Haiyong Chen

Steel Surface Defect Detection Using a New Haar–Weibull-Variance Model in Unsupervised Manner

Automatic defect detection on the steel surface is a challenging task in computer vision, owing to miscellaneous patterns of the defects, low contrast between the defect and background, the existence of pseudo defects, and so on. In this paper, a new Haar-Weibull-variance (HWV) model is proposed for steel surface defect detection in an unsupervised manner. First, an anisotropic diffusion model is utilized to eliminate the influence of pseudodefects. Second, a new HWV model is established to characterize the texture distribution of each local patch in the image. The proposed model can project the texture distribution of each patch into the low-dimensional space with only two parameters. The parameter distribution of the whole image can also be unified into the form of linear radiation in an Euclidean space. The reliable background can be extracted via the formation of parameter distribution, by which the model parameter can be optimized further. Finally, the adaptive threshold can be determined to distinguish the defect from the background. Experimental results show that the proposed method can detect an arbitrary type of defect on the homogeneously textured surface and achieve an average detection rate of 96.2% on the data set, which outperforms the previous methods.

Trajectory tracking control of a Macro-Micro welding robot based on the vision navigation

A trajectory tracking control scheme of a Macro-Micro welding robot based on the vision navigation is proposed. The robot system is composed of macro manipulator and micro manipulator attached to the tip of the macro part. Based on the visual sense, the robot is capable of effectively realizing the teaching mode and real-time tracking mode simultaneously, which is quite advantage to achieve better seam tracking performance. When using the teaching mode, several points sampled by the visual system can be used to build seam model that can be regarded as a standard seam curve of trajectory plan motion of the macro part. The correction bias control of the micro manipulator is to compensate the seam errors between standard seam position and real measures. The coordinated motion control is adopted to keep the joint values of the micro manipulator in desired range. Experiments results show the effectiveness of these control schemes and the good tracking results are obtained.

Networked H ∞ synchronization of bilateral teleoperation systems

This brief is mainly concerned with the networked H∞ synchronization problem of master-slave bilateral teleoperation systems with nonuniform sampling. The networked-induced delays and packet dropouts are also considered under a unified framework. The sufficient condition in terms of linear matrix inequality (LMI) is provided to guarantee H∞ performance for the error dynamics. A simple optimization algorithm is given to obtain the smaller attenuation level γ. Lastly, a numerical example illustrates the validity of the proposed method.

Visual feature extraction and tracking control for narrow seam in three dimension space

An algorithm on adaptive visual features extraction for three dimension micro narrow weld seam is proposed. And the reliable extraction of the seam is realized. Then a morphological profile recognition method and fast straight line fitting method are proposed to robustly extract the skeleton of structured light stripe. Furthermore, the orthogonal characteristics of the weld features and the structure light skeleton is utilized to complete the decoupling calibration. Thus, the high precision position on three-dimensional Cartesian space is reached by only using two-dimensional information in image space. Finally, satisfactory results are obtained.

Visual servo control system for narrow butt seam

In this paper a visual servo control system is designed to realize the accurate tracking of the narrow butt seam of the thin plate. An effective “and” operation and statics method is proposed to reduce the image noises caused by various disturbances in the welding process. And the ROI of the image is determined to improve the efficiency of the image processing. The feature line of the thinned seam is extracted by using combination of hough transform and least square line fitting technique. Moreover the method used in this study ensured the safety and reliability. In terms of controller, to guarantee the reliability, image error filtering and output pulse verification is added to the controller. Finally, a series of experiments were conducted to verify the good performance of the proposed seam tracking system.

Dispatching algorithm design for elevator group control system with Q-learning based on a recurrent neural network

A dispatching algorithm of elevator group control system is proposed based on reinforcement learning. Elevator dispatching is modeled by a Markov Decision Process. Then an internally recurrent neural network based reinforcement learning method is designed to find the optimal dispatching policy while the state-action value function is iteratively approximated. Finally, several simulated experiments are done to compare the trained dispatching policy with other traditional ones. The experimental results demonstrate the effectiveness of proposed dispatching method.

A three dimension reconstruction method on a kind of micro and thin laser seam

The laser weld is becoming more and more popular in the steel industrial production. However, the varying illumination, reflection and splatter lead to the irregular seam shape, which deteriorates the seam quality. In order to evaluate the seam shape and its quality, a 3D reconstruction method about micro and thin laser seam is proposed. The line structured light stripe is projected on the laser seam to be measured by a projector, and deformation of the stripe is captured by a CCD camera with industrial microscope lens. An image processing method that can efficiently locate the deformation of the stripe in the image plane is presented. Also, a novel procedure to automatically define the region of interest in the image is proposed. And then a straight line and curve fit is used to reduce the harm of the various disturbances and accurately gain the centre line of stripe. Furthermore, the characteristic points of the seam are obtained by using the distance search method. Finally, the proposed reconstruction method is applied to laser seam specimen, the desired performances are gained, and the results are satisfying.

A novel learning control method of macro-micro robot

A novel control scheme on learning frame is proposed for control of uncertain robot manipulator performing periodic tasks. In the paper, the uncertainties in the robot system may follow non-Gaussian distribution or Gaussian distribution, but the premise condition is that they are bounded. The nonlinear relation formulation between controller gain and tracking trajectory error of robot in each batch has been given. The optimal process of controller gain has been presented by using the proposed ILC frame. In the end, the convergence condition is analyzed and the experimental result is given. It is shown that all the error signals in the learning control system are bounded and the robot motion trajectory converges to the desired one asymptotically. Experimental results demonstrate the feasibility and effectiveness of the proposed control method for repetive robot motion.