Self-localisation mechanism in an unknown territory has been an interest area for humans since ages. Image matching is an obvious contender due to advancements in imaging devices and compute technologies. Deep learning methods have proven to be state-of-art in recent times but require large volumes of relevant data. Aerial image matching re-mains a challenging task due to the quality of images (e.g. platform disturbances, atmospheric effects), multiple types of on-board sensors (e.g. visual, thermal), variations in scales and look angles etc. To address these challenges, we present a cross-platform path planning dataset composed of images acquired from an aircraft and the Google Earth Engine (GEE). The proposed dataset contains manually aligned frames, corresponding match region, and se-mantic labeling of the images. Multiple galleries representing historical and instantaneous paths are generated. Our dataset envisages several realistic scenarios in cross-platform matching and semantic segmentation. We evaluate the performance of state-of-the-art image matching and segmentation algorithms on the proposed dataset. We will make our dataset freely available at https://www.iith.ac.in/~lfovia/downloads.html. Further, a case study on utilizing an existing open-source dataset for cross-platform path planning is also presented. © 2021 IEEE.