Processing Pipeline of Sugarcane Spectral Response to Characterize the Fallen Plants Phenomenon

20th Argentinean Bioengineering Society Congress

物理学;生物科学

Solano, Agustin^1 ; Kemerer, Alejandra^2 ; Hadad, Alejandro^1

Information Systems Laboratory, Faculty of Engineering, National University of Entre Rios, Argentina^1

INTA-EEA Paraná, Natural Resources and Abiotic Factors Group, Faculty of Agronomy, National University of Entre Rios, Argentina^2

关键词: Cultivated lands;    Electromagnetic spectra;    Extraction process;    Homogeneous regions;    Multispectral images;    Remote sensing tools;    Segmentation results;    Spectral response;   

Others  :  https://iopscience.iop.org/article/10.1088/1742-6596/705/1/012025/pdf
DOI  :  10.1088/1742-6596/705/1/012025

学科分类：生物科学（综合）

来源: IOP

【 摘 要 】

Nowadays, in agronomic systems it is possible to make a variable management of inputs to improve the efficiency of agronomic industry and optimize the logistics of the harvesting process. In this way, it was proposed for sugarcane culture the use of remote sensing tools and computational methods to identify useful areas in the cultivated lands. The objective was to use these areas to make variable management of the crop. When at the moment of harvesting the sugarcane there are fallen stalks, together with them some strange material (vegetal or mineral) is collected. This strange material is not millable and when it enters onto the sugar mill it causes important looses of efficiency in the sugar extraction processes and affects its quality. Considering this issue, the spectral response of sugarcane plants in aerial multispectral images was studied. The spectral response was analyzed in different bands of the electromagnetic spectrum. Then, the aerial images were segmented to obtain homogeneous regions useful for producers to make decisions related to the use of inputs and resources according to the variability of the system (existence of fallen cane and standing cane). The obtained segmentation results were satisfactory. It was possible to identify regions with fallen cane and regions with standing cane with high precision rates.