{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,3]],"date-time":"2026-09-03T17:51:46Z","timestamp":1788457906920,"version":"build-2803163510"},"reference-count":36,"publisher":"MDPI AG","issue":"2","license":[{"start":{"date-parts":[[2026,1,22]],"date-time":"2026-01-22T00:00:00Z","timestamp":1769040000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100017691","name":"Guangxi Key Research and Development Program","doi-asserted-by":"crossref","award":["GuikeAB24010338"],"award-info":[{"award-number":["GuikeAB24010338"]}],"id":[{"id":"10.13039\/501100017691","id-type":"DOI","asserted-by":"crossref"}]},{"DOI":"10.13039\/501100017691","name":"Guangxi Key Research and Development Program","doi-asserted-by":"crossref","award":["GuikeAB25069340"],"award-info":[{"award-number":["GuikeAB25069340"]}],"id":[{"id":"10.13039\/501100017691","id-type":"DOI","asserted-by":"crossref"}]},{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["32360374"],"award-info":[{"award-number":["32360374"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]},{"name":"Innovation Project of Guangxi Graduate Education","award":["YCSW2024383"],"award-info":[{"award-number":["YCSW2024383"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Sensors"],"abstract":"<jats:p>Soil organic matter (SOM) is a key indicator for evaluating soil fertility and ecological functions, and hyperspectral technology provides an effective means for its rapid and non-destructive estimation. However, in practical soil systems, the spectral response of SOM is often highly covariant with mineral composition, moisture conditions, and soil structural characteristics. Under small-sample conditions, hyperspectral SOM modeling results are usually highly sensitive to spectral preprocessing methods, sample perturbations, and model architecture and parameter configurations, leading to fluctuations in predictive performance across independent runs and thereby limiting model stability and practical applicability. To address these issues, this study proposes a multi-strategy collaborative deep learning modeling framework for small-sample conditions (SE-EDCNN-DA-LWGPSO). Under unified data partitioning and evaluation settings, the framework integrates spectral preprocessing, data augmentation based on sensor perturbation simulation, multi-scale dilated convolution feature extraction, an SE channel attention mechanism, and a linearly weighted generalized particle swarm optimization algorithm. Subtropical red soil samples from Guangxi were used as the study object. Samples were partitioned using the SPXY method, and multiple independent repeated experiments were conducted to evaluate the predictive performance and training consistency of the model under fixed validation conditions. The results indicate that the combination of Savitzky\u2013Golay filtering and first-derivative transformation (SG\u20131DR) exhibits superior overall stability among various preprocessing schemes. In model structure comparison and ablation analysis, as dilated convolution, data augmentation, and channel attention mechanisms were progressively introduced, the fluctuations of prediction errors on the validation set gradually converged, and the performance dispersion among different independent runs was significantly reduced. Under ten independent repeated experiments, the final model achieved R2 = 0.938 \u00b1 0.010, RMSE = 2.256 \u00b1 0.176 g\u00b7kg\u22121, and RPD = 4.050 \u00b1 0.305 on the validation set, demonstrating that the proposed framework has good modeling consistency and numerical stability under small-sample conditions.<\/jats:p>","DOI":"10.3390\/s26020741","type":"journal-article","created":{"date-parts":[[2026,1,26]],"date-time":"2026-01-26T16:13:48Z","timestamp":1769444028000},"page":"741","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":2,"title":["Stability-Oriented Deep Learning for Hyperspectral Soil Organic Matter Estimation"],"prefix":"10.3390","volume":"26","author":[{"given":"Yun","family":"Deng","sequence":"first","affiliation":[{"name":"Guangxi Key Laboratory of Embedded Technology and Intelligent System, Guilin University of Technology, 12 Jiangan Road, Guilin 541004, China"},{"name":"College of Computer Science and Engineering, Guilin University of Technology, 319 Yanshan Street, Yanshan District, Guilin 541006, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yuxi","family":"Shi","sequence":"additional","affiliation":[{"name":"Guangxi Key Laboratory of Embedded Technology and Intelligent System, Guilin University of Technology, 12 Jiangan Road, Guilin 541004, China"},{"name":"College of Computer Science and Engineering, Guilin University of Technology, 319 Yanshan Street, Yanshan District, Guilin 541006, China"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1968","published-online":{"date-parts":[[2026,1,22]]},"reference":[{"key":"ref_1","doi-asserted-by":"crossref","unstructured":"Heller Pearlshtien, D., and Ben-Dor, E. (2020). Effect of organic matter content on the spectral signature of iron oxides across the VIS\u2013NIR spectral region in artificial mixtures: An example from a red soil from Israel. Remote Sens., 12.","DOI":"10.3390\/rs12121960"},{"key":"ref_2","first-page":"1050","article-title":"Effects of natural and anthropogenic factors on soil organic matter content in cultivated land","volume":"54","author":"Li","year":"2023","journal-title":"Soil Bull."},{"key":"ref_3","doi-asserted-by":"crossref","first-page":"730","DOI":"10.1108\/GS-05-2023-0041","article-title":"The modified model for hyperspectral estimation of soil organic matter using positive and inverse grey relational degree","volume":"13","author":"Xu","year":"2023","journal-title":"Grey Syst. Theory Appl."},{"key":"ref_4","first-page":"1093","article-title":"Spectral analysis methods and applications of active soil organic matter components","volume":"52","author":"Zhou","year":"2020","journal-title":"Soils"},{"key":"ref_5","doi-asserted-by":"crossref","unstructured":"Ding, Z., Liu, K., Grunwald, S., Smith, P., Ciais, P., Wang, B., Wadoux, A.M., Ferreira, C., Karunaratne, S., and Shurpali, N. (2025). Advancing soil organic carbon prediction: A comprehensive review of technologies, AI, process-based and hybrid modelling approaches. Adv. Sci., 12.","DOI":"10.1002\/advs.202504152"},{"key":"ref_6","first-page":"1320","article-title":"Construction of hyperspectral prediction models for soil organic matter content in soils developed from loessial parent materials in Liaoning Province","volume":"53","author":"Wang","year":"2022","journal-title":"Soil Bull."},{"key":"ref_7","first-page":"869","article-title":"Hyperspectral estimation of soil organic carbon content in reforested land in the middle Yellow River Basin: A case study of Daning County","volume":"50","author":"Deng","year":"2022","journal-title":"Shanxi Agric. Sci."},{"key":"ref_8","first-page":"3738","article-title":"Hyperspectral prediction of soil organic carbon content in Jiangxi Province based on CARS\u2013BPNN","volume":"55","author":"Wu","year":"2022","journal-title":"Sci. Agric. Sin."},{"key":"ref_9","doi-asserted-by":"crossref","unstructured":"Bai, Z., Xie, M., Hu, B., Luo, D., Wan, C., Peng, J., and Shi, Z. (2022). Estimation of soil organic carbon using Vis\u2013NIR spectral data and spectral feature band selection in southern Xinjiang, China. Sensors, 22.","DOI":"10.3390\/s22166124"},{"key":"ref_10","first-page":"1","article-title":"Attention-based CNN ensemble for soil organic carbon content estimation with spectral data","volume":"19","author":"Zhao","year":"2022","journal-title":"IEEE Geosci. Remote Sens. Lett."},{"key":"ref_11","doi-asserted-by":"crossref","unstructured":"Wang, H., Sun, Q., Niu, X., Liu, K., Zhang, J., Hao, Z., and Xu, D. (2025). Soil organic carbon prediction using an efficient channel attention-enhanced CNN\u2013LSTM model with LUCAS spectral library. Eur. J. Soil Sci., 76.","DOI":"10.1111\/ejss.70202"},{"key":"ref_12","doi-asserted-by":"crossref","first-page":"116867","DOI":"10.1016\/j.geoderma.2024.116867","article-title":"Soil organic carbon mapping utilizing convolutional neural networks and earth observation data: A case study in Bavaria, Germany","volume":"444","author":"Tziolas","year":"2024","journal-title":"Geoderma"},{"key":"ref_13","first-page":"3144","article-title":"Estimation of soil organic carbon content based on UAV hyperspectral remote sensing in black soil regions","volume":"50","author":"Yang","year":"2025","journal-title":"Earth Sci."},{"key":"ref_14","first-page":"833","article-title":"Construction of rapid detection models for soil organic matter based on hyperspectral data","volume":"52","author":"Li","year":"2021","journal-title":"J. Shandong Agric. Univ. (Nat. Sci. Ed.)"},{"key":"ref_15","first-page":"2022","article-title":"Origin traceability of Angelica dahurica based on near-infrared spectroscopy combined with data augmentation CNN algorithms","volume":"59","author":"Guo","year":"2024","journal-title":"Chin. Pharm. J."},{"key":"ref_16","doi-asserted-by":"crossref","unstructured":"Wang, L., Peng, J., and Sun, W. (2019). Spatial\u2013spectral squeeze-and-excitation residual network for hyperspectral image classification. Remote Sens., 11.","DOI":"10.3390\/rs11070884"},{"key":"ref_17","doi-asserted-by":"crossref","first-page":"1627","DOI":"10.1021\/ac60214a047","article-title":"Smoothing and differentiation of data by simplified least squares procedures","volume":"36","author":"Savitzky","year":"1964","journal-title":"Anal. Chem."},{"key":"ref_18","doi-asserted-by":"crossref","first-page":"1201","DOI":"10.1016\/j.trac.2009.07.007","article-title":"Review of the most common pre-processing techniques for near-infrared spectra","volume":"28","author":"Rinnan","year":"2009","journal-title":"Trends Anal. Chem."},{"key":"ref_19","doi-asserted-by":"crossref","unstructured":"Gao, Y., Ma, L., Zhang, Z., Pan, X., Yuan, Z., Wang, C., and Yu, D. (2025). The influence of viewing geometry on hyperspectral-based soil property retrieval. Remote Sens., 17.","DOI":"10.3390\/rs17142510"},{"key":"ref_20","first-page":"738","article-title":"Near-infrared spectroscopic determination of soluble solid content in watermelon based on the SPXY algorithm","volume":"39","author":"Wang","year":"2019","journal-title":"Spectrosc. Spectr. Anal."},{"key":"ref_21","doi-asserted-by":"crossref","unstructured":"Zhao, F., Zhang, J., Meng, Z., and Liu, H. (2021). Densely connected pyramidal dilated convolutional network for hyperspectral image classification. Remote Sens., 13.","DOI":"10.3390\/rs13173396"},{"key":"ref_22","doi-asserted-by":"crossref","first-page":"60","DOI":"10.1186\/s40537-019-0197-0","article-title":"Image data augmentation for deep learning: A survey","volume":"6","author":"Shorten","year":"2019","journal-title":"J. Big Data"},{"key":"ref_23","first-page":"1227","article-title":"Progress in spectral data augmentation methods and their applications","volume":"44","author":"Tang","year":"2025","journal-title":"J. Instrum. Anal."},{"key":"ref_24","doi-asserted-by":"crossref","first-page":"2011","DOI":"10.1109\/TPAMI.2019.2913372","article-title":"Squeeze-and-excitation networks","volume":"42","author":"Hu","year":"2020","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."},{"key":"ref_25","doi-asserted-by":"crossref","first-page":"110904","DOI":"10.1016\/j.asoc.2023.110904","article-title":"Hyperparameter optimization for convolutional neural network by opposite-based particle swarm optimization and an empirical study of photomask defect classification","volume":"148","author":"Hong","year":"2023","journal-title":"Appl. Soft Comput."},{"key":"ref_26","doi-asserted-by":"crossref","first-page":"7229","DOI":"10.1007\/s10489-024-05502-1","article-title":"Nonlinear crossing strategy-based particle swarm optimizations with time-varying acceleration coefficients","volume":"54","author":"Watanabe","year":"2024","journal-title":"Appl. Intell."},{"key":"ref_27","doi-asserted-by":"crossref","first-page":"109","DOI":"10.1016\/S0169-7439(01)00155-1","article-title":"PLS-regression: A basic tool of chemometrics","volume":"58","author":"Wold","year":"2001","journal-title":"Chemom. Intell. Lab. Syst."},{"key":"ref_28","doi-asserted-by":"crossref","first-page":"5","DOI":"10.1023\/A:1010933404324","article-title":"Random forests","volume":"45","author":"Breiman","year":"2001","journal-title":"Mach. Learn."},{"key":"ref_29","doi-asserted-by":"crossref","unstructured":"Vapnik, V.N. (1995). The Nature of Statistical Learning Theory, Springer.","DOI":"10.1007\/978-1-4757-2440-0"},{"key":"ref_30","doi-asserted-by":"crossref","unstructured":"Ribeiro, S.G., Teixeira, A.S., de Oliveira, M.R.R., Costa, M.C.G., Ara\u00fajo, I.C.d.S., Moreira, L.C.J., and Lopes, F.B. (2021). Soil organic carbon content prediction using soil-reflected spectra: A comparison of two regression methods. Remote Sens., 13.","DOI":"10.3390\/rs13234752"},{"key":"ref_31","doi-asserted-by":"crossref","unstructured":"Shen, L., Gao, M., Yan, J., Li, Z.-L., Leng, P., Yang, Q., and Duan, S.-B. (2020). Hyperspectral estimation of soil organic matter content using different spectral preprocessing techniques and PLSR method. Remote Sens., 12.","DOI":"10.3390\/rs12071206"},{"key":"ref_32","doi-asserted-by":"crossref","first-page":"32","DOI":"10.1016\/j.geoderma.2018.09.003","article-title":"Prediction of soil organic carbon stock by laboratory spectral data and airborne hyperspectral images","volume":"337","author":"Guo","year":"2019","journal-title":"Geoderma"},{"key":"ref_33","doi-asserted-by":"crossref","first-page":"127","DOI":"10.1177\/09670335241269014","article-title":"Non-linear machine learning coupled near-infrared spectroscopy enhanced model performance and insights for coffee origin traceability","volume":"32","author":"Sim","year":"2024","journal-title":"J. Near Infrared Spectrosc."},{"key":"ref_34","doi-asserted-by":"crossref","unstructured":"Rebecca, J.V., Bhogilal, H.V., Doug, A., Chen, S., Gillespie, A., Adamchuk, V., and Biswas, A. (2021). Evaluation of optimized preprocessing and modeling algorithms for prediction of soil properties using VIS\u2013NIR spectroscopy. Sensors, 21.","DOI":"10.3390\/s21206745"},{"key":"ref_35","unstructured":"Nalepa, J., Myller, M., and Kawulok, M. (2019). Hyperspectral data augmentation. arXiv."},{"key":"ref_36","doi-asserted-by":"crossref","unstructured":"Wang, W., Liu, X., and Mou, X. (2021). Data augmentation and spectral structure features for limited samples hyperspectral classification. Remote Sens., 13.","DOI":"10.3390\/rs13040547"}],"container-title":["Sensors"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/1424-8220\/26\/2\/741\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,1,27]],"date-time":"2026-01-27T05:20:17Z","timestamp":1769491217000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/1424-8220\/26\/2\/741"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,1,22]]},"references-count":36,"journal-issue":{"issue":"2","published-online":{"date-parts":[[2026,1]]}},"alternative-id":["s26020741"],"URL":"https:\/\/doi.org\/10.3390\/s26020741","relation":{},"ISSN":["1424-8220"],"issn-type":[{"value":"1424-8220","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,1,22]]}}}