Roundness deviation evaluation method based on statistical analysis of local least square circles

Today, the quantitative evaluation of the quality of circular or cylindrical workpieces is becoming increasingly important for the relevant industrial production sectors. Although there are already some roundness deviation evaluation algorithms available to accomplish this task, these methods are always done in a holistic way. In many industrial scenarios, however, fine evaluation of the roundness variation of local segments is often more practical than the global assessment. By performing a fine evaluation of roundness variation of local segments, crucial information that can reveal intrinsic quality characteristics of both the workpiece and the production machine can be retrieved. However, this important issue has not been well studied. To deal with this problem, a roundness deviation evaluation method based on statistical analysis of local least square circles was proposed. Experimental results illustrated that the proposed method can stably and reliably evaluate the local and global roundness deviations effectively.