# Orthogonal matrix

A real matrix is orthogonal (or, more precisely, orthonormal) when it has an inverse equal to its transpose[1][2]

PT=P-1

The term comes from the fact that the canonical orthonormal basis of the  is transformed by any orthonormal matrix (and only by orthonormal matrices) into another orthonormal basis.

Each orthonormal matrix represents one orthonormal basis, and reciprocally:

• If the columns of an orthonormal matrix are taken to be a set of vectors, then this set is an orthonormal basis
• If an orthonormal basis is written as the columns of a matrix, then this matrix is orthonormal [1]
• These properties are also valid for lines instead of columns

The concept generalizes to complex matrices as unitary matrices; however, in this case, instead of the transpose it's necessary to use the conjugate-transpose.

## References

1. Vecteurs et matrices, at http://benhur.teluq.uqam.ca, in French
2. Rowland, Todd. "Orthogonal Matrix." From MathWorld--A Wolfram Web Resource, created by Eric W. Weisstein.