Reduction and reconstruction strategy of active thermal control system based on unsupervised learning and thermo-optics for spaceborne high-resolution remote sensor

•An intelligent improvement method for active thermal control system (ATCS) of spacecrafts.•A reduction and reconstruction strategy based on an unsupervised learning framework.•The optimal spacecrafts’ ATCS can be identified automatically at low computational cost.•The reconstructed ATCS has a 78.6 % reduction rate and take only 3 h.•The feasibility/validity of the strategy are verified by orbital thermal and thermo-optical analyses.