Ensemble of pruned bagged mixture density networks for improved water quality retrieval using Sentinel-2 and Landsat-8 remote sensing data
(2024) In IEEE Geoscience and Remote Sensing Letters 21.- Abstract
- Remote sensing (RS) data provide large-scale observations to measure water quality parameters (WQPs) like turbidity (Turb), which indicates the haziness of the water. Accurately estimating these parameters solely from RS data is inherently complex due to various factors, necessitating the use of advanced models capable of capturing the intricate relationships between RS spectral bands as an input and the target parameter as an output. One promising approach is the use of ensemble machine learning (ML) models, which construct more complex models by leveraging the complementary strengths of multiple base models. In this letter, a novel method known as pruned bagged mixture density network (PBMDN) was proposed. First, using a bootstrap-based... (More)
- Remote sensing (RS) data provide large-scale observations to measure water quality parameters (WQPs) like turbidity (Turb), which indicates the haziness of the water. Accurately estimating these parameters solely from RS data is inherently complex due to various factors, necessitating the use of advanced models capable of capturing the intricate relationships between RS spectral bands as an input and the target parameter as an output. One promising approach is the use of ensemble machine learning (ML) models, which construct more complex models by leveraging the complementary strengths of multiple base models. In this letter, a novel method known as pruned bagged mixture density network (PBMDN) was proposed. First, using a bootstrap-based bagging approach, a pool of base mixture density network (MDN) models was generated. Then, a forward pruning scheme was utilized to find an optimal subset of the base models for final ensemble aggregation. The coincident in situ measurements of Turb and multispectral Sentinel-2 (S2) and Landsat-8 (L8) RS data for three water bodies in USA were used to evaluate the performance of PBMDN. Results showed that PBMDN could achieve lower estimation errors [mean absolute percentage error (MAPE) of 25.25% for S2 and 35.71% for L8] compared to the single MDN model (MAPE of 40.53% for S2 and 47.92% for L8). PBMDN also performed significantly better than other widely used Turb estimation techniques, including other ML models and semi-empirical algorithms, indicating its strong potential in the estimation of WQPs using RS data. (Less)
Please use this url to cite or link to this publication:
https://lup.lub.lu.se/record/07cd1373-5ef8-4eb0-8395-1d0fdfefc764
- author
- Taheri Dehkordi, Alireza LU ; Hashemi, Hossein LU ; Naghibi, Seyed Amir LU and Mehran, Ali
- organization
- publishing date
- 2024-08-01
- type
- Contribution to journal
- publication status
- published
- subject
- in
- IEEE Geoscience and Remote Sensing Letters
- volume
- 21
- pages
- 5 pages
- publisher
- IEEE - Institute of Electrical and Electronics Engineers Inc.
- external identifiers
-
- scopus:85200222056
- ISSN
- 1545-598X
- DOI
- 10.1109/LGRS.2024.3436920
- language
- English
- LU publication?
- yes
- id
- 07cd1373-5ef8-4eb0-8395-1d0fdfefc764
- date added to LUP
- 2024-08-09 12:30:57
- date last changed
- 2024-08-12 16:19:30
@article{07cd1373-5ef8-4eb0-8395-1d0fdfefc764, abstract = {{Remote sensing (RS) data provide large-scale observations to measure water quality parameters (WQPs) like turbidity (Turb), which indicates the haziness of the water. Accurately estimating these parameters solely from RS data is inherently complex due to various factors, necessitating the use of advanced models capable of capturing the intricate relationships between RS spectral bands as an input and the target parameter as an output. One promising approach is the use of ensemble machine learning (ML) models, which construct more complex models by leveraging the complementary strengths of multiple base models. In this letter, a novel method known as pruned bagged mixture density network (PBMDN) was proposed. First, using a bootstrap-based bagging approach, a pool of base mixture density network (MDN) models was generated. Then, a forward pruning scheme was utilized to find an optimal subset of the base models for final ensemble aggregation. The coincident in situ measurements of Turb and multispectral Sentinel-2 (S2) and Landsat-8 (L8) RS data for three water bodies in USA were used to evaluate the performance of PBMDN. Results showed that PBMDN could achieve lower estimation errors [mean absolute percentage error (MAPE) of 25.25% for S2 and 35.71% for L8] compared to the single MDN model (MAPE of 40.53% for S2 and 47.92% for L8). PBMDN also performed significantly better than other widely used Turb estimation techniques, including other ML models and semi-empirical algorithms, indicating its strong potential in the estimation of WQPs using RS data.}}, author = {{Taheri Dehkordi, Alireza and Hashemi, Hossein and Naghibi, Seyed Amir and Mehran, Ali}}, issn = {{1545-598X}}, language = {{eng}}, month = {{08}}, publisher = {{IEEE - Institute of Electrical and Electronics Engineers Inc.}}, series = {{IEEE Geoscience and Remote Sensing Letters}}, title = {{Ensemble of pruned bagged mixture density networks for improved water quality retrieval using Sentinel-2 and Landsat-8 remote sensing data}}, url = {{http://dx.doi.org/10.1109/LGRS.2024.3436920}}, doi = {{10.1109/LGRS.2024.3436920}}, volume = {{21}}, year = {{2024}}, }