The Dunkl transform generalizes the classical Fourier transform in the context of finite reflection groups, introducing a family of operators parameterized by a multiplicity function \(\textbf{k} \ge 0\) . Hypercomplex extensions of integral transforms, particularly based on quaternion and octonion algebras, have gained prominence for processing multi-dimensional signals. In this paper, we introduce the Octonionic Dunkl Transform (ODT), which unifies the algebraic structure of the octonions—the largest normed division algebra—with the analytic framework of Dunkl operators. We provide its explicit definition, accounting for the non-associative nature of octonion multiplication through a fixed left-nested parenthesization convention. Fundamental properties are established, including linearity, a Bessel function series representation, and a detailed parity decomposition that separates the transform into eight real-valued components. We prove a sharp inversion formula and an isometry (Plancherel theorem) in the associated weighted \(L^2\) space. Furthermore, we establish two uncertainty principles: a Heisenberg-type inequality relating the weighted dispersions of a signal and its transform, and a Donoho-Stark-type concentration principle. These results extend earlier work on the quaternion Dunkl transform and the octonion Fourier transform, providing a new theoretical tool for the analysis of octonion-valued signals in settings with reflection symmetry.
