Virtual laser scanning of dynamic scenes (VLS-4D): Current approaches and future perspectives in remote sensing

Virtual laser scanning (VLS) has proven to be a useful tool for survey planning, method development and training data generation in a variety of areas of Earth and environmental sciences. Until recently, most applications have used static representations of the real or a fictive environment, neglecting the inherent dynamics of our world, that also affect Light Detection and Ranging (LiDAR) measurements. Given the enormous potential of LiDAR simulation to support the monitoring of dynamic phenomena such as landslides, tree sway or urban change, this review provides an overview of current approaches to virtual laser scanning of dynamic scenes (VLS-4D). We first build a theoretical framework that includes the relevant types of changes to scene objects and the strategies by which they are implemented in the simulation. Furthermore, we review methods for generating dynamic scenes as input to VLS, present existing frameworks supporting VLS-4D, and highlight the main scientific objectives of the VLS-4D studies published so far. Despite the established use of VLS-4D in robotics and autonomous driving, there are few examples in environmental science where high fidelity is required not only of the scene and its dynamics, but also of the simulated ray-scene interaction. With our work, we aim to direct future research and encourage geoscientific disciplines to adopt VLS-4D as an environment for experimentation, method development and data generation and permutation.