A common problem in Computer Vision is to separate and analyse the
objects of interest in a sequence of images in order to extract some physic character-
istics that typically will be used for any kind of study or control process. Nowadays,
many of these kind of applications have been successfully applied in numerous ar-
eas such as navigation systems, °ow analysis, image diagnosis in medicine, structure
analysis in civil engineering and others. In these scenarios, an e±cient computational
approach provides the means to achieve accurate results at a very low cost.
This paper puts forward a Computer Vision based algorithm to analyse the
properties of water streams in a scale model of an over°ow channel with the objective
of simulate di®erent conditions and obtaining a more e®ective channel design.
The proposed algorithm obtain pixel to real distance transformation in the im-
age, calculates physical properties of the stream by using several Arti¯cial Vision
techniques and analyses the resulting data.
A set of assays have been made the performance and evaluate the the algorithm.