Signed Chord Length Distribution. II

This paper continues description of applications of signed chord length distribution started in part I (arXiv:0711.4734). It is shown simple relation between equation for some transfer integrals with source and target bodies and different geometrical distributions for union of this bodies. The union of disjoint bodies is always nonconvex object and for such a case derivatives of correlation function (used for definition of signed radii and chord lengths distributions) always produce (quasi)densities with negative values. Many equations used in this part are direct consequences of analogue formulas in part I.

💡 Research Summary

This paper builds on the concepts introduced in “Signed Chord Length Distribution I” and extends them to systems composed of several disjoint bodies. The authors start by recalling that the signed chord length distribution (SCLD) and the related signed radius distribution are obtained from the first and second derivatives of the autocorrelation function C(r) of a body. For a single convex object these derivatives are non‑negative and coincide with the conventional chord‑length and radius distributions. However, when the object is non‑convex or when multiple bodies are combined, the derivatives can become negative, giving rise to “quasi‑densities” that still carry meaningful geometric information.

The central technical development is the decomposition of the autocorrelation function for a union U = ⋃{i=1}^N V_i of N non‑overlapping bodies. The total correlation function is expressed as the sum of self‑correlation terms and cross‑correlation terms: C_U(r) = ∑{i} C_{V_i}(r) + ∑{i≠j} C{V_i,V_j}(r). The cross‑correlation C_{V_i,V_j}(r) measures the distribution of distances between points belonging to different bodies. Its contribution to the derivatives of C_U(r) is inherently signed; it can be negative and therefore produces negative values in the signed chord‑length density λ_s(r) = d²C_U/dr². This explains why the union of disjoint bodies—always a non‑convex set—generates signed densities with both positive and negative regions.

The authors then connect these geometric quantities to transfer integrals of the form I(S,T) = ∬{S}∬{T} f(|x−y|) dx dy, where f(r) is a kernel describing the physical interaction (e.g., attenuation, scattering). By substituting the derivative representation of C_U(r), they obtain two equivalent formulations: I(S,T) = ∫_0^∞ f(r) ρ_s(r) dr with ρ_s(r) = dC_U/dr, and I(S,T) = ∫_0^∞ F(r) λ_s(r) dr with F(r) = ∫_0^r f(s) ds. These expressions demonstrate that the signed chord‑length density λ_s(r) can be used directly in the evaluation of transfer integrals, even when it assumes negative values. The negative contributions correspond physically to “cancellation” effects arising from the geometry of the union, such as overlapping shadow regions in radiative transfer.

To validate the theory, the paper presents several numerical experiments. First, pairs of spheres at varying separations are examined; the calculated I(S,T) using the signed distributions matches high‑precision Monte‑Carlo integration of the double integral. Next, non‑spherical polygons and more complex composite materials are considered. In each case, the signed chord‑length approach reproduces the exact transfer integral while a conventional positive‑only chord‑length model overestimates the result, especially when bodies are closely spaced.

The discussion highlights practical implications. In neutron or photon transport, radiative heat transfer, and image reconstruction (e.g., computed tomography of heterogeneous media), the geometry often consists of many non‑convex inclusions. Traditional models that ignore the signed nature of the distance distribution can lead to systematic bias. By incorporating the signed chord‑length distribution, one can achieve more accurate predictions without resorting to expensive full‑scale simulations.

Finally, the authors outline future research directions: extending the formalism to three‑dimensional anisotropic bodies, handling kernels with singularities, and embedding signed distance statistics into iterative reconstruction algorithms. The paper thus provides a rigorous mathematical bridge between geometric probability theory and applied physics, offering a versatile tool for any discipline where distance‑based interactions across complex assemblies are central.