Telecentric optical path and computer vision detection system were used to detect the space azimuth of a specific board. The number of feature points was studied about its influence on the detection, and further more about its determination to increase the detection reliability. The pictures of a specific board were captured by charge-coupled device(CCD), and the space azimuth of the specific board was obtained by image processing. A large number of experiments show that the detection accuracy has a certain relationship with the number of feature points. If the distance between specific board and CCD is from 2 m to 5 m, the calculated space azimuth is more accurate when the number of effective feature points on the specific board is greater than 28.