Collaborative scene sharing for overcoming visual obstructions is provided. Video feeds of a scene are obtained from multiple devices viewing the scene from different angle. An obstruction obstructs portion(s) of the scene from view by a device. An aggregation grid that partitions views of the scene into cells is established and used in providing an aggregated view of the scene to the device. This includes identifying cell(s) of the aggregation grid for which the view by the device is obstructed by the obstruction, and providing to the device video portion(s) obtained from video feed(s) from the multiple devices. The video portion(s) correspond to the cell(s) for which the view by the device is obstructed, and the video feeds from which the video portions are obtained are from other device(s), of the multiple devices, for which the view for the cell(s) is unobstructed.