提出了一种利用Ｈｏｐｆｉｅｌｄ神经网络进行交叉线识别的算法。在本算法中，一个含有交叉线的图象被映射到Ｈｏｐｆｉｅｌｄ神经网络之上，每一个象素与一个神经元相对应，神经元间的权重由其对应的象素间的关系（包括距离、斜率等）所决定。在Ｈｏｐｆｉｅｌｄ神经网络的收敛过程中，神经元的状态根据一定的规则不断地被调整直到收敛；同时随着神经元状态的确定，交叉线段也被识别出来。在实验模拟中，此算法显示了优越的性能。 A new algorithm of Hopfield neural network based on Hopfield neural network is proposed. In this algorithm, an image with crossed lines is mapped onto the Hopfield neural network. Each pixel corresponds to a neuron. The weight between neurons is determined by the relationship between the corresponding pixels (including the distance, Slope, etc.). During the convergence process of Hopfield neural network, the state of neurons is continuously adjusted according to certain rules until convergent. At the same time, as the state of neurons is determined, the intersecting line segments are also identified. In experimental simulation, this algorithm shows superior performance.