A fast collocation method for the radiosity equation, based on the hierarchical algorithm of Hanrahan and Salzman: the 1D case (Q1774028)

From MaRDI portal
scientific article
Language Label Description Also known as
English
A fast collocation method for the radiosity equation, based on the hierarchical algorithm of Hanrahan and Salzman: the 1D case
scientific article

    Statements

    A fast collocation method for the radiosity equation, based on the hierarchical algorithm of Hanrahan and Salzman: the 1D case (English)
    0 references
    0 references
    29 April 2005
    0 references
    The author studies a modified collocation method for the solution of the planar radiosity equation \[ u(x)=E(x)+\frac{\rho(x)}{2}\int_LV(x, y)\frac{n(x)\cdot(y-x)\,n(y)\cdot(x-y)}{\| x-y\|}\,u(y)\,dy,\quad x\in L, \tag{1} \] where \(L\) is a curve in \(\mathbb R^2\), \(n(\cdot)\) is the curve normal which should exist almost everywhere, \(E(x)\) is the emissivity, \(\rho:L\to[0,1]\) is the reflection coefficient and \(V(x,y)\) is visibility function. Only the unoccluded case \(V = 1\) is considered. For the approximate solution of (1) the author uses midpoint collocation with piecewise constant trial functions. The complicated regularity structure of the solution of the radiosity equation makes the efficient use of higher order trial functions very difficult. So the author considers a modified collocation method in which the matrix of the collocation method is approximated by a method developed by \textit{P.~M.~Hanrahan, D.~Salzman} and \textit{L.~Aupperle} [A rapid hierarchical radiosity algorithm, in: SIGGRAPH '91 Proceedings, Computer Graphics 25, No.~4, 197--206 (1991)]. The author proves that the modified collocation method results in a reduction of work while the order of convergence stays the same. Numerical examples demonstrate the theoretical results for two model problems.
    0 references
    0 references
    0 references
    0 references
    0 references
    planar radiosity equation
    0 references
    modified collocation method with piecewise constant trial functions
    0 references
    hierarchical method
    0 references
    order of convergence
    0 references
    numerical examples
    0 references
    0 references