School of Artificial Intelligence, Xi'an Aeronautical Polytechnic Institute, Xi'an 710089, China.
School of Information Engineering, Chang'an University, Xi'an 710064, China.
Comput Intell Neurosci. 2022 Aug 21;2022:7316133. doi: 10.1155/2022/7316133. eCollection 2022.
The necessity of predicting and estimating river velocity motivates the development of a prediction method based on GAN image enhancement and multifeature fusion. In this method, in order to improve the image quality of river velocity, GAN network is used to enhance the image, so as to improve the integrity of image data set. In order to improve the accuracy of prediction, the image is extracted and fused with multiple features, and the extracted multiple features are taken as the input of CNN, so as to improve the prediction accuracy of convolution neural network. The results show that when the velocity is 0.25 m/s, 0.50 m/s, and 0.75 m/s, the accuracy of improved method can reach 85%, 90%, and 92%, which are higher than SVM, VGG-16, and BPNET algorithms. The above results indicate that the improvement has certain positive value and practical application value.
预测和估计河速的必要性促使人们开发了一种基于 GAN 图像增强和多特征融合的预测方法。在该方法中，为了提高河速图像的质量，使用 GAN 网络对图像进行增强，从而提高图像数据集的完整性。为了提高预测的准确性，提取并融合了多个特征，将提取的多个特征作为 CNN 的输入，从而提高卷积神经网络的预测精度。结果表明，当速度为 0.25 m/s、0.50 m/s 和 0.75 m/s 时，改进方法的准确率可达 85%、90%和 92%，高于 SVM、VGG-16 和 BPNET 算法。上述结果表明，改进具有一定的积极价值和实际应用价值。
