The radiative cooling technology is a method of directly transmitting radiative energy to the cold sink of outer space through the existing atmospheric transparent window of 8–13​ $\mathrm{\mu }$m. Therefore, it is necessary to find a selective emitter that can match the infrared atmospheric window. Here, we combine the selective emitter with the inverse design of the direct binary search (DBS) algorithm numerically. The periodic unit is glass photonics crystal (GPC) modified by DBS algorithm and a bottom layer of silver separated by a 11-$\mathrm{\mu }$m-thickness layer of glass. The simulation results reveal that the average emissivity of 8–13​ $\mathrm{\mu }$m has been optimized from 71.8% to 94%. In addition, the proposed emitter remains a high reflectivity of 90.1% for the wavelength range from 1 $\mathrm{\mu }$m to 6 $\mathrm{\mu }$m in order to reduce the energy of absorption outside the transparent window. Both of the incident-angle insensitivity and polarization insensitivity are verified. With the above merits, the proposed emitter opens a novel approach to applications of radiative cooling based on inverse design.

辐射冷却技术是一种通过现有的8-13号大气透明窗将辐射能直接传输到外层空间冷沉的方法 $\mathrm{\mu }$米。因此，有必要寻找一种能够匹配红外大气窗口的选择性发射器。在这里，我们在数值上将选择性发射器与直接二分搜索（DBS）算法的逆向设计相结合。周期单位是由 DBS 算法修改的玻璃光子晶体 (GPC) 和由 11-$\mathrm{\mu }$m 厚的玻璃层。仿真结果表明，平均发射率为 8-13$\mathrm{\mu }$m 已从 71.8% 优化到 94%。此外，所提议的发射器在波长范围从 1$\mathrm{\mu }$米到 6 $\mathrm{\mu }$m 以减少透明窗外的吸收能量。验证了入射角不敏感和偏振不敏感。凭借上述优点，所提出的发射器开辟了一种基于逆向设计的辐射冷却应用的新方法。