An Efficient and Precise Parallel Implementation Scheme for 3D Stolt Interpolation

High-resolution radar imaging systems necessitate higher-order Taylor expansions of the range equation to achieve high-resolution imaging results, requiring enhanced processing and imaging efficiency from both the system and the algorithm. This paper proposes an efficient and precise three-dimensional Stolt interpolation parallel implementation scheme, which avoids lengthy and complex calculations. The nested interpolation scheme is implemented using GPU dynamic parallelism. To balance parallelism and complexity, the nested interpolation scheme is further optimized through multi-stream processing, reducing the demands on system performance and hardware resources while improving interpolation efficiency and maintaining high interpolation accuracy. This approach was validated through practical experiments using the GB-SAR imaging system, effectively demonstrating its practicality and reliability.