介绍了一种基于中分辨雷达回波序列的目标轮廓成像识别方法。该方法实现了一维回波序列到目标二维径向轮廓像的变换。在二维轮廓像基础上，可以提取到较一维回波更加有效、稳定的特征。选择ＡＲＴ２神经网络模型对这些特征进行分类，得到了平均９１％的正确识别率。 A method of target contour recognition based on mid-resolution radar echo sequence is introduced. The method realizes the transformation from one-dimensional echo sequence to the target two-dimensional radial contour image. Based on the two-dimensional contour images, more efficient and stable features than the one-dimensional echo can be extracted. The ART2 neural network model was chosen to classify these features, resulting in an average recognition rate of 91%.