Measurement of sole activity by digital image analysis

Abstract

Evaluating flatfish activity can be a useful tool for studying fish behavior and welfare. The aim of this work was to obtain a quantitative index for measuring flatfish activity using image analysis. Accordingly, motor activity of a sole population was recorded by digital video for three nights, bearing in mind the nocturnal lifestyle of the species. Subsequent image analysis was done by image subtraction of consecutive frames. The result was a “difference frame” showing the changes in the image area due to fish movement. Using these data, an image processing activity index (IPAI) was determined by measuring the percentage of area altered due to fish movement and by taking this percentage as an indicator of fish activity. Typical sole behavioral acts (take-off and surface swimming) were recorded by direct observation during the same intervals in order to calibrate the IPAI. A direct observation activity index (DOAI) was determined by weighting each kind of attitude according to its average duration. Results obtained from image analysis (IPAI) were compared with results obtained by direct observation (DOAI). A linear relationship between the two indexes was found with a correlation coefficient of r 2 = 0.80 for 92% of coverage area tank and r 2 = 0.90 for 210% of coverage area tank. Thus, this digital video-based index can be a highly reliable and accurate method for objectively measuring activity levels in sole with a low consumption of labor and time. Nevertheless, the application to farm conditions would need further research and an accurate calibration for other species.