Terrestrial laser scanning (TLS) is commonly used for acquiring dense point clouds, used to generate high-quality 3D models, which supports Building Information Modelling (BIM) and digital twinning. However, not only is careful TLS planning necessary to ensure data completeness while minimising scanning time, but real-world conditions often introduce occlusions that prevent original scan plans from achieving the intended coverage. This study presents a method for determining optimal scanner locations through efficient discretisation of target object surfaces into key-points. In addition, it addresses the challenge of dynamically and efficiently adapting scan plans by evaluating whether new scan locations are necessary when occlusions limit visibility. This research uses 3D surface discretisation (into key-points) as the primary method for efficiently evaluating object 3D surface coverage, used during both initial scan planning and dynamic re-planning. A next-best-view algorithm is applied for overall scan plan generation. The method is validated through simulation using a model from the International Society for Photogrammetry and Remote Sensing (ISPRS) dataset, Helios++, and random insertion of clutter. This work contributes to automating TLS workflows, making them more adaptive to real-world uncertainties. Future work will focus on real-world implementation and integration with robotic scanning systems for enhanced automation.