Evaluation and experimental validation of the recognition distance for image-spectrum collaborative IRST systems

At long distances, aerial targets typically appear as point targets, making their recognition a significant challenge in both industrial and military domains. The image-spectrum collaborative detection strategy offers a key advantage by overcoming the limitations of traditional imaging-based infrared search and track (IRST) systems, which suffer from restricted spatial resolution and contrast. However, systematic modeling and experimental validation of its recognition performance remain limited. To address this gap, this paper proposes a spectral recognition distance model tailored to image-spectrum collaborative IRST systems, establishing a quantitative relationship between recognition distance and the spectral signal-to-interference ratio (SIR). Experimental validation was conducted using both cooperative and non-cooperative targets. For the cooperative scenario, the model achieved a normalized mean square error (NMSE) below 0.1, while for the non-cooperative case, the relative deviation was 0.09. Furthermore, the spectral channel successfully recognized the target at a distance of 25.5 km using a smaller 42 mm entrance pupil, even though the image channel failed with a larger 52 mm entrance pupil due to a signal-to-noise ratio (SNR) below 1. These findings highlight the superior recognition capability of the image-spectrum collaborative IRST system compared to conventional imaging-only methods.