One of the challenges of the snowflake divertor (SFD) configuration is finding a reliable means of reconstructing the magnetic field geometry in the divertor. Since the SFD (and other advanced divertors) have multiple field nulls, there is a large region with shallow flux gradients that is difficult to resolve accurately using external diagnostics. In this work we present a technique that uses heat flux measured by the infrared television (IRTV) camera to improve SFD reconstruction. This is relevant for purposes of control, since the SFD is topologically unstable and requires active feedback on the shape [E. Kolemen, et. al., Nucl. Fusion, 58, 6 (2018)], and analysis, since reconstructions provided by other algorithms such as EFIT [L. Lao, et. al., Nucl. Fusion, 25, 11 (1985)] can mis-characterize the shape and even the snowflake type (plus or minus). The technique identifies the spatial position of the two x-points located in the SFD based on characteristics of the heat flux such as the strike point location and power distribution. The inferred x-point positions are then used as a constraint in fitting new equilibria using the TokSys suite of software. This procedure is applied to ~800 DIII-D SFD timeslices and reduces the summed strike point errors from an average 9.4 cm to 0.9 cm. The newly-created x-point constrained equilibria are compared to kinetic reconstructions and an average 16% reduction in the edge current is observed. This is correlated via a simple linear relationship to the shape constraints. Other changes in the pedestal structure are observed, but more work must be done to incorporate the IRTV constraint directly into kinetic solvers to obtain integrated solutions.