Research of Sub-pixel Sub-division Location Algorithm for Spot Center of Digital Aerial Survey Camera Image

摘要

So far, the geometric laboratory calibration system of optical aerial survey camera has been well established, while there is not a perfect system for digital aerial survey camera and the calibration of digital camera has not been included in the metrological verification system. In order to make the photogrammetry project rule-based, it's imperative to establish geometric laboratory calibration system of digital survey camera. Exact measuring angle method is the most common method to calibrate the camera in lab. The sub-pixel sub-division location of spot center for point-source image of digital camera is one of the key techniques to calibrate the camera based on the exact measuring angle method in lab. A novel location algorithm is proposed on account of the realities of the lab system. First, the Gauss filter is introduced to reduce the noise from the stray light and so on. Then the integration of iterative threshold segmentation and seed region growing is adopted to identify the general location of the spot. Thirdly, using the bilinear interpolation, the pixels used to estimate the location are increased. Finally, we use the proposed improved moment method to determinate the precise centroid location of the digital image. Experimental results show that the spots are identified with 93 percent accuracy and the centroid accuracy is superior to 1/40 of a pixel and improves about 2 to 4 times than other algorithms, which indicates that the proposed method is effective, convenient and precision.