Retraction Note: Discussion on ecological design and construction of river landscape based on fine-grained image classification

The fine-grained image classification method does not have a very good training effect when performing random weight for neural networks. For the under condition of convolutional network, main reason is that its upgrading can better improve stability network. Fine-grained through bilinear networks to effectively merge several networks, so as obtain high-resolution advantage classification. With development society and improvement urban construction, artificial reconstruction rivers in city will strengthen connectivity smoothness reconstructed river certain extent, but it also affect self-healing river. At present, treatment pollution major aspect country’s environmental pollution. Obtain transformation restoration ecological environment around data exchange other methods. Establishing scientific complete plant community with naturalized revetment system reduce impact water on riverbank. same time, these plants play role flood prevention discharge, polluted quality give purification. While carrying out protection river, pays attention design planning landscape, withstand flowing while having beautiful landscape design, realize civilization construction landscape.