The quality of underwater video photography is limited by the visibility of the water column. Generally speaking, it is difficult to tell the dimension of the target directly from the underwater images. To overcome this problem, one may project a laser stripe onto the target and measure the displacement of the laser scan lines relative to a straight baseline. The displacement reveals the profile of the object at that location. With a calibrated CCD camera, the displacement expressed in pixels can be converted into the dimension of the target in physical units. To obtain a broader view at a closer distance to reduce the influence of suspension particles, a wide-angle lens is usually used. Therefore the image is non-linear and distorted in the corners. Rather than finding out the optical and geometrical parameters of the CCD camera and the environment, the authors propose building several mappings between the world coordinate system and the image frame to cover the scope of the CCD. The authors lay vertical and horizontal grid lines of 5 cm span on an acrylic plate that is aligned with the laser beam scan plane. These grid lines serve as the "longitudes" and "latitudes" of the map. On the captured image, the authors curve-fit the controlled points of the grid lines in pixels. A pair of interpolated longitude and latitude which passes through the target point are used to estimate the location of the point in the world coordinate system. Coins and one bottle of known sizes are used for verification of this approach. The results indicate that the error varies between 1/spl sim/2% measured from a distance of one metre.
[1]
Ware Myers,et al.
Industry Begins to Use Visual Pattern Recognition
,
1980,
Computer.
[2]
Stephen J. Maybank,et al.
On plane-based camera calibration: A general algorithm, singularities, applications
,
1999,
Proceedings. 1999 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No PR00149).
[3]
H. Hatze,et al.
High-precision three-dimensional photogrammetric calibration and object space reconstruction using a modified DLT-approach.
,
1988,
Journal of biomechanics.
[4]
A. Kiessling,et al.
Methods For Analyzing Three Dimensional Scenes
,
1975,
IJCAI.
[5]
Donna M. Kocak,et al.
Real-time 3D underwater imaging and mapping using a laser line scan technique
,
1997,
Optics & Photonics.
[6]
H. M. Karara,et al.
Direct Linear Transformation from Comparator Coordinates into Object Space Coordinates in Close-Range Photogrammetry
,
2015
.
[7]
Roger Y. Tsai,et al.
A versatile camera calibration technique for high-accuracy 3D machine vision metrology using off-the-shelf TV cameras and lenses
,
1987,
IEEE J. Robotics Autom..