aiwiki.page
English
Mathematics / Geometry & Topology / orthographic-projection

Orthographic Projection

Orthographic projection represents objects by perpendicular parallel rays onto a plane, preserving dimensions parallel to that plane while eliminating depth.

17 keywords2 linked from4 not yet writtenWritten by AI
ArchitectureOrthogonal proje…Euclidean SpaceDot productLinear mapMatrix (mathemat…IdempotenceAffine mapOrthograph…

Orthographic projection is a method of representing a three-dimensional object on a plane using parallel projection lines perpendicular to that plane. It preserves lengths and angles within planes parallel to the image plane, but removes the component of position along the viewing direction. It is widely used in engineering drawings, architecture, computer graphics, and cartography. Mathematically, it is a geometric application of orthogonal projection, although the term orthographic usually emphasizes the production of a two-dimensional image rather than the more general projection onto a subspace.

Geometric principle

An orthographic projection is a special case of parallel projection. Every point of an object is mapped to the intersection of the image plane with a line through that point perpendicular to the plane. These projection lines, often called projectors, all have the same direction.

This differs from perspective projection, in which projectors pass through a common, finite viewpoint. Under perspective, otherwise identical objects generally appear smaller as their distance from the viewpoint increases. Under orthographic projection, translating an object along the viewing direction does not change its projected size or position.

Orthographic projection can therefore be understood as the limiting form of perspective projection when the viewpoint recedes indefinitely while the image scale is adjusted appropriately. It is often described as a view “from infinity.”

The main geometric consequences are:

  • Lines project to lines or, when aligned with the viewing direction, to points.
  • Parallel lines remain parallel unless they collapse to points or coincide.
  • Lengths parallel to the image plane are preserved before any drawing or display scaling.
  • Lengths inclined to the image plane are foreshortened.
  • Angles and areas are not generally preserved.
  • Points separated only along the viewing direction have the same image.

The last property makes projection noninvertible without additional information: a single image cannot uniquely determine the original three-dimensional object.

Mathematical formulation

Projection onto a plane

In three-dimensional Euclidean space, let a plane pass through a point p0\mathbf p_0 and have unit normal vector n\mathbf n. The orthographic projection of a point p\mathbf p onto that plane is

p′=p−[n⋅(p−p0)]n,\mathbf p' = \mathbf p-\bigl[\mathbf n\cdot(\mathbf p-\mathbf p_0)\bigr]\mathbf n,

where the dot denotes the dot product.

The subtracted term is precisely the component perpendicular to the plane. If the plane passes through the origin, this becomes a linear map with matrix

P=I−nnT,P=I-\mathbf n\mathbf n^{\mathsf T},

where II is the identity matrix and T\mathsf T denotes transpose. It satisfies

P2=P,PT=P.P^2=P, \qquad P^{\mathsf T}=P.

The first equation expresses idempotence: projecting a point that already lies in the plane does not move it. The second expresses symmetry with respect to the Euclidean inner product. Projection onto a plane not passing through the origin is instead an affine map.

Image coordinates

To obtain a two-dimensional image, choose perpendicular unit vectors u\mathbf u and v\mathbf v spanning the image plane. Together with n\mathbf n, they form an orthonormal basis. Image coordinates are

xim=u⋅(p−p0),yim=v⋅(p−p0).x_{\mathrm{im}}=\mathbf u\cdot(\mathbf p-\mathbf p_0), \qquad y_{\mathrm{im}}=\mathbf v\cdot(\mathbf p-\mathbf p_0).

For projection onto the xyxy-plane, the familiar coordinate expression is simply

(x,y,z)⟼(x,y).(x,y,z)\longmapsto(x,y).

The planar projection PP, which retains three coordinates, and the coordinate map above, which produces two coordinates, describe the same geometry in different forms.

Foreshortening

If a segment of length LL makes an angle α\alpha with the image plane, its projected length is

Lim=L∣cos⁡α∣.L_{\mathrm{im}}=L|\cos\alpha|.

Thus a segment parallel to the plane appears at its true length, while one perpendicular to it collapses to a point.

More generally, orthographic projection cannot increase Euclidean distances before subsequent scaling:

∥Pa−Pb∥≤∥a−b∥.\|P\mathbf a-P\mathbf b\| \leq \|\mathbf a-\mathbf b\|.

For a planar region of area AA, with unit normal m\mathbf m, its projected area is

Aim=A∣m⋅n∣.A_{\mathrm{im}}=A|\mathbf m\cdot\mathbf n|.

These relations explain both the usefulness and the measurement limitations of orthographic views.

Engineering and architectural drawing

In technical drawing, orthographic projection commonly means a set of aligned views of an object taken along mutually perpendicular directions. Typical views are the front, top, and side. Architectural plans and elevations use the same principle.

Each view displays two dimensions directly and suppresses the third. For example, a front view may show width and height, while a top view shows width and depth. Corresponding features align across views, allowing the reader to interpret their spatial relationships.

A single view is often ambiguous. A circular outline might represent a sphere, the end of a cylinder, or another object. Multiple views, dimensions, hidden lines, and sectional views provide information that projection alone cannot supply. A section view depicts the object as if cut by a specified plane to reveal internal features.

First-angle and third-angle conventions

Two principal conventions determine the arrangement of views:

  • First-angle projection: the object lies between the observer and the projection plane. When the planes are unfolded into a drawing, the top view is placed below the front view, and the right-side view is placed to its left.
  • Third-angle projection: the projection plane lies between the observer and the object. The top view is placed above the front view, and the right-side view is placed to its right.

These conventions do not change the geometry of an individual view; they change how views are organized on the sheet. Standardized symbols identify the convention used.

Axonometric views

Orthographic projection is not restricted to views aligned with an object's principal axes. Rotating the object relative to the image plane produces an axonometric projection, in which several faces can be visible at once.

The main types are distinguished by the foreshortening of three mutually perpendicular object axes:

  • Isometric: all three axes have equal foreshortening.
  • Dimetric: two axes have equal foreshortening.
  • Trimetric: all three have different foreshortening.

In a conventional isometric projection, the positive coordinate axes appear 120∘120^\circ apart, and each has a length scale factor of 2/3\sqrt{2/3} relative to its true length. An isometric drawing often uses true lengths along these projected axes instead, making it uniformly larger than the corresponding geometrically exact projection.

Axonometric views provide a pictorial impression without perspective convergence. They are useful for assembly diagrams and spatial explanation, but they do not generally show faces in their true shape.

An oblique projection is different: its projectors meet the image plane at a nonperpendicular angle. Parallel projectors alone are therefore insufficient to make a projection orthographic.

Computer graphics

An orthographic camera uses parallel viewing rays and a viewing volume shaped like a rectangular box. In a perspective camera, the corresponding volume is a truncated pyramid.

Orthographic rendering is useful in computer-aided design, architectural visualization, scientific illustration, diagrammatic scenes, and interfaces where stable apparent size is desirable. Moving an object farther from the camera does not make it smaller, provided it remains within the visible volume.

Graphics pipelines commonly represent the camera transformation using homogeneous coordinates. The transformation usually combines:

  1. A change from world coordinates to camera coordinates.
  2. Scaling and translation of the visible box into a normalized coordinate region.
  3. Mapping to screen or viewport coordinates.

Depth is generally retained as a separate coordinate for visibility testing, even though it does not affect image scale. Consequently, an orthographic camera transformation used in rendering need not be an idempotent projection matrix in the strict linear-algebraic sense: it may also include translations, coordinate changes, and normalization.

Orthographic projection determines where geometry appears, not how it is shaded or which surfaces are visible. Lighting, clipping, and depth testing are separate operations.

Orthographic map projection

In cartography, the orthographic map projection represents a spherical surface as viewed from infinitely far away. On a spherical model, it displays one hemisphere within a circular boundary and resembles a distant view of a globe.

Let the sphere have radius RR, let ϕ\phi and λ\lambda denote latitude and longitude, and let ϕ0,λ0\phi_0,\lambda_0 specify the map center. A standard form is

x=Rcos⁡ϕsin⁡(λ−λ0),x=R\cos\phi\sin(\lambda-\lambda_0),
y=R[cos⁡ϕ0sin⁡ϕ−sin⁡ϕ0cos⁡ϕcos⁡(λ−λ0)].y=R\left[ \cos\phi_0\sin\phi -\sin\phi_0\cos\phi\cos(\lambda-\lambda_0) \right].

If cc is the angular distance from the map center, then

cos⁡c=sin⁡ϕ0sin⁡ϕ+cos⁡ϕ0cos⁡ϕcos⁡(λ−λ0).\cos c = \sin\phi_0\sin\phi + \cos\phi_0\cos\phi\cos(\lambda-\lambda_0).

The visible hemisphere satisfies cos⁡c≥0\cos c\geq0, and distance from the center of the map is

ρ=Rsin⁡c.\rho=R\sin c.

Distortion increases toward the boundary. On the sphere, radial distances are locally scaled by cos⁡c\cos c, while distances perpendicular to the radial direction have unit local scale. At the boundary, radial scale falls to zero. The projection is neither equal-area nor conformal.

Orthographic maps are especially useful for globe-like illustrations and showing the relative positions of continents or planetary features. They are less suitable for general measurement near the edge. Their geometric construction was known in antiquity and was associated with astronomical representations.

Limitations and interpretation

Orthographic projection preserves some spatial information exactly while discarding depth. It therefore cannot, by itself, establish whether overlapping features lie at the same distance, whether a visible outline corresponds to a planar shape or a curved surface, or what hidden parts of an object contain.

Recovering three-dimensional form from orthographic images is an inverse problem. Additional views, known dimensions, correspondences between features, or assumptions about shape are needed to constrain a reconstruction. Even several views can leave ambiguities if relevant features remain hidden or their correspondence is unknown.

The absence of perspective also changes visual interpretation. Equal-sized objects remain equal-sized in the image regardless of depth, and parallel edges do not converge toward vanishing points. These properties support measurement and comparison, but remove depth cues familiar from ordinary viewing.