该算法将雷达波形通过时频分析转换为二维时频图像，利用图像处理算法提取时

The algorithm is a two-dimensional frequency analysis of converting video images, with the main component frequency information is image processing algorithm of FIG extraction, and the combined use of convolutional neural network support vector machine implementation feature extraction and recognition radar wave by the radar waveform; then, for LPI radar waveforms training sample number of small, hard training parameters CNN's deep problems, the idea of ​​learning to migrate into the radar waveform recognition algorithms to analyze the relationship and deep network support vector machine classifier model, proposed two transfer learning algorithm to identify the model (Inception-v3-SVM, ResNet-V2-152-SVM)

The algorithm converts radar waveform into two-dimensional time-frequency image through time-frequency analysis, extracts the main component information in the time-frequency map by image processing algorithm, and uses a convolution neural network to jointly support the characteristic extraction and recognition of radar waveform, and then, the number of training samples for LPI radar waveform is small. Deep CNN's parameter training difficult problem, the idea of migration learning into the radar waveform recognition algorithm, the analysis of the relationship between support vector machine and deep network model classifier, two migration learning algorithm recognition models (Inception-v3-SVM, ResNet-V2-V2-152-SVM)

The algorithm transforms radar waveform into two-dimensional time-frequency image by time-frequency analysis, extracts the principal component information of time-frequency image by image processing algorithm, and realizes the feature extraction and recognition of radar waveform by convolutional neural network and support vector machine. Then, aiming at the problem that LPI radar waveform training samples are few and deep CNN parameters training is difficult, the idea of transfer learning is put forward In the radar waveform recognition algorithm, the relationship between SVM and deep network model classifier is analyzed, and two migration learning algorithm recognition models (perception-v3-svm, resnet-v2-152-svm) are proposed<br>