r2+k2​r4+k3​r6)+2p1​xn​yn​+p2​(r2+2xn2​)+λxn​+δyn​ $$y’ = y_n + y_n (k_1 r^2 + k_2 r^4 + k_3 r^6) + 2 p_2 x_n y_n + p_1 (r^2 + 2 y_n^2) + \lambda y_n + \delta x_n$$y′=yn​+yn​(k1​r2+k2​r4+k3​r6)+2p2​xn​yn​+p1​(r2+2yn2​)+λyn​+δxn​

Where:

• $x_n, y_n$xn​,yn​ are normalized coordinates relative to the principal point
• $r = \sqrt{x_n^2 + y_n^2}$r=xn2​+yn2​​ is the radial distance from the principal point

These formulas account for lens distortion, decentering, and scaling effects, converting the real image to the ideal coordinate system.

Why Interior Orientation Matters

1. Accuracy – Correcting lens distortions ensures precise mapping of image points to ground coordinates.
2. Collinearity Condition – Interior orientation allows the use of linear geometric models to relate images and objects in space.
3. Foundation for 3D Reconstruction – Any further steps, like relative and absolute orientation, depend on a properly oriented image coordinate system.

By applying interior orientation, photogrammetrists effectively “normalize” the camera, turning a real-world image into a mathematically consistent, idealized model that can be used for mapping, 3D modeling, and analysis.