| Because of its superior physical and optical properties,sapphire is widely used in substrate materials,consumer electronics,and military industries.During the growth of sapphire crystals,bubble defects are generated due to factors such as temperature and pulling speed.The presence of bubbles seriously affects the optical properties of the sapphire material and the sapphire ingot slice position.Therefore,the bubble detection in the sapphire ingot stage is an important link in the sapphire processing process.In the current production and application field,for the sapphire ingot bubble detection,the human eye detection method is mainly used,but the detection efficiency and detection accuracy are low,and the damage to the human eye will be caused.Supported by the project of Jinzhong City Science and Technology Bureau in Shanxi Province and the joint project of sapphire inoculation testing technology based on machine vision,this paper proposes a machine vision inspection method for sapphire ingots based on machine vision technology and the laser scattering effect between green laser and sapphire crystal.The main research contents are as follows:(1)According to the surface characteristics of the sapphire ingot and its detection requirements,the optimal wavelength range of the visual inspection light source and the imaging principle of the inspection method are studied.The selected laser source is used to irradiate the sapphire ingot in full field.The generated laser scattering effect is the main basis of the imaging principle of this detection method.(2)According to the above imaging principle,a sapphire ingot bubble machine vision inspection system was built.A series of pre-processing is performed on the detected image of the collected sapphire ingot sample,including image denoising,image enhancement,image segmentation,and extraction of bubble defect information.The visual effect of the preprocessed image is better,and the bubble target is prominent,which makes a good preparation for the subsequent identification and positioning of the bubble.(3)Based on the obtained pre-processed image,this paper uses edge pixel-based ROI detection method to define the pixels within the gray range of the bubble in the form of pixel set;According to the analysis of the positions of the bubble spots and the light source spots,the light source spots are removed,so as to realize the identification and positioning of the bubble spots.The experimental results show that the sapphire ingot bubble machine vision detection method proposed in this paper can identify bubbles and determine the location of the bubbles.The detection error rate is less than 10%,the average detection speed of a single sample is 0.3min /.Compared with the human eye detection method,it has obvious advantages and is a reliable,labor-saving and material-saving detection method. |
