In this paper, we present a novel linear method for the simultaneous estimation of the homography matrix and one-sided radial lens distortion. Initially, we highlight that Fitzgibbon’s method, commonly recognized as a DLT method for this problem, is inadequate for handling noisy data. Subsequently, we formulate a new DLT method incorporating lens distortion by considering the inverse homography transformation. The proposed method, termed invDLT, provides two solutions: the minimal case with 4.5 point pairs and the least-squares case with more than five point pairs. We conduct extensive experiments on both synthetic and real image data, revealing that invDLT substantially outperforms conventional methods in terms of estimation accuracy, robustness to outliers, and computational efficiency.