Pengembangan wahana pencitraan bawah air guna identifikasi dan kuantifikasi terumbu karang dengan metode jaringan syaraf tiruan
Abstract
Initially finger printing methods was used by police for criminal case investigation; then expanded to other practical application. Artificial Neural Network is one of the technique to process data to identify and recognize a particular pattern similar to the way human brain functions. In this thesis, I investigate and identify various coral features using artificial neural network (ANN) back-propagation method. The underwater vehicle was constructed to collect coral imagery, using; stainless steel body with two floats on the left and right and equipped with camera, GPS-map sounder, and light sources. The 3000 images were collected and pre-processed, to extract descriptor or key parameters of the features, during field measurements. The ANN back-propagation applied to identify the life form as sub-massive coral has an accuracy of 92.86%. Therefore, the application of this technique for features identification of the coral life form is recommended..
Collections
- MT - Fisheries [3007]