低对比度场景下冰水特征与冰封率智能识别研究OA

Ice hazards are prevalent in rivers,canals,and reservoirs in cold regions.As an important parameter influencing water heat loss and ice formation,ice concentration is used for evaluating ice hazards.Its efficient moni-toring and accurate identification are crucial for preventing ice floods.Compared with natural rivers,water convey-ance projects exhibit less variation in boundary and hydrodynamic conditions,as well as superior water quality.This makes it challenging to distinguish between ice and water,leading to greater errors in image-based ice concentration recognition methods.To address this challenge,an intelligent ice concentration recognition algorithm based on a deformable convolutional neural network is proposed.This algorithm incorporates deformable convolutional layers,which can adaptively adjust the sampling positions of convolutional kernels to achieve more accurate capture of com-plex ice and water features in low-contrast scenarios.A floating ice dataset containing 330 images from the Middle Route of South-to-North Water Diversion Project was constructed,and a five-fold cross-validation method was used to optimize the algorithm parameters.Experimental results of 16 typical floating ice images show that the average accuracy(ACC)of ice concentration identification reaches 0.96,while the mean intersection over union(IoU)achieves 0.91.Compared with commonly used ice concentration recognition algorithms such as the Otsu and SVM,the proposed algorithm improves the average ACC by 16％and 10％,and the average IoU by 19％and 9％,respec-tively.The findings of this study provide an alternative method for ice concentration recognition in water conveyance channels.