Instrumentation & Measurement Magazine 25-5 - 17

An Adaptive Sampling Strategy for
Surface Measurement of Complex
Curved Workpieces
Yi Yang, Hairong Fang, Kai Huo, and Yufei Chen
L
arge complex workpieces are the core components
in aerospace, shipbuilding, transportation and
other fields. Such components usually have large
size, complex shape characteristics and poor stiffness of
workpieces, such as large common bottom components of
launch vehicles, rocket panels, aircraft radomes, etc. When
machining such high-performance parts, it is usually necessary
to copy the parts according to the task requirements.
The integration technology of complex surface measurement
and machining is the organic combination of digital
measurement technology and machining technology and
is an effective means to ensure the high-precision copying
machining of large complex workpieces. Data acquisition
technology is the most critical part. For the data acquisition
of some large and complex workpieces, laser scanning measurement
technology with high reliability and local accuracy
is usually used. In the process of data acquisition, the measuring
track and the measuring point interval are planned
according to the structural characteristics of the workpiece.
The quality of the measurement data will directly affect the
accuracy and overall quality of the follow-up profiling processing.
The two most important factors in curved surface
measurement are the scale of sampling quantity and the distribution
of sampling points [1]. In the point cloud model,
the larger the number of sampling points, the more appearance
features are included, and the more accurate the 3D
reconstruction results are [2]. However, the sampling scale
is directly proportional to the time of data acquisition. An
excessive sampling scale will lead to an increase of the measurement
time cost, overload of data processing and the
increase of overall product manufacturing cost. Therefore,
the research on intelligent sampling point planning of complex
surfaces is of great significance.
At present, scholars have carried out a lot of research on the
sampling algorithm of complex surface measurement. First,
the adaptability of the sampling algorithm is the focus of some
research. Liu et al. [3] proposed an adaptive sampling method
for point laser sensor based on the section method. Wang et
al. [4] proposed an adaptive sampling method based on a
August 2022
6-DOF Industrial Robot for curve and arc extension prediction,
which realized the digital sampling of unknown surfaces
of CAD models. Second, some studies pay more attention to
the feature information of the sampling model. Luca et al. [5]
proposed a sampling method on different scales of free-form
surfaces to find the best sampling strategy in the process of
surface reconstruction. Third, some research focuses on the
adaptability of on-line measurement. Yang et al. [6] proposed a
surface measurement method combining a spatial filling scanning
path with adaptive sampling. Lu et al. [7] proposed an
adaptive sampling method based on front-end path detection
in order to digitize unknown free-form surfaces accurately and
efficiently.
As described, the research of sampling point planning for
curved surfaces is mainly divided into two parts. One part of
the research focuses on the planning of the sampling trajectory.
The planning of sampling point distribution is based on
the planning of sampling trajectory, and the sampling points
are generally evenly distributed in these applications. The
other part of the research focuses on the sampling planning of
the workpiece with known CAD modeling. The distribution
of sampling points can be planned according to the curvature
characteristics of the original CAD model. However, there
are few studies on the sampling point planning of unknown
model workpieces in the existing literature. In this paper,
based on the principle of section line scanning measurement,
an adaptive sampling planning algorithm is proposed for
the surface data measurement of large complex surface components.
Based on the B-spline wavelet function, the curve
complexity relation is constructed, and the original sampling
points are planned by combining the arc length parameters
and the curve complexity parameters. This method can intelligently
plan the distribution of sampling points according
to the complexity of the curve shape, which improves the
disadvantage of manual planning of the sampling path in
the traditional uniform sampling, and solves the problem of
over dense distribution of sampling points in the low curvature
region of the uniform sampling result of spherical crown
surface.
IEEE Instrumentation & Measurement Magazine
1094-6969/22/$25.00©2022IEEE
17
Instrumentation & Measurement Magazine 25-5

Table of Contents for the Digital Edition of Instrumentation & Measurement Magazine 25-5
