Monitoring the growth of farmed fish is an important task which is currentl
y difficult to carry out. An underwater stereo image analysis technique off
ers the potential for estimating key dimensions of free-swimming fish, from
which the fish mass can be estimated. This paper describes the development
of a three-dimensional point distribution model to capture the typical sha
pe and variability of salmon viewed from the side. The model was fitted to
stereo images of test fish by minimizing an energy function, which was base
d on probability distributions. The minimization was an iterated two-step m
ethod in which edges were selected for magnitude, direction, and proximity
to the model, and the model was then fitted to the edges. A search strategy
for locating the edges in 3D was devised. The model is tested on two image
sets. In the first set 19 of the 26 fish are located in spite of their var
iable appearance and the presence of neighboring fish. In the second set th
e measurements made on 11 images of fish are compared with manual measureme
nts of the fish dimensions and show an average error in length estimation o
f 5%. (C) 2000 Academic Press.