A vision-based processing methodology for profile grinding of contour surfaces

Abstract

On-machine direct detection of profile errors is vital to improve accuracy and efficiency in profile grinding. However, achieving such detection processes is difficult because of harsh machining conditions. This study presents a novel machine-vision-based processing methodology for the profile grinding of contour surfaces instead of the traditional optical-enlargement-based profile grinding which is manual dependent and low efficient. Grinding errors were efficiently detected online through machine vision. A specific vision system was coordinately designed with the profile grinding system to ensure distortionless measurement of workpiece contour and overcome the interferences of machining environment during profile grinding. A machining error detection principle was proposed based on the online captured workpiece contour image. Real-time error identification and compensation algorithms were developed through the synthetic error measurement. Simulations and experiments were conducted successively. The results indicated that profile errors were considerably reduced and measurement efficiency was improved, validating the effectiveness of the proposed methodology for profile grinding of contour surface. The findings can also provide a reference for the direct measurement of machining errors in other machines.