Byeong-Mook Chung

Development of real-time vision-based fabric inspection system

Quality inspection of textile fabric products is an important problem for fabric manufacturers. This paper presents an automatic vision-based system for the quality control of web textile fabrics. Typical web material is 1-3 m wide and is driven with speeds ranging from 20 to 200 m/min. At present, the quality inspection process is manually performed by experts. However, they cannot detect more than 60% of the overall defects for the fabric if it is moving faster than 30 m/min. To increase the quality and homogeneity of fabrics, an automated visual inspection system is needed for better productivity. Currently, the existing inspection systems are too expensive for small companies. In this paper, a PC-based real-time inspection system is proposed with benefits of low cost and high detection rate. The proposed algorithm showed good results for several types of fabric defects.

Active tension control of high speed splitting machines using fuzzy PID

In the textile manufacturing process, it is very important to uniformly maintain the thickness of yarn, because the quality of texture is affected by the dint of the quality of yarn. However, the production of chemical yarn is done using more than 10 strands because of productivity. The united yarn has to be split into independent strands before making textures. In this process, most splitting machines have too heavy and varied tension because they use a passive brake system to control the tension of the yarn. This article proposes a splitting machine with an active tension controller to uniformly maintain the tension of the fine yarn. In real experiments, to measure the tension of the yarn, a tension sensor is developed using a potentiometer and a tension bar. However, because it is difficult to use constant PID gains for the tension control, the fuzzy logic based PID controller is proposed.

Fuzzy logic modeling and control for drilling of composite laminates

In drilling of composite laminates, it is important to minimize or reduce occurrences of delaminations. In particular, a peel-up delamination at entrance and push-out delamination at exit are common. Delaminations may be avoided by regulating the drill thrust force which can be controlled by adjusting the feedrate of the drill. Dynamics involved in drilling of composite laminates is time varying and nonlinear. In this paper, a fuzzy logic model and control strategy are proposed. Simulation results show that the fuzzy model can describe the nonlinear time-varying process well. The fuzzy controller realizes a fast rise time and a little overshoot of drilling force.

Control of Nonlinear Multivariable Systems Using Direct Fuzzy Learning Method

A learning method that can acquire the control rules in the multi-input multi-output system is presented. Generally, in the control of a multivariable system, the error of the outputs must be transformed into the corrections of the inputs to get the control rules. If we do not know the mathematical modeling equation of the system, we cannot find out the Jacobian matrix to show the output variance with respect to the input variance. However, because it is not necessary to use the exact variable matrix in the learning of the fuzzy rules, this article introduces a method to get a representative constant Jacobian matrix that assures the convergence. Using this constant matrix, the fuzzy rules are successfully obtained. In particular, because many fuzzy subsets must be used in the case of a multivariable system, the optimization technique to minimize the fuzzy rule structure is also applied.

Color inspection of printed texture using scanner and NN model

It is very important to inspect color of printed texture in the textile process. Because a scanner is less sensitive to environmental light in view of color distinction, it is more effective than CCD camera in the color inspection. To distinguish the color of the printed texture, RGB color values obtained from a scanner must be transformed to the standard colorimetric system used in the textile industry. It is XYZ color system that is defined by CIE (Commission Internationale de lEclairage). The mapping from RGB to XYZ color values is not simple and the scanner has even a positional deviation of RGB colors. In this paper an automatic color inspection method using a general scanning machine is presented. We used a NN (neural network) model to map RGB to XYZ and compensate the positional error. In the real experiments, this inspection system shows to get very exact XYZ values from the traditional scanner regardless of the measuring position.

A tuning method of multi variable PID gains using Jacobian

Generally, a PID controller is not suitable to control a multi variable system because it is very difficult to tune their gains. However, this paper shows that it is not hard to tune the PID gains if we can find a Jacobian matrix of the system. Because the Jacobian matrix expresses the ratio of output variations according to input variations, it is possible to adjust the input values in order to reduce the output error. If the cost function is composed of error related terms, the gradient approach can tune the PID gains to minimize the function. In simulation, a hydrofoil catamaran with two inputs and two outputs is considered as a multi variable system. We can easily obtain the multi variable PID controller by the proposed tuning method. When it was compared with a LQR controller with a modelling equation, its performance was very good.

Knowledge-based control for RGB convergence in deflection yoke of CDT

The color display tube used in computer monitors consists of many components. The deflection yoke (DY) among them supplies the vertical and horizontal magnetic fields so that the spatial trajectories of electron beams can be deflected according to the synchronization signals. If the magnetic fields are not correctly formed, color blurring or blooming where the color image on screen may not be clear will occur. It is misconvergence. In the manufacture of the DY, its quality is strictly controlled to achieve the desired convergence. If the misconvergence happens it has to be cured by sticking the ferrite sheets on the inner part of DY. However, the adjustment is not easy because this process requires an experts knowledge and experience. A knowledge-based intelligent controller using a neuro-fuzzy model will be introduced in this paper.

Phase-height relationship by plane analysis in 3D shape measurement using fringe pattern projector

In a three-dimensional(3-D) measurement system based on a digital light processing (DLP) projector and a camera, a height estimating function is proposed based on geometric analysis. The proposed 3-D shape measurement method is a hybrid method that combines the geometric parameter measuring method and the least squares method. This method uses the phase-to-height relationship for one line by plane analysis, and the related parameters are estimated using the least squares method. Sinusoidal fringe patterns of the projector are projected on the object, and the phase of the measuring point is calculated from the camera image. Then, the relationship between the phase by fringe patterns and the height of the measuring point is described as a parameter of the horizontal coordinate on the image plane. Thus, the 3-D shape of the object can be obtained. Our experiments show that the modeling errors for the given x-z plane are within ±0.1 mm. The model for the entire working space can be represented by several plane models because all the x-z plane models along the y-axis are nearly the same. Therefore, the proposed method can dramatically reduce the number of mapping functions needed for 3-D measurement.