The detection and tracking of weak and small targets in the aerospace background are of great significance to fields such as national defense security, space exploration, and unmanned aerial vehicle (UAV) supervision. In complex scenarios, the detection of infrared small targets is rather difficult due to factors such as the low grayscale value of the targets, small pixel size, and frequent background changes. Therefore, there is an urgent need to design an effective algorithm for detecting small targets. Meanwhile, it is also of great importance to develop a virtual simulation platform that integrates background modeling, energy enhancement, and detection algorithms.

To address the above issues, first of all, considering the limitations of anisotropic filtering in scenarios with a large amount of edge noise, an improved anisotropic filtering algorithm is proposed. By using the difference in image grayscale gradients to accurately represent background information, the accuracy and reliability of the background information can be improved. Then, aiming at the problem that common high-order or multi-frame accumulation leads to the simultaneous enhancement of the energy of both the target and the background, a motion estimation energy enhancement model is designed with the help of particle filter technology. A pipeline is constructed based on the inter-frame positions of the targets detected by the particle filter, and the targets are accumulated through the pipeline information to achieve the energy enhancement of the targets. Next, in order to filter out false targets, based on the energy accumulation, an algorithm based on multi-directional energy estimation is proposed. By taking advantage of the continuity of the targets and the randomness of the noise, the maximum value of the candidate targets is calculated and screened to accurately extract the target information. Finally, a virtual experimental platform for the space small target detection system is designed and developed based on Matlab's APP Designer. This platform has the excellent function of detecting and tracking weak and small targets in the aerospace background. The proposed weak and small target detection algorithm, as well as several classic small target detection algorithms, are introduced into the constructed platform. Users can flexibly select algorithms according to the characteristics of different aerospace scenarios, and can also set key parameters such as the starting serial number and the threshold value by themselves. With the help of this virtual simulation platform, users can intuitively perceive the impact differences of different methods on the detection results.

Through experimental comparisons of various algorithms in three different aerospace scenarios, the proposed algorithm demonstrates good detection effects on the developed virtual platform. In the detection experiments of sequential images, under the same τ value, it outperforms other algorithms. Its detection rate is greater than 80%, and the false alarm rate is less than 0.006.

This algorithm can meet the detection requirements for small targets. In addition, the developed virtual simulation experimental platform is conducive to users' parameter design and test analysis according to their own needs, making it convenient for users to perform a series of operations and understand the relevant detection principles.