|
In order to design and manufacture the Five-hundred-meter Aperture Spherical Radio Telescope (FAST) active reflector
measuring equipment, measurement on each reflector panel surface shape was presented, static measurement of the
whole neutral spherical network of nodes was performed, real-time dynamic measurement at the cable network dynamic
deformation was undertaken. In the implementation process of the FAST, reflector panel surface shape detection was
completed before eleven-meter reflector panel installation. Binocular vision system was constructed based on the method
of binocular stereo vision in machine vision, eleven-meter reflector panel surface shape was measured with
photogrammetry method. Cameras were calibrated with the feature points. Under the linearity camera model, the lighting
spot array was used as calibration standard pattern, and the intrinsic and extrinsic parameters were acquired. The images
were collected for digital image processing and analyzing with two cameras, feature points were extracted with the
detection algorithm of characteristic points, and those characteristic points were matched based on epipolar constraint
method. Three-dimensional reconstruction coordinates of feature points were analyzed and reflective panel surface shape
structure was established by curve and surface fitting method. The error of reflector panel surface shape was calculated
to realize automatic measurement on reflector panel surface shape. The results show that unit reflector panel surface
inspection accuracy was 2.30mm, within the standard deviation error of 5.00mm. Compared with the requirement of
reflector panel machining precision, photogrammetry has fine precision and operation feasibility on eleven-meter
reflector panel surface shape measurement for FAST.
|
