Dike swarms are ubiquitous on terrestrial planets and represent the frozen remnants of magma transport networks. However, spatial complexity, protracted emplacement history, and uneven surface exposure typically make it difficult to quantify patterns in dike swarms on different scales. In this study, we address this challenge using the Hough Transform to objectively link dissected dike segments and analyze multiscale spatial structure in dike swarms. We apply this method to swarms of three scales: the Spanish Peaks, USA; the Columbia River Flood Basalt Group (CRBG), USA; the Deccan Traps Flood Basalts, India. First, we cluster dike segments in Hough Transform space, recognizing prevalent linearly aligned structures that represent single dikes or dike packets, with lengths up to $10-30x$ the mapped mean segment length. Second, we identify colinear and radial dike segment mesoscale structures within each data set, using the Hough Transform to segment swarms into constituent spatial patterns. We show that for both the CRBG and Deccan Traps, a single radial swarm does not well characterize the data. Instead, multiple and sometimes overlapping mesoscale linear and radial features are prevalent. This suggests a time-evolving transport network where structural inheritance of dike pathways over an extended time is likely common, but large-scale reorganizations of the plumbing system that imply state shifts in crustal stresses or mantle melt supply also occur. We expect that the Hough Transform may find useful applications in a variety of geologic settings where many quasi-linear features, at any scale, are superimposed spatially.