Research on a Radio-Laser Composite Imaging Dataset Construction Method for Aircraft Target

To satisfy the requirements of radio-laser composite imaging and target recognition, this paper proposes a bimodal dataset construction method for radio-laser composite imaging, aiming to alleviate the lack of available datasets for bimodal target recognition research. A three-dimensional target geometric model and a missile-target encounter model is established to ensure consistent imaging geometry and motion conditions. Laser imaging and radio-frequency imaging methods are employed to generate bimodal images of the target under the same observation framework. By varying the target azimuth, pitch, and roll angles, bimodal images corresponding to different target attitudes are obtained, and a radio-laser composite imaging dataset is constructed. Simulation experiments are conducted on an aircraft target to verify the effectiveness of the proposed method. The results show that the laser images can clearly preserve the geometric contour and overall structural profile of the target, whereas the radio images can effectively characterize the scattering-center distribution features. Principal-axis direction analysis is further introduced to evaluate the consistency of the constructed bimodal data. The average directional error is approximately 3.15°, indicating that the generated dataset maintains good structural consistency between the two modalities. The proposed dataset provides an effective data foundation for subsequent studies on multimodal information fusion, composite imaging, and target recognition.