Search

Morphological filters are widely used in data pre-processing, whereas the existing morphological filters are not optimal in terms of performance, since they are designed by combining the basic morphological operators together artificially. Hence, we proposed an automatic morphological neural network design method based on differentiable architecture search (DARTS), which is a neural architecture search (NAS) method. Firstly, the mechanism is revealed that the morphological dilation with flat structural elements behaves the same as Maxpooling without down-sampling. Secondly, morphological dilation and erosion neural network layers were defined. Thirdly, using DARTS and the search space with morphological neural network layers, end-to-end morphological neural network was implemented. Finally, the morphological neural network was successfully applied in bearing fault diagnosis. Experiments on the three datasets show that the proposed method improved the feature extraction capabilities of neural networks, then achieved end-to-end bearing fault diagnosis, and the diagnosis accuracy was improved by more than 2.0%. • For the first time, the morphological operator is embedded in the neural network, which improves the signal feature extraction and suppresses the background noise. • For the first time, the DARTS algorithm is applied to search for the optimal morphological operator combination, which avoids the dependence of artificial experience and improves efficiency. • The morphological neural network obtained based on DARTS was applied in bearing fault diagnosis, and improved the diagnosis accuracy. [ABSTRACT FROM AUTHOR]

In order to obtain high precision in camera calibration, a target should be large enough to cover the whole field of view (FOV). For cameras with large FOV, using a small target will seriously reduce the precision of calibration. However, using a large target causes many difficulties in making, carrying and employing the large target. In order to solve this problem, a calibration method based on the virtual large planar target (VLPT), which is virtually constructed with multiple small targets (STs), is proposed for cameras with large FOV. In the VLPT-based calibration method, first, the positions and directions of STs are changed several times to obtain a number of calibration images. Secondly, the VLPT of each calibration image is created by finding the virtual point corresponding to the feature points of the STs. Finally, intrinsic and extrinsic parameters of the camera are calculated by using the VLPTs. Experiment results show that the proposed method can not only achieve the similar calibration precision as those employing a large target, but also have good stability in the whole measurement area. Thus, the difficulties to accurately calibrate cameras with large FOV can be perfectly tackled by the proposed method with good operability. [ABSTRACT FROM AUTHOR]

The accuracy of camera calibration is of great importance to vision measurements. Target-based calibration methods should cover the whole field of view (FOV), which leads to complex operation in large FOVs and rely on personal experiences. To overcome the difficulty in accurate calibration, a calibration method based on reinforcement learning is proposed. First, a Markov decision process (MDP) model of the calibration procedure is established. Then, the reward function is designed, combining the requirement of calibration accuracy and the state-space constraint. Finally, the optimized target locations and poses are obtained by continuous interaction with the calibration environment based on Q-learning, which plays a key guiding role in camera calibration. Simulation experiments and real experiments are performed, which indicate that the proposed method effectively improves the success rate of large FOV camera calibration, and solves the problems of relying on personal experiences, low calibration accuracy, and poor stability caused by target placement in the camera calibration process. • Reinforcement learning theory is innovatively applied to optimize the target locations and poses which are the crucial factors that determine the calibration accuracy of the large field of view (FOV) cameras. • The target placement in large FOV camera calibration is modeled as a Markov Decision Process (MDP) for the first time. • An improved Q-learning based method is proposed to solve the MDP model to obtain the optimized target locations and poses. The proposed method avoids the reliance upon personal experiences, greatly improves the success rate and the accuracy of camera calibration. [ABSTRACT FROM AUTHOR]