基于不同卷积尺度融合与近红外光谱的土地分类模型研究

doi:10.3964/j.issn.1000-0593(2023)02-0460-08

Abstract
References
Related Citation (15)

Download: PDF (6924 KB)
Export: BibTeX | EndNote (RIS)

Abstract In order to improve the accuracy of the land cover classification model based on near-infrared spectroscopy, the soil near-infrared spectrum data released by Eurostat was adopted as the research object in this study, and the land cover classification model based on short-time Fourier transform (STFT) preprocessing method and different convolution scale fusion are studied to achieve the rapid identification of cultivated land, forest land and grassland. In order to meet the requirements of two-dimensional convolution, 4200 wavelength points in the 400~2 500 nm band of the one-dimensional spectrum were transformed into two-dimensional images by SFTF, so the spectrum information about spectral data was extractedfor the later modeling. The samples were randomly divided into the training set, validation set and test set according to the ratio of 6∶2∶2. Two kinds of convolution neural network(CNN) models, including the single-size kernel CNN and themulti-size kernel fusion CNN, were established to classify the land cover. The CNN models adopted the ReLU activation function, batch normalization (BN), and dropout methods to prevent the model’s gradient disappearance. The Early Stopping method trains the network to prevent the model from overfitting. Firstly, this paper discussed the influence of different STFT window lengths (64, 100 and 128) and convolution kernel sizes (3×3, 5×5 and 7×7) on the classification effect of the different models. The experimental results show that the overall classification accuracy of all model shit a high point when the STFT window length was 100 and the window overlap length was 50%. The model’s classification accuracy decreases with the convolution kernel size. That is, the accuracy of the model with a smaller convolution kernel size is relatively higher. The overall classification accuracy of the CNN model with a convolution kernel size of 3×3 reaches 78.76%, which is higher than that of the CNN model with a convolution kernel size of 5×5 and 7×7. The CNN models with different convolution kernel sizes have good classification results for a certain land cover type.That is, the model with 3×3 convolution size has the best classification effect for cultivated land, 5×5 convolution size has the best classification effect for woodland, 7×7 convolution size has the best classification effect for grassland. Secondly, a fusion-CNN model based on the hybrid of multi-size convolution kernels is proposed. The model integrates the advantages of different size convolution kernels. The classification accuracy of the fusion-CNN for three land cover types has been improved to varying degrees, with an overall classification accuracy of 84.39%. The Fusion-CNN model overcomes the shortcomings of the single size convolution kernel CNN model, such as a long selection period for appropriate convolution kernel size and cumbersome parameter adjustment steps, and can simplify and speed up the modeling process. Fusion-CNN convolution fusion network can more effectively extract the internal characteristic information on soil near-infrared spectroscopy to obtain high and stable land cover classification accuracy.

Key words：Near infrared spectroscopy; Convolution fusion; Convolution neural network; Land cover classification; Short-time Fourier transform

Received: 2021-08-26 Accepted: 2022-05-27

ZTFLH:

O657.33

Corresponding Authors: CHEN Zheng-guang E-mail: ruzee@sina.com

Cite this article:

WEI Jin-shan,CHEN Zheng-guang,JIAO Feng. Research on Land Classification Model Based on Fusion of Different Convolution Scales and Near-Infrared Spectroscopy[J]. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2023, 43(02): 460-467.

URL:

https://www.gpxygpfx.com/EN/10.3964/j.issn.1000-0593(2023)02-0460-08 OR https://www.gpxygpfx.com/EN/Y2023/V43/I02/460

[1] Foody G M. Remote Sensing of Environment, 2002, 80(1): 185.
[2] LI Zhen, ZHANG Yan, YANG Song, et al(李镇, 张岩, 杨松, 等). Transactions of the Chinese Society of Agricultural Engineering(农业工程学报), 2014, 30(20): 179.
[3] WANG Zhi-jie(王志杰). Journal of Nanjing Forestry University·Natural Sciences Edition(南京林业大学学报·自然科学版), 2015, 39(3): 1.
[4] HUANG Li-wen, MAO Zheng-yuan, LI Er-zhen, et al(黄利文, 毛政元, 李二振, 等). Journal of Image and Graphics(中国图象图形学报), 2007，(4): 633.
[5] LI Lin, TIAN Xin, WENG Yong-ling, et al(李林, 田馨, 翁永玲, 等). Journal of Southeast University·Natural Science Edition(东南大学学报·自然科学版), 2021, 51(3): 529.
[6] ZHANG Ying, ZHANG Xiao-li, LI Hong-zhi, et al(张莹, 张晓丽, 李宏志, 等). Scientia Silvae Sinicae(林业科学), 2018, 54(8): 88.
[7] Zhang F, Yang X. Remote Sensing of Environment, 2020, 251: 112105.
[8] Talukdar S, Singha P, Mahato S, et al. Remote Sensing, 2020, 12(7): 1135.
[9] Lepers E, Lambin E F, Janetos A C, et al. BioScience, 2005, 55(2): 115.
[10] DENG Wan-gang, WU Wei-dong, CHEN Ming-zhi, et al(邓万刚, 吴蔚东, 陈明智, 等). Ecology and Environmental Sciences(生态环境), 2008，(3): 1130.
[11] Zhang X, Liu H, Zhang X, et al. Geoderma, 2018, 320: 12.
[12] Xie H, Zhao J, Wang Q, et al. Scientific Reports, 2015, 5(1): 1.
[13] Li X, Fan P, Li Z, et al. Journal of Spectroscopy, 2021, 2021: 1508267.
[14] CHAI Hua-bin, YAN Chao, ZOU You-feng, et al(柴华彬, 严超, 邹友峰, 等). Geomatics and Information Science of Wuhan University(武汉大学学报·信息科学版), 2021, 46(8): 1224.
[15] SHI Zhou, WANG Qian-long, PENG Jie, et al(史舟, 王乾龙, 彭杰, 等). SIENTIA SINICA：Terrae(中国科学: 地球科学), 2014, 44(5): 978.
[16] Long J, Shelhamer E, Darrell T. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2015: 3431.
[17] CHEN Chao, QI Feng(陈超, 齐峰). Computer Science(计算机科学), 2019, 46(3): 63.
[18] LI Ji-long, DUAN Xiang-yang, FAN Chen, et al(李基隆，段向阳，范忱，等). Infrared and Laser Engineering(红外与激光工程), 2021, 50(7): 88.
[19] ZHENG Yuan-pan, LI Guang-yang, LI Ye(郑远攀, 李广阳, 李晔). Computer Engineering and Applications(计算机工程与应用), 2019, 55(12): 20.
[20] ZHOU Fei-yan, JIN Lin-peng, DONG Jun(周飞燕, 金林鹏, 董军). Chinese Journal of Computers(计算机学报), 2017, 40(6): 1229.
[21] LI Hai, LI Yang, ZUO Zheng-rong(李海, 李洋, 左峥嵘). Journal of Infrared and Millimeter Waves(红外与毫米波学报), 2021, 40(3): 369.
[22] He K M, Sun J. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2015: 5353.
[23] Wu L, Jia X Y, Zhu C S. Journal of Physics: Conference Series, 2020, 1629(1): 012002.
[24] LI Qing-xu, WANG Qiao-hua, MA Mei-hu, et al(李庆旭, 王巧华, 马美湖, 等). Spectroscopy and Spectral Analysis(光谱学与光谱分析), 2021, 41(6): 1800.
[25] Padarian J, Minasny B, Mcbratney A B. Geoderma, 2019, 340: 279.
[26] Huang Y L, Meng S Y, Zhao P, et al. Applied Optics, 2019, 58(18): 5122.