Abstract: In this paper, an isotropic resolution down-looking scene matching algorithm based on SNN kernel method has been proposed. By simulating the charge attract model, the SNN kernel function is presented to compute similarity of high dimensional data with unequal dimensions. By mapping image character points to radial basis vector (RBV) space, the transition matrix and similarity between two character point sets are constructed by SNN kernel function. Finally, we use a permutation testing module to enforce the stability of SNN kernel to ensure the output reliability. The test results show that the proposed algorithm has high accuracy, high real-time performance, and strong robustness against the image aberrance, noise, and signal concealment.