V-STARS Measurement at the National Radio Astronomy Observatory (NRAO)
The following is a summary of the results from the demonstration of the Geodetic Systems, Inc. V-STARS system. V-STARS is a powerful, portable and precise system that uses one or more high-resolution digital cameras to measure objects via photogrammetry. The system is widely used throughout industry for part and tool inspection. GSI performed the demonstration at NRAO’s Very Large Array (VLA) facility in Socorro, New Mexico.
For the demonstration, a 25 meter diameter VLA reflector (similar to the one shown above) was measured in three different orientations; 90˚, 45˚, and 0˚. The intention was to measure the main surface of the reflector five times; two times at 90˚, twice at 45˚ (once before adjustment, once after adjustment), and once at 0˚. All these measurements were accomplished on the first night. In addition, the VLA’s subreflector was measured twice on the first afternoon. Also, a VLBA subreflector located on the ground was measured on the first afternoon three times. As the geometry turned out to be less than optimal for the 45˚ measurements (due to crane limitations), the visit was extended and the main surface of the reflector was measured six more times the following day and night. In total 11 measurements were made of the main reflector over the two day period of the visit.
Summary of Measurement and Results
Main Surface Measurements
Prior to photography, five retro-reflective targets were applied to each of the 172 panels that make up the reflector surface. One target was placed in each corner and a target was placed in the center of each panel. Therefore, there were a total of 860 surface targets. A few targets were also applied to the quadropod legs, and to the top of the feed structure. In addition, a three meter scale bar and a two meter scale bar were attached to the antenna to provide scale for the measurement. A small reference bar with five known points (called the AutoBar) was also placed on the reflector. Finally, about 40 special targets, called coded targets, were applied to the surface of the antenna. Each coded target is surrounded by a unique pattern of retro-reflective squares so it can be automatically identified. With the AutoBar and the coded targets, the pictures can be processed completely automatically. Targeting time took about three hours.
The antenna was measured eleven times. Each measurement of the antenna used about 100 photographs taken from various locations and heights all around the antenna. The on-site cranes were used for the photography. The different camera locations were needed to ensure all points on the object were seen from enough geometrically diverse locations to get good intersection angles for triangulation.
The photography time varied depending on the number of pictures and the orientation of the antenna but usually took 60 to 90 minutes. The processing time also varied depending on whether the measurement was an initial or a repeat measurement. In a repeat measurement, the processing time is about half since the approximate locations of the targets are already known. Processing time varied from about 30 to 60 minutes.
Each measured point is assigned a label to help identify it. The prefix of the label identifies what kind of point it is. The following labeling scheme was adopted.
• R#_# – Reference points on the main surface of the reflector. The first number of the label represents the ring number with the outermost ring being ring #1 and the innermost ring being ring #18. The second number represents the point’s location on the ring. The topmost point on a ring is #1, and the numbers increase in a clockwise direction around the ring (as you face the front of the reflector). For example, R3_12 is the 12th point clockwise from the top on the third ring from the edge.
6• QUAD – Points on the quadropods
• FEED – Points on the top of the Feed
• CODE – Special coded targets used to automate the measurement
• AUTOBAR – Points on the reference bar used to initialize the first measurement
• A – Points on the “A” scale bar
• B – Points on the “B” scale bar
The photogrammetric measurements produced results in an arbitrary working coordinate system defined by the AutoBar reference bar. These coordinates were transformed into a more meaningful coordinate system defined in the following manner. A best-fit parabola was fit to all the surface points from measurement #9 (File: 45˚ #4B.XYZ) because this was the most representative of all the measurements. The origin of the coordinate system is at the vertex, and the positive Z axis goes thru the focus. Thus, the X-Y plane is parallel to the face of the reflector. Finally, the Y axis goes through the perpendicular projection of the reference point at the top of the reflector (Point R1_1) on the X-Y plane. The coordinate system is illustrated in the two figures below.
As our demonstration shows, the V-STARS system provides a broad and powerful range of capabilities for meeting the demanding requirements for industrial measurement at NRAO.
- Measurement in unstable environments, or from unstable platforms.
- Non-contact measurement so the reflector is not deformed.
- Minimal interference with production resulting in low downtime.
- Ability to work in a wide variety of environmental conditions.
- High portability.
- Fast setup with little warm-up time and no pre-calibration required
- Fast, automatic measurement and very efficient repeat measurements.
- High accuracy and repeatability.
- Operation from a notebook computer.
- Fully integrated transformation and analysis capabilities.
We thank NRAO for the opportunity to demonstrate our system and its capabilities to you. We believe there are many applications at NRAO that are well suited to photogrammetric measurement. Please contact us at any time to discuss possible applications.