In order to solve the problem of classification and recovery of floating objects on the water surface

a classification and recognition model for floating objects on water surface is proposed based on deep convolutional neural network（DCNN）under the framework of deep learning.The CCD camera is used to collect the images of various floating objects

and the training and test datasets are established through the operation of translation and rotation

random clipping

affine transformation and adding noise.Through the automatic hierarchical learning of the low-level features to extract abstract high-level features

based on the extracted high-level features

the SoftMax classifier is used for training.The experimental results show that the classification accuracy of floating objects can reach 98.46 %

which verifies the feasibility and effectiveness of the model for the classification and recognition of floating objects on the water surface.