Abstract:Ice formation is widespread in industrial and natural environments and often causes various adverse effects such as equipment failures and traffic hazards. This paper systematically reviewed the classification, principles, and development trends of existing ice detection technologies, and horizontally compared the detection ranges and accuracy indicators of contact and non-contact ice detection sensors under each property based on the classification of five physical properties of ice, namely, acoustics, optics, thermology, electricity, and mechanics. The analysis results show that acoustic and optical contact ice detection technologies are applicable to the detection of ice layers or water films at the micrometer to millimeter scale, achieving a maximum thickness resolution of 8 μm, which can meet the detection requirements of 0.3 mm ice layers on aircraft surfaces. Significant scattering and interference induced by micro-scale bubbles and brine channels within ice reduce the thickness detection accuracy, but simultaneously provide a quantifiable acoustic identification basis for ice type differentiation. Thermal contact ice detection technologies achieve detection by monitoring the latent heat release at the moment of freezing or the changes in thermal resistance and heat capacity of the ice body, but it is difficult to precisely distinguish millimeter-level thickness differences and different ice types. Electrical and mechanical contact ice detection technologies can stably achieve an ice layer thickness detection accuracy of 0.10.3 mm; wind tunnel experiments show that they can distinguish ice layers of 0.060.45 mm, and they have been widely used in ice monitoring of aircraft and roads. For the non-contact real-time monitoring requirements of centimeter-level ice thickness on wind turbine blades and transmission lines, camera-based ice detection technologies with a detection accuracy of 110 mm are mostly adopted. The meter-level sea ice detection required for polar scientific expeditions and maritime transportation relies on microwave and LiDAR remote sensing technologies with a spatial resolution of 1 m and the capabilities of ice-water differentiation and dynamic inversion of ice thickness, providing key safety guarantee data. Furthermore, the optimization paths of existing methods in improving the coupling accuracy of multiple schemes and enhancing the generalization ability of intelligent algorithms were further refined, providing theoretical support and practical paths for the prevention and control of ice disasters in key infrastructure such as aerospace vehicles, cold region power grids, and offshore wind power. In the future, ice detection technologies will integrate multimodal sensing technologies such as acoustic, optical, thermal, electrical, and mechanical sensing to achieve high-precision dynamic detection of multiple parameters, including ice type, ice thickness, and ice area.