Vegetation classification algorithm using convolutional neural network ResNet50 for vegetation mapping in Bandung district area

Vegetation classification algorithm using convolutional neural network ResNet50 for vegetation mapping in Bandung district area

Bandung District is one of crop provider for West Java Province. About 31.158,22 ha is used for crop. However, some of them are not maintained well due to lack of vegetation map information. Local authority has tried to map the vegetation in their area by using free license satellite images, and aer...

Full description

Saved in:
Bibliographic Details
Published in:Jurnal infotel (Online) Vol. 14; no. 2; pp. 146 - 153
Main Authors: Astuti, Rina Pudji, Rachmawati, Ema, Edwar, Edwar, Siregar, Simon, Sardi, Indra Lukmana, Fahmi, Arfianto, Agustian, Yayan, Yoga Putra, Agus Cahya Ananda, Daffa, Faishal
Format: Journal Article
Language:English
Indonesian
Published: LPPM Institut Teknologi Telkom Purwokerto 31-05-2022
Subjects:
aeral images
resnet50
vegetation mapping
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Bandung District is one of crop provider for West Java Province. About 31.158,22 ha is used for crop. However, some of them are not maintained well due to lack of vegetation map information. Local authority has tried to map the vegetation in their area by using free license satellite images, and aerial images from Unmanned Aerial Vehicle (UAV). Despite both images being able to provide large plantation area images, both are unable to classify the vegetation type in those images. Telkom University with Bandung Agriculture Regional Office (Dinas Pertanian Kabupaten Bandung) has conducted joint research to develop algorithm based on 50-layer residual neural network (ResNet50) to classify the vegetation type. The input is of this algorithm is primarily aerial images are captured from different type, height, and position of crops. Seven different ResNet50 configurations have been set and simulated to classify the crop images. The result is the configuration with resized images, employing triangular policy of cyclic learning rate with rate 1.10−7 – 1.10−4 comes out as the best setup with more than 95% accuracy and relatively low loss.
ISSN:2085-3688
2460-0997
DOI:10.20895/infotel.v14i2.756