Research On Image Deblurring Algorithm Based On Sapphire Surface Defects

In recent years, due to the excellent properties in physics, chemistry and optics, sapphire has been widely used in many fields such as industry, manufacturing, and so on. It is the high hardness, high melting point, not soluble in water, wear, corrosion and good light transmission that make the sapphire can be used as the screens of mobile phones, television and other daily household appliances, and even will replace the incandescent lamp which now in everyone’s life. However, the application market of the sapphire is still not able to expand to the civilian population, this because, the ordinary people can’t afford to the high costs from its production and processing. This caused by the special characteristics of the sapphire(high hardness, etc.), and thus result in some of sapphire getting some certain defects more or less.In order to improve the qualified rate of the sapphire, not only the improvement in the production and processing technologies, but also the quality detection technologies are needed. The researches on sapphire quality detection technologies is helpful for substandard products elimination, and what more important is that it can improve the processing technology and the production of high quality sapphire products.Currently, the sapphire detection methods are limited in two main basic categories: one is detecting through the artificial eyes under some certain lights, which has many shortcomings such as high negative rate, low efficiency and eye damages; the other one is using different types of microscopes for different types of defects, then classifying the defects and record them manually, this method followed by some disadvantages like low efficiency and high technical requirements for the operators. Therefore, it is urgent to make the sapphire defect technologies be automatic and efficient.About the sapphire defects detection model, what this paper mainly do are listed as follow: 1) The paper presents a novel laser based detection model for detecting the sapphire defects, then qualitatively analyzes the defects classifications, and quantitatively deals with the severity of the defects with some hot spots such as image processing, machine learning, pattern recognition and so on, lastly determines the quality of the sapphire. 2) This paper makes a detailed analysis on the feasibility of the model proposed. 3) Based on the characteristics of the sapphire defects image, this paper comes up with an effective algorithm to restore the two kinds of predictable and high incidence motion blurred images during processing the feasibility verification, they are linear motion blurred images and out-of-focus blurred images.The novel image deblurring algorithm we propose has the following innovative ideas : 1) This algorithm works under the improved cepstrum, bit plane pixel separation and Radon projection transform, it makes the PSF parameters extraction more accurate. 2) This algorithm includes an important statistical iteration based inverse filter, it helps restoring the blurred image with high quality. 3) The algorithm we proposed has the advantages of low complexity, high efficiency, high accuracy, good stability and strong expansibility. It is suitable for the industrial application image preprocessing such as the defect detection of sapphire, and provides high quality source images for further defect detection.This paper makes a lot of experiments and data tests, they all show that, this algorithm can efficiently recover sapphire defect blurred images and eliminate the ringing effect during the restoration. These prove this algorithm a simple and efficient one. In order to demonstrate the high stability of the algorithm we proposed, an analys is on its performance is made at the end, it includes some comparisons with other similar algorithms and the research on anti noise performances of the algorithm.