The Online Visual Inspection System For Ceramic Parts Defects

Ceramics are indispensable materials in life and production,defects of ceramics will seriously affect the quality of ceramics,ceramic ring and square ceramic tubes due to small size,large quantities bring more difficult to detect in the production process.Traditional ceramic parts detection relies on artificial eye detection,which there is low detection efficiency,poor stability,prone to the phenomenon of false detection and missed.In recent years,non-destructive testing methods based on sound and light,etc.caused great concern in the ceramic defect detection.In this paper,machine vision approach is used,for detecting defects in a ceramic ring or square ceramic tube,discussion on a ceramic parts online visual defect detection method,the main contents include: 1)Design of online detection system,the main contents include: image acquisition unit,motion control unit,image processing unit,human computer interaction unit,etc.By combining the relevant theoretical knowledge of mechanical,electrical,optics,machine vision and practical application,design and realization the machine vision inspection system of ceramic parts defects,can detection the ceramic ring and square ceramic tube defects.2)Research of ceramic ring detection algorithm,use the projection method + Hough transform to detecting circular ceramic,according to the circular geometric feature,Two-dimensional scanning is applied in image by the projection method,which can calculate circular center coordinates,according to Hough Transform the algorithm can solve the radius.3)Research of square ceramic tube detection algorithm,use projection method + local patch to study detection method for square ceramic tube,use the region of interest to obtain an image of square ceramic tube,adjust the inclination of square ceramic tube through the center of gravity weighting,square ceramic tube is divided into multiple blocks by local patch,analyze characteristics of each block to detection defect.Experimental results show,the method using projection + Hough transform can facilitate detection concentric circles,can convert the three-dimensional accumulation in Hough transform to one-dimensional accumulation,reduce time-consuming of algorithms.The method using the projection + local patch simplifies the testing process of square ceramic tube,and ensure the correct detection rate.Experimental data show,ceramic parts defect detection algorithm takes 83 ms,which can reach 10 Hz detection speed requirements in ceramic parts.