Applications of shearography in industry include the detection of strain anomalies which result when engineering components containing defects are subjected to stress. The output derived from shearographic apparatus is a fringe pattern which is used to confirm the integrity of, or characterise defects within, the component under test. A step towards the automation of the process is to convert the fringe lines into a mathematical representation that a computer can use for analysis. Modelling can be achieved by fitting B-spline curves to the fringe patterns and using a search to find a best fit. The paper compares the results of the run time performance of three search methods applied to this problem namely; discrete hill-climbing, random mutation hill-climbing and genetic algorithm.