Estimation of Potato LAI Using UAV Multispectral and Multiband
Combined Textures
GUO Li-xiao1, 2, CHEN Zhi-chao1*, MA Yan-peng1, 2, BIAN Ming-bo1, 2, FAN Yi-guang2, CHEN Ri-qiang2, LIU Yang2, FENG Hai-kuan2, 3*
1. School of Surveying and Land Information Engineering, Henan Polytechnic University, Jiaozuo 454000, China
2. Key Laboratory of Quantitative Remote Sensing in Agriculture of Ministry of Agriculture and Rural Affairs, Information Technology Research Center, Beijing Academy of Agriculture and Forestry Sciences, Beijing 100097, China
3. National Engineering and Technology Center for Information Agriculture, Nanjing Agricultural University, Nanjing 210095, China
Abstract：The leaf area index (LAI) is an important indicator for characterizing crop growth, so an efficient and accurate estimation of crop LAI can guide field production management. Spectral features can provide information about the reflected and absorbed wavelengths of crops, while texture features can provide information about the gray-scale attributes and spatial location relationships of crops. Previous studies have shown some limitations in estimating crop LAI using only spectral features, and at high LAI levels, the “saturation phenomenon” occurs, resulting in an underestimation of LAI. To fully explore the information of multispectral images from UAVs, texture information of multiple bands was combined to obtain multiband combined texture and to explore whether the fusion of spectral features with multiband combined texture can improve the accuracy of LAI estimation. Firstly, we obtained multispectral data and ground-truthed LAI data of three key fertility stages of potato; then we extracted the texture features of each fertility stage using the gray-level co-occurrence matrix (GLCM) and combined the texture features of multiple bands; then we analyzed the correlation between the vegetation index, the texture features, and the multi-band combination of textures and LAI, and synthesized the correlations and correlations with LAI, and investigated whether the fusion of spectral information and multiband combination of textures could improve the accuracy of LAI estimation. Then, we analyzed the correlation between vegetation index, texture features, and multiband combined texture and LAI and combined the correlation and variance expansion factors to select the preferred vegetation index; finally, we integrated the multiband combined texture and used partial least squares regression (PLSR), ridge regression (RR) and K-nearest neighbors regression (KNR) with parameter tuning to determine the correlation between the vegetation index and LAI, and then used KNR to estimate the correlation between the vegetation index and LAI. KNR will estimate potato LAI at each fertility stage and compare it with the model using only the vegetation index to verify the feasibility of inverting LAI using a multiband combined texture. The results showed that: (1) the correlation between single-band texture, two-band combined texture and three-band combined texture and LAI increased sequentially; (2) the preferred multiband combined texture at each fertility stage of potato showed highly significant correlation with LAI, with correlation coefficients ranging from 0.79 to 0.83; and (3) compared with the model using only the vegetation index, the addition of the multiband combined texture could significantly increase the model's accuracy and stability. The KNR model had the highest accuracy in estimating potato LAI during the tuber formation period, with a modeling R2 of 0.83，an RMSE of 0.23 m2·m-2, and a validation R2 of 0.75 and an RMSE of 0.25 m2·m-2; the PLSR model had the highest accuracy during the tuber growth period, with a modeling R2 of 0.73 and an RMSE of 0.26 m2·m-2, and a validation R2 of 0.87 and an RMSE of 0.20 m2·m-2; and the PLSR model had the highest accuracy during the starch accumulation period, with a modeling R2 of 0.73 and an RMSE of 0.26 m2·m-2; and the PLSR model had the highest accuracy during the starch accumulation period. The PLSR model had the highest estimation accuracy, with modeling R2 of 0.73 and RMSE of 0.31 m2·m-2, and validation R2 of 0.84 and RMSE of 0.25 m2·m-2. This method can provide a reference for the UAV multispectral combination of texture features to estimate potato LAI.
Key words：UAV; Multispectral remote sensing; Multiband combined texture; Leaf area index
[1] LUO Qi-you, GAO Ming-jie, ZHANG Shuo, et al(罗其友，高明杰，张 烁，等). Chinese Journal of Agricultural Resources and Regional Planning(中国农业资源与区划), 2021, 42(7): 1.
[2] Yang Guijun, Liu Jiangang, Zhao Chunjiang, et al. Frontiers in Plant Science, 2017, 8: 1111.
[3] SU Wei, HOU Ning, LI Qi, et al(苏 伟，侯 宁，李 琪，等). Transactions of the Chinese Society for Agricultural Machinery(农业机械学报), 2018, 49(1): 151.
[4] Li Z, Li Z, Huang H, et al. PhotoniX 4, 23(2023). https://doi.org/10.1186/s43074-023-00100-9.
[5] Lv J, Shen S, Chen L, et al. PhotoniX 4, 28(2023). https://doi.org/10.1186/s43074-023-00108-1.
[6] WANG Xiao-xuan, LU Xiao-ping, MENG Qing-yan, et al(王枭轩，卢小平，孟庆岩，等). Spectroscopy and Spectral Analysis(光谱学与光谱分析), 2022, 42(7): 2278.
[7] MA Jun-wei, CHEN Peng-fei, SUN Yi, et al(马俊伟，陈鹏飞，孙 毅，等). Acta Agronomica Sinica(作物学报), 2023, 49(12): 3364.
[8] FAN Jun-liang, WANG Han, LIAO Zhen-qi, et al(范军亮，王 涵，廖振棋，等). Transactions of the Chinese Society for Agricultural Machinery(农业机械学报), 2023, 54(7): 347.
[9] HANG Yan-hong, SU Huan, YU Zi-yang, et al(杭艳红，苏 欢，于滋洋，等). Transactions of the Chinese Society of Agricultural Engineering(农业工程学报), 2021, 37(9): 64.
[10] GU Xiao-bo, CHENG Zhi-kai, ZHOU Zhi-hui, et al(谷晓博，程智楷，周智辉，等). Transactions of the Chinese Society for Agricultural Machinery(农业机械学报), 2023, 54(6): 148.
[11] FENG Hai-kuan, YUE Ji-bo, FAN Yi-guang, et al(冯海宽，岳继博，樊意广，等). Spectroscopy and Spectral Analysis(光谱学与光谱分析), 2023, 43(9): 2876.
[12] Ma Yanpeng, Chen Zhichao, Fan Yiguang, et al. Frontiers in Plant Science, 2023, 14: 1265132.
[13] Liu Yang, Feng Haikuan, Yue Jibo, et al. Computers and Electronics in Agriculture, 2023, 213: 108229.
[14] Meng Lin, Yin Dameng, Cheng Minghan, et al. Drones, 2023, 7(4): 254.
[15] Susan Barati, Behzad Rayegani, Mehdi Saati, et al. The Egyptian Journal of Remote Sensing and Space Science, 2011, 14(1): 49.
[16] Sui Juan, Qin Qiming, Ren Huazhong, et al. IEEE International Geoscience and Remote Sensing Symposium, 2018, 7354(https://doi.org/10.1109/IGARSS.2018.8517909).
[17] Sripada R P, Heiniger R W, White J G, et al. Agronomy Journal, 2006, 98(4): 968.
[18] Van Beek J, Tits L, Somers B, et al. Remote Sensing, 2014, 6(2): 1760.
[19] Erdle K, Mistele B, Schmidhalter U. Field Crops Research, 2011, 124(1): 74.
[20] Sun Xinkai, Yang Zhongyu, Su Pengyan, et al. Frontiers in Plant Science, 2023, 14: 1158837.
[21] Ustin S L, Jacquemoud S. Remote Sensing of Plant Biodiversity, 2020, 349(https://doi.org/10.1007/978-3-030-33157-3_14).
[22] YANG Qi, YE Hao, HUANG Kai, et al(杨 琦，叶 豪，黄 凯，等). Transactions of the Chinese Society of Agricultural Engineering(农业工程学报), 2017, 33(8): 104.
