Airborne laser scanning (ALS) point clouds are employed throughout the country for the generation of digital surface models and to derive further information about forested areas. This acquisition method is considered state-of-the-art up to now, especially in forested areas. However, as these are regularly but not frequently collected in Austria and Europe, modelling based on ALS data of forest parameters in high temporal resolution is difficult. In particular, the derivation of dynamic information such as biomass or condition of a tree population after environmental events such as storms or forest fires or the monitoring of protected areas requires relatively high temporal resolution. Aerial images, along with image-based point clouds derived from them, provide a further option for the creation of surface models. This data is recorded at shorter intervals, such as annually in Vienna or every three years for the entirety of Austria. Especially in areas with high vegetation cover such as forests, the two modelling approaches yield different elevation values. The aim of this study is to systematically quantify these differences and to investigate strategies to approximate IM models to the ALS models. For this investigation, a specific area within the Wienerwald, in the area of the Lainzer Tiergarten, was selected for the development and evaluation of such a process to minimise the height differences. Initially, topographic models, such as the normalised digital surface model (nDSM), were derived from the available point clouds. Statistical parameters for different kernel sizes of the image matching nDSM were then calculated within a specially defined canopy mask. These parameters, along with the known deviation between the laser scanning and image matching model, were used to train a random forest regression to create a model to fit the image matching with the airborne laser scanning data. The validation, conducted on three distinct areas, showed an approximation of the elevation values to the laser scanning nDSM utilised as a reference within the canopy mask. This improvement demonstrates a remarkable approximation of the two models of about 77% in relation to the median of the deviations between the adjusted and the given model compared to the initial situation. The image matching data shows its limitations in elongated gaps in the canopy, as the closing effects of small canopy gaps in forested areas pose challenges for the matching of the images. In such instances, the regression function cannot make any improvements.