基于模糊集的神经网络景象匹配算法

Fuzzy Sets-based Neural Network Algorithm for Scene Matching

将神经网络思想引入到景象匹配,提出了基于模糊集的神经网络景象匹配算法。该算法将图像模糊集作为特征空间,尝试了在模糊域中采用神经网络学习算法进行精确寻优。实验结果表明,设计的算法不但较好的满足了景象匹配系统对算法的性能要求,而且比传统算法具有更高的抗干扰能力。

Neural network theory is used for scene matching, and a novel neural network algorithm based on fuzzy sets is proposed. First, image fuzzy sets were taken as character space, and then the exact position was optimized by the 3 layer forward neural network algorithm. Results show that this method has merits in accuracy, speed, and robustness, and its anti-jamming performance is also better than some traditional scene matching algorithms.