Display Omitted A surface defect detection system is proposed for mobile phone screen glass. This system achieves 94% in sensitivity and 97.33% in specificity.The proposed system takes approximate 1.6601s to detect a MPSG. The detection accuracy and speed can meet the needs of online detection for MPSG.Compared with other methods used in the experiment, the proposed improved fuzzy c-means can segment the surface defects in MPSG more accurately. Defect detection using machine vision technology plays an important role in the manufacturing process of mobile phone screen glass (MPSG). This study proposes an improved detection algorithm for MPSG defect recognition and segmentation. Considering the problem of MPSG image misalignment caused by vibrations in the mobile stages, a contour-based registration (CR) method is used to generate the template image used to align the MPSG images. Based on this registration result, the combination of subtraction and projection (CSP) is used to identify defects on the MPSG image, which can eliminate the influence of fluctuation in ambient illumination. To segment the defects with a fuzzy grey boundary from a noisy MPSG image, an improved fuzzy c-means cluster (IFCM) algorithm is developed in this study. A defect detection system is developed, and the proposed algorithms are validated using a number of experimental tests on MPSG images. The testing results demonstrate that the approach proposed in this study can effectively detect various defects on MPSG and that it has better performance than other methods.

Automatic surface defect detection for mobile phone screen glass based on machine vision

The emergency of surface defect would significantly influence the quality of MPCG (Mobile Phone Cover Glass) Therefore, efficient defect detection is highly required in the manufacturing process Focusing on the problem, an automatic detection system is developed in this paper The system adopts backlight imaging technology to improve the signal to noise ration and imaging effect Then, a modified segmentation method is presented for defect extraction and measurement based on deep neural networks In the method, a novel data generation process is provided, with which the drawback that huge amount of data is required for training deep structured networks can be overcome Finally, experiments are well conducted to verify that satisfactory performance is achieved with the proposed method

Vision-Based Defect Detection for Mobile Phone Cover Glass using Deep Neural Networks

The modelling of volumetric errors of machine tools has been widely used by the method of Homogeneous Transformation Matrix (HTM) based on rigid body kinematics analysis for a long time. Without spindle induced errors and thermal errors, to establish a closed-loop HTM for a three-axis machine tool, the total of 21 geometric errors are the primary elements to be known. It is well known that the measured points of translational errors are directly related to the volumetric error at the tool cutting point through rigid body kinematics. In the generalized HTM method, however, this relationship is missing. This report, therefore, proposes a new comprehensive approach to formulate the volumetric errors based on the famous Abbe principle in order to derive the error term in the motion direction, and Bryan principle to derive the error terms in orthogonal to the motion direction. The proposed methodology is simple in concept, rational in physical meaning and easy in implementation. Experimental results, including faster multi-degree-of-freedom error measurements, volumetric error analysis and compensation on a testbed of small machine tool, validate the correctness and effectiveness of this method.

A novel modeling of volumetric errors of three-axis machine tools based on Abbe and Bryan principles

Abbe error is the inherent systematic error in linear displacement measurement due to the measuring axis being out of line with the moving axis. The resulting gap is called the Abbe offset, which w...

A novel method of angular positioning error analysis of rotary stages based on the Abbe principle

This paper presents a robust and low-cost roll angle measurement system (RAMS) on the basis of two parallel beams in association with two position detectors. The commonly occurring influences of beam drift, beam diameter, and intensity variations, and non-parallelism of dual-beam on the roll angular error measurement of precision linear stages are thoroughly considered and reduced. The effectiveness of the designed system and proposed methods were demonstrated by a series of experiments. It has been verified that the designed system’s measurement accuracy is within ± 1.2 arcsec over a measurement range of 1 m. The designed system is easy to construct both in the laboratory environment and factory field.

Robust roll angular error measurement system for precision machines.

Levels are fundamental metrological tools for machine leveling and angle measurements. Current levels are all based on the pendulum principle and bubble vial principle. The former is precise but complicated, while the latter is not precise enough. This paper presents a novel, simple and precise optoelectronic level. It is based on the principle of light refraction in the transparent and viscous fluid. The fluid surface is always in leveling condition regardless whether the ground is level or not. Measuring the refraction angle with precise dual-axis autocollimator can directly reflect the inclined angle of the level. A leakage prevention design is also proposed to allow the level to be placed in any pose. Calibrated by an angular interferometer, the error of the dual-axis optoelectronic level is better than ?0.7 arcsec in the measuring range of ?100 arcsec, and better than ?0.4 arcsec for the range of ?30 arcsec, for both axes. Without the need of a pendulum mechanism, this is a simpler design for precision level.

A novel dual-axis optoelectronic level with refraction principle

3-D measurement and evaluation of surface texture produced by scraping process

Smart mobile phone is the most popular consumer product nowadays. The quality of the mobile phone panel is important as it gives the first image to potential buyers. Many mobile phone glass panels have the black thin film coating surround the boundary. Some surface defects and uneven coating thickness will degrade the appearance quality of the mobile phone. This research aims to develop a multi-function automated optical inspection system for the mobile phone panels in the production line.Different from some commercial systems that can only identify the 2D surface defects or the 3D film thickness, the developed system is capable of detecting both the 2D defects on the whole panel and the uniformity of 3D film thickness of the glass coating. The 2D surface defects are inspected by the image processing system that can not only extract the defects through the technique of pattern recognition but also classify the defect types by the technique of morphology. The film thickness is measured by a focusing probe utilizing the DVD pickup head as a low cost sensor.Experimental results show that five types of surface defects, including the hole, oil stain, scratch, stripe and edge, can be classified with 100% success rate. The focus probe processed by normalized focus error signal (NFES) technique can achieve both the measurement accuracy and standard deviation less than 0.1 μm.

Development of an Automated Optical Inspection System for Mobile Phone Panels

DVD player is regarded as a mature and low-cost commercial product for multimedia and computer data storage. Fundamentally the pickup head is a key part of the DVD player and it uses focused laser beam to read the data on the DVD disk tracks whose width is approximately 0.74μm. This research focused on the control aspect applied to a DVD pickup head for the development of an auto-focusing laser probe. Practically, the laser pickup head has the S-curve linearity property with the measuring range of about 7μm [3]. To extend the measurement range, the feedback control of VCM in the focusing direction is needed. This study used Genetic Algorithms (GAs) to search for appropriate PID gains for VCM control. In MATLAB simulation, the GAs tuning scheme could produce the optimised gains within less than 0.02 sec. Moreover, the experimental results demonstrated that the VCM auto-focusing measurement range could be extended to 1350μm.

Jingsyan Torng

Papers

A novel dual-axis optoelectronic level with refraction principle

3-D measurement and evaluation of surface texture produced by scraping process

Development of an Automated Optical Inspection System for Mobile Phone Panels

A Genetic Approach on the PID Control of VCM for Auto-focusing Laser Probe