Comparing the performance of two scientific tools for obtaining fish length measurements

Precise descriptions of size structure are crucial for the adaptive management of marine fish populations influenced by human activity and environmental factors. Stereo-video systems are powerful tools for monitoring fish populations. Yet, due to the high investment required in software and equipment, stereo-videography can have some financial issues. This study compared the performance of the commercial software EventMeasure from SeaGIS and the open-access R package StereoMorph. We evaluated the effect of distance to the system, rotation movement of the object, and true length on the accuracy and precision of each software. Additionally, we used a diver-operated stereo-video system to obtain in situ measurements of reef fish species. EventMeasure was generally more accurate (error = 0.53%) and precise (CV = 0.35%) than StereoMorph (error = 4.54%; CV = 0.65%). However, the latter showed errors < 5% when measuring objects at distances up to 4 m and close to the plane axis. The precision for in situ measurements for EventMeasure (CV = 15.0%) was similar to that of StereoMorph (Mean CV = 15.8%). We found a high correlation (ρ = 0.93, p < 0.001) between paired fish length estimation from both software, although StereoMorph was slightly more precise than EventMeasure (Mean CV = 2.07% and Mean CV = 2.11%, respectively). This open-access software provided suitable accuracy and precision results despite some limitations, offering the option to reduce software costs without compromising accuracy for affordability.