| Small-size curved lenses are widely used in various imaging systems.However,in the process of production,However,in the production process,how to control the quality and defect detection of the processed optical lenses is a technical problem that optical manufacturing enterprises need to solve urgently.At present,in industrial mass production,the detection of small defects of small-sized curved lenses is usually done by manual observation,which is difficult to observe,low in detection efficiency,and impossible to quantitatively analyze the measurement data.Based on the principle of transmissive stripe deflection,dark field imaging and backlight illumination,this paper designs and builds an automatic detection system to realize comprehensive,efficient and accurate detection of various defects of small size lenses.The specific research work lists as follows:(1)A comprehensive detection method combining transmissive stripe deflection,multi-angle side point source illumination and backlight illumination is proposed to realize image enhancement and acquisition of various lens defects,especially for the micro-defects,such as watermark defects and haze defects generated during processing.(2)Based on the image sharpness evaluation algorithm,the defect information is automatically focused and clearly collected,and the edge contour extraction algorithm is used to efficiently acquire the region of interest(ROI)of multiple sets of lenses,so as to realize multi-type defects efficient synchronous detection of lenses.Experiments have shown that for a lens with a diameter of 6-14 mm,a detection speed of 12 pieces / 17 s can be achieved.(3)Integrate phase solving algorithm,image enhancement algorithm and image segmentation algorithm for real-time processing of ROI.According to the image characteristics of each type of defect,an appropriate defect extraction and classification algorithm is designed,lens defect classification and size calculation are realized.(4)The construction of the small-scale curved lens comprehensive defect detection system is carried out,and the effectiveness of the proposed method on the detection of various fine defects of the lens is verified by experiments.Defect size inspection for different types of lenses with diameters from 6mm to 14 mm proves that the system defect recognition accuracy can reach 6μm,which meets the requirements of US military standard 40-20.The proposed detection method is applied to the experiment of glass molding process parameter adjustment,which proves that the detection method plays a good guiding role in optimizing the process parameters in the actual production process. |