In this study, a methodology is proposed for the detection of artifacts by a robotic system against a complex background. The artifact under scrutiny is a diminutive unmanned aircraft of the airplane category. The information channel of the robotic system is a passive optoelectronic system that operates within the 8–13 µm wavelength range (LWIR). The atmospheric background, on which the robotic system performs the detection of unmanned aircraft, is formed by the cloud atmosphere's own radiation when observed from the ground surface. Of particular interest is the complex (inhomogeneous) atmospheric background created by clouds of different classes, characterized by the presence of discontinuities and/or a significant water gradient with small angular values. The method is predicated on the background detection principle and comprises three approaches for obtaining the necessary optical-information components: spatial, spatio-temporal, and temporal. The fundamental principle underlying this method is the capacity to detect local variations in parameters and characteristics of the spatial radiation pattern of atmospheric background inhomogeneities, even in the absence of priori information regarding the presence of an unmanned aircraft within the field of view of the passive optoelectronic system (Yakimenko et al. in Range Light & Engineering 32–3:38–42, 2024; Yakimenko in Ranges Light & Engineering 32–5:17–22, 2024). Changes in these parameters and characteristics are attributable to the distortion of the natural regularities of the spatial structure of atmospheric background radiation by the presence of unmanned aircraft radiation in this region of space.