Position measurement of flight hardware components, such as Solid Rocket Boosters, External Tank, and Shuttle Orbiter, is a critical step in the stacking and processing flow of the Shuttle Space Transport System. Surveyors who deal with land, construction, and highway surveying often deal with similar position measurement problems. Even though the surveying techniques used to locate Shuttle components and launch support equipment make use of similar mathematical techniques (based on the fundamentals of trigonometry and plane geometry), the instrumentation used by NASA and United Space Alliance (USA) in these kinds of operations greatly surpasses the measurement precision capabilities of typical surveying equipment.

Figure 1. Example Survey Results of Two Targets

A low-cost alternative to expensive laboratory-grade surveying instrumentation was proposed and evaluated by the Engineering Development Contractor, Dynacs Inc., at KSC. This alternative is a digital vision system, composed of a digital camera, computer, and flat black and white targets. Figure 1 shows an example of relative error between true target positions and measured positions during a test of the vision measurement system. True target positions were determined by a precision survey performed by the NASA and USA Optics Group, using an array of high-quality laboratory-grade theodolites (see figure 2). The measurement error of the NASA/USA survey is less than 1 millimeter (mm).

In our test of the camera position measurement system, the camera coordinate system’s origin is a point on the camera axis at a distance f in front of the charged couple device (f is the focal length of the lens). The z-axis is positive in front and away from the camera origin. The x-axis is horizontal and positive to the left of the camera origin. The y-axis is vertical and positive up from the camera origin.

The formulas described in figure 3 estimate the camera position measurement resolution limit. Note that the numerical values give error in millimeters with z in feet. The x, y, z error can be larger than that predicted by the equations in figure 3. Lens distortion may be the primary cause. An example of y error, shown in figure 4, is approximately within the theoretical error, since the target positions in this test were all near the y-axis, where lens distortion has minimal effect. As shown in figure 4, the z measurement resolution is usually much lower than the x and y measurement resolutions. However, at close range (distance between camera and targets), the z resolution can surpass the x and y resolution.

The low-cost advantage of the digital camera vision position measurement system may offset the lower position resolution disadvantages in many applications.