Automatic Catfish Sorter and Counter Based on Weight Classification

Abstract

Catfish is one of the main raw materials in the fisheries sector which also supports the Indonesian economy. However, manual sorting of catfish makes the weight of catfish not uniform when marketed. Therefore, tools and machines are needed that automatically sort catfish in large quantities, making it more effective and easier for catfish breeders to use. In this study the process of making the system was realized in 3 stages, namely needs analysis, implementation, and testing of tools. The purpose of this system is to be able to classify the weight of catfish in the large category with fish weighing > 140 grams, medium with fish weighing 80-140 grams and small with fish weighing 20-80 grams. From the system test results, the average accuracy is 99.56. % and a precision of 98.75% for load cell sensor readings, while for E18 infrared sensor readings the results are always known, but data is not always sent and displayed on the LCD screen. Classification of fish categories carried out by the system has worked well, where the system is able to sort and count correctly 14 times out of 15 tests with an accuracy of 93%. The average computing time required for this tool is 3.67 seconds for 15 tests.