Development Of Jouvenile Catfish Counting Method With Image Processing And Mass Measurement

Abstract

Alternative methods for jouvenile fish counting to replace the manual method are image processing method and mass measurement method. Image processing is one of the computer algorithms that can observe and analyze an object without direct contact with the object being observed. The image of jouvenile fish was taken by using a container attached to a holding frame with a digital camera on its top. The resulted image was processed with a Visual Basic program to determine the area of jouvenile fish. The jouvenile fish area was then used to determine the number of fish. Based on the image analysis on the jouvenile size of 7.8 cm, the average area obtained was of 871 pixels with the standard deviation of 112 pixels. The jouvenile mass was measured manually on each sample. The average mass of jouveniles was 2.17 g with a standard deviation of 0.66 and an average body length of 6.48 cm. The measurement results were then used for the simulation data with Microsoft Excel. The simulation was to calculate the probability of certain jouvenile mass occurrences. The accuracy of image processing was quite high, that is, the calculation error was 4% for image capturing 0 second after fish release, 3% for 5 second after release, 5% for 10 second after release and 10% for 15 seconds after release in the grade 3-4 cm. In the grade 5-7 cm the average error was 6% for image capturing 0 second after fish release, 4% for 5 second after release, 7% for 10 second after release and 14% for 15 second after release. The best accuracy was obtained when image was taken at an interval of 5 seconds. The error of mass measurement method as it is compared to manual counting was 3.5 %, while the simulation of mass measurement method error was 2.3%.