TY - JOUR
T1 - Identification of land use for crops in Jawa Timur province based on multispectral imaging by combining multi-layer density-based spatial clustering of applications with noise and time-weighted dynamic time wrapping
AU - Ferdiyansyah, Ahmad Rifan
AU - Prastyo, Dedy Dwi
AU - Fithriasari, Kartika
N1 - Publisher Copyright:
© 2024 Author(s).
PY - 2024/11/15
Y1 - 2024/11/15
N2 - Up-to-date information about land use from crops can improve the quality of agricultural statistical data in Indonesia. However, updating activities through ground checks generally requires considerable resources so that various alternative methods continue to be developed to identify land use for agriculture. One alternative to identify land use is to utilize remote sensing data using the time-weighted Dynamic Time Wrapping (TWDTW) method. Several studies have proven that the TWDTW method can produce fairly accurate predictions. However, their studies are limited to small areas or only focused on a single food crop. When applied to a larger area with highly varied cropping patterns, the TWDTW method tends to produce inaccurate predictions. Therefore, this study combines methods to predict land use in large areas with various cropping patterns. This combination method involves creating subclasses using Multi-Layer DBSCAN to capture variations in cropping patterns, followed by predicting land use using the TWDTW method. In general, prediction results using the TWDTW method have better accuracy than those using DTW or Euclidean. In addition, combining them with the Multi-Layer DBSCAN method will increase the accuracy of the predictions.
AB - Up-to-date information about land use from crops can improve the quality of agricultural statistical data in Indonesia. However, updating activities through ground checks generally requires considerable resources so that various alternative methods continue to be developed to identify land use for agriculture. One alternative to identify land use is to utilize remote sensing data using the time-weighted Dynamic Time Wrapping (TWDTW) method. Several studies have proven that the TWDTW method can produce fairly accurate predictions. However, their studies are limited to small areas or only focused on a single food crop. When applied to a larger area with highly varied cropping patterns, the TWDTW method tends to produce inaccurate predictions. Therefore, this study combines methods to predict land use in large areas with various cropping patterns. This combination method involves creating subclasses using Multi-Layer DBSCAN to capture variations in cropping patterns, followed by predicting land use using the TWDTW method. In general, prediction results using the TWDTW method have better accuracy than those using DTW or Euclidean. In addition, combining them with the Multi-Layer DBSCAN method will increase the accuracy of the predictions.
UR - http://www.scopus.com/inward/record.url?scp=85210253368&partnerID=8YFLogxK
U2 - 10.1063/5.0230583
DO - 10.1063/5.0230583
M3 - Conference article
AN - SCOPUS:85210253368
SN - 0094-243X
VL - 3201
JO - AIP Conference Proceedings
JF - AIP Conference Proceedings
IS - 1
M1 - 060007
T2 - 9th SEAMS-UGM International Conference on Mathematics and its Applications 2023: Integrating Mathematics with Artificial Intelligence to Broaden its Applicability through Industrial Collaborations
Y2 - 25 July 2023 through 28 July 2023
ER -