Analisis Perbandingan Akurasi DTM Berbasis Data Uav-Lidar Menggunakan Metode Filtering Cloth Simulation Filter (CSF) dan Macro Terrasolid (Studi Kasus: Desa Bungurasih dan Tambakoso, Kecamatan Waru)

Dewi, Zuniarni Kusuma (2026) Analisis Perbandingan Akurasi DTM Berbasis Data Uav-Lidar Menggunakan Metode Filtering Cloth Simulation Filter (CSF) dan Macro Terrasolid (Studi Kasus: Desa Bungurasih dan Tambakoso, Kecamatan Waru). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5016221007-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (10MB) | Request a copy

Abstract

Ketersediaan Digital Terrain Model yang akurat pada kawasan urban fringe dengan karakteristik lahan yang kompleks sangat ditentukan oleh kualitas tahap filtering point cloud dalam memisahkan titik ground dari objek non-tanah. Penelitian ini dilaksanakan di Desa Bungurasih yang mewakili permukiman padat dan infrastruktur kompleks serta Desa Tambakoso yang mewakili kawasan tambak dan lahan basah pada Kecamatan Waru menggunakan data UAV-LiDAR yang diakuisisi dengan DJI Matrice 350 RTK Zenmuse L2. Tahapan pengolahan dengan penyaringan point cloud menggunakan metode Macro TerraSolid melalui perangkat lunak Spatix berbasis 13 susunan macro, serta metode CSF melalui perangkat lunak CloudCompare dilanjutkan dengan pemeriksaan visual dan statistik titik tanah, pembentukan DTM melalui interpolasi TIN, serta uji ketelitian geometrik menggunakan Independent Check Point (ICP) berbasis GNSS sesuai SNI 9135-2:2023. Metode Macro menggunakan parameter terrain angle, iteration angle, iteration distance, dan max building size yang disesuaikan melalui pendekatan trial and error, sedangkan metode CSF menggunakan parameter scene, cloth resolution, class threshold, dan max iteration yang disesuaikan dengan kerapatan objek pada masing-masing lokasi. Hasil uji akurasi menunjukkan bahwa metode Macro secara konsisten menghasilkan ketelitian yang lebih tinggi dibandingkan CSF pada kedua lokasi. Di Desa Bungurasih metode Macro memperoleh CE90 sebesar 0.169 m dan LE90 sebesar 0.149 m yang memenuhi standar Kelas 1 Skala 1:1.000, sedangkan CSF hanya memperoleh CE90 sebesar 0.294 m dan LE90 sebesar 0.268 m yang memenuhi standar Kelas 2. Di Desa Tambakoso kedua metode memenuhi standar Kelas 1 Skala 1:1.000, dengan metode Macro mencatatkan nilai CE90 sebesar 0.110 m dan LE90 sebesar 0.124 m lebih presisi dibandingkan CSF yang memperoleh CE90 sebesar 0.173 m dan LE90 sebesar 0.183 m. Berdasarkan hasil tersebut, metode Macro TerraSolid terbukti lebih presisi dan stabil dalam memisahkan titik ground dari berbagai objek non-tanah dan noise.
=======================================================================================================================================
Digital elevation modeling in urban fringe areas with complex characteristics heavily relies on the quality of the point cloud filtering stage to generate an accurate Digital Terrain Model (DTM). This study aims to establish the stages of UAV-LiDAR-based DTM generation, analyze algorithm performance, and evaluate as well as compare the vertical accuracy of DTMs produced by the Cloth Simulation Filter (CSF) and Macro TerraSolid methods. Case studies were conducted in Bungurasih Village representing dense residential areas and complex infrastructure and Tambakoso Village representing aquaculture ponds and wetlands, Waru District. The primary data were acquired using a DJI Matrice 350 RTK UAV equipped with a Zenmuse L2 sensor, which were subsequently validated using GNSS-based Independent Check Points (ICPs) in accordance with the SNI 9135-2:2023 standard. In the Tambakoso area, despite the detection of negative elevation anomalies caused by sensor wave absorption over the aquaculture ponds, the Macro method still demonstrated superior object separation performance compared to CSF. Based on the Linear Error 90% (LE90) vertical accuracy test, the Macro TerraSolid method proved to be significantly more precise in both study areas. In the densely populated area of Bungurasih Village, the Macro method achieved an LE90 value of 0.149 m, satisfying the stringent Class 1 standards for a 1:1,000 scale, whereas CSF obtained 0.268 m, which only met the Class 2 standards. Meanwhile, in the wetland area of Tambakoso Village, both methods successfully met the Class 1 qualifications for a 1:1,000 scale, with the Macro method recording a higher accuracy of 0.124 m compared to 0.183 m for CSF. Overall, this study concludes that the Macro TerraSolid method is highly recommended for large-scale DTM extraction across various morphological typologies, as it has proven to be more precise, stable, and reliable in filtering out various types of non-ground objects and noise.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Cloth Simulation Filter (CSF), Digital Terrain Model (DTM), filtering, Macro Terrasolid, UAV-LIDAR
Subjects:	T Technology > TA Engineering (General). Civil engineering (General) > TA590 Topographical surveying
Divisions:	Faculty of Civil, Planning, and Geo Engineering (CIVPLAN) > Geomatics Engineering > 29202-(S1) Undergraduate Thesis
Depositing User:	Zuniarni Kusuma Dewi
Date Deposited:	21 Jul 2026 02:37
Last Modified:	21 Jul 2026 02:37
URI:	http://repository.its.ac.id/id/eprint/135903

Actions (login required)

View Item