Most learning-based methods previously used in image dehazing employ a supervised learning strategy,which is timeconsuming and requires a large-scale dataset.However,large-scale datasets are difcult to obtain.Here,we ...Most learning-based methods previously used in image dehazing employ a supervised learning strategy,which is timeconsuming and requires a large-scale dataset.However,large-scale datasets are difcult to obtain.Here,we propose a selfsupervised zero-shot dehazing network(SZDNet)based on dark channel prior,which uses a hazy image generated from the output dehazed image as a pseudo-label to supervise the optimization process of the network.Additionally,we use a novel multichannel quad-tree algorithm to estimate atmospheric light values,which is more accurate than previous methods.Furthermore,the sum of the cosine distance and the mean squared error between the pseudo-label and the input image is applied as a loss function to enhance the quality of the dehazed image.The most signifcant advantage of the SZDNet is that it does not require a large dataset for training before performing the dehazing task.Extensive testing shows promising performances of the proposed method in both qualitative and quantitative evaluations when compared with state-of-the-art methods.展开更多
This paper addresses the predefined-time bipartite tracking problem for second-order Multi-Agent Systems(MASs)with undirected signed topologies.A group of observers,which can estimate the state tracking errors for eac...This paper addresses the predefined-time bipartite tracking problem for second-order Multi-Agent Systems(MASs)with undirected signed topologies.A group of observers,which can estimate the state tracking errors for each follower in a pre-specified time,is proposed based on the time-varying function.In order to deal with the uncertainties caused by the unknown disturbances and the unknown input signal of the leader,we propose a predefined-time distributed control protocol based on the sliding mode control method.In addition,an auxiliary dynamic sliding variable is designed to reduce system chattering.Wetheoretically prove that the two control protocols can drive the state trajectories of each follower to reach the corresponding sliding surface within a specified time,and finally ensure that the prescribed-time bipartite tracking consensus is achieved for the MASs.Simulations are provided to verify the proposed schemes,and the simulation results further confirm the superiority of the adaptive control protocol.展开更多
基金supported in part by the National Natural Science Foundation of China(Grant No.61705127)Degree Construction Project of Detection Technology and Automation Devices,Shanghai University of Engineering Science(No.19XXK003)。
文摘Most learning-based methods previously used in image dehazing employ a supervised learning strategy,which is timeconsuming and requires a large-scale dataset.However,large-scale datasets are difcult to obtain.Here,we propose a selfsupervised zero-shot dehazing network(SZDNet)based on dark channel prior,which uses a hazy image generated from the output dehazed image as a pseudo-label to supervise the optimization process of the network.Additionally,we use a novel multichannel quad-tree algorithm to estimate atmospheric light values,which is more accurate than previous methods.Furthermore,the sum of the cosine distance and the mean squared error between the pseudo-label and the input image is applied as a loss function to enhance the quality of the dehazed image.The most signifcant advantage of the SZDNet is that it does not require a large dataset for training before performing the dehazing task.Extensive testing shows promising performances of the proposed method in both qualitative and quantitative evaluations when compared with state-of-the-art methods.
基金the National Natural Science Foundation of China[grant number 61705127].
文摘This paper addresses the predefined-time bipartite tracking problem for second-order Multi-Agent Systems(MASs)with undirected signed topologies.A group of observers,which can estimate the state tracking errors for each follower in a pre-specified time,is proposed based on the time-varying function.In order to deal with the uncertainties caused by the unknown disturbances and the unknown input signal of the leader,we propose a predefined-time distributed control protocol based on the sliding mode control method.In addition,an auxiliary dynamic sliding variable is designed to reduce system chattering.Wetheoretically prove that the two control protocols can drive the state trajectories of each follower to reach the corresponding sliding surface within a specified time,and finally ensure that the prescribed-time bipartite tracking consensus is achieved for the MASs.Simulations are provided to verify the proposed schemes,and the simulation results further confirm the superiority of the adaptive control protocol.