Outlier removal method for the refinement of optically measured displacement field based on critical factor least squares and subdomain division

Abstract

As a representative type of outlier, the abnormal data in displacement measurement often inevitably occurred in full-field optical metrology and significantly affected the further evaluation, especially when calculating the strain field by differencing the displacement. In this study, an outlier removal method is proposed which can recognize and remove the abnormal data in optically measured displacement field. A iterative critical factor least squares algorithm (CFLS) is developed which distinguishes the distance between the data points and the least square plane to identify the outliers. A successive boundary point algorithm is proposed to divide the measurement domain to improve the applicability and effectiveness of the CFLS algorithm. The feasibility and precision of the proposed method are discussed in detail through simulations and experiments. Results show that the outliers are reliably recognized and the precision of the strain estimation is highly improved by using these methods.

A rapid outliers detection and correction method for external ballistic velocity measurement data

“…Then the B-spline is modeled by equation ( 5), and the estimated values of the first-degree, second-degree, and third-degree B-spline fitting coefficients are calculated by equation ( 6) corresponding as d (1) , d (2) , and d (3) , respectively. The coefficients 2) , and d (3) are substituted into equation (7) with the next time t p+L , respectively, so the estimated values of y p+L are ŷ (1) p+L , ŷ (2) p+L , and…”

“…Step 3: The coefficients d(1) , d (2) , and d (3) are substituted into equation (7) with the time {tp, t p+1 , . .…”

“…Scholars have done much research on handling outliers, and there were many practical algorithms [5][6][7][8]. In [9], the outlier detection was based on the median relative height of the surrounding data within a defined detection window.…”

Cross-medium imaging model and calibration method based on refractive optical path for underwater morphology measurement

“…In addition, we calculated the distance from each measured point to the center of the fitting sphere, and made a difference with the radius of the fitting sphere to obtain the deviation of each data point. In order to remove the influence of noise points on the measurement results, we used the outlier removal method [36] and selected points with a deviation range of [−0.05, 0.05] to ensure that the number of remaining points still accounts for more than 95% of the total number of points. The measurement results are shown in table 4.…”

Experiments and Modelling of Composite–Aluminium Bolted Joints

“…Other authors have made efforts to reduce the noise from DIC measurements and our method of fitting measurement data to the equation of a surface is similar to the pointwise least squares (PLS) method by Pan et al [40]. Also Wang et al [41] reduce measurement noise by outlier removal based on a critical distance between a data point and a least squares plane fitted to measurement data. Smoothing element analysis (SEA), introduced by Tessler et al [42], is another technique for smoothing noisy displacement fields.…”