Abstract. Three hundred and nine width measurements at various heights to 10 m on a metal light pole were calculated from video images captured with a prototype video-laser rangefinder instrument. Data were captured at distances from 6 to 15 m. The endpoints for the width measurements were manually selected to the nearest pixel from individual video frames.

Chi-square analysis shows that 95 percent of measurements can be expected to be within 13 mm of the actual. All errors lower than -13 mm were on nearly horizontal images. Poor contrast causing poor edge detection was the source of most of this error. All errors greater than 13 mm occurred between rows 280 and 390 on the image. Improper focus seemed to be the main contributor to positive errors. Improvements to the camera and lens system should improve these results considerably.