You are on page 1of 16

Athens, Greece

whispers Oct 31 – Nov 02, 2023

Sub-pixel Discrimination of Soil and Crop in


Drone-based Hyperspectral Imagery

Manohar Kumar C. V. S. S.
Rama Rao Nidamanuri
Vinay Kumar Dadhwal

Department of Earth and Space Sciences


Indian Institute of Space Science Technology
Thiruvananthapuram, India
Contents
Crop and Soil Methodology

Is it necessary? Unmixing models

Objective Results and Analysis

Study area Conclusion and future scope

Spectral uniqueness References & Acknowledgements


Crop and Soil: Remote Sensing
Plant: Soil:
Chlorophyll Organic Carbon
Leaf area Organic Matter
Biomass Nitrogen
Disease Phosphorous
pH
Moisture

Source: https://ohioline.osu.edu/factsheet/fabe-5541
Motivation:
Literature:
Mapping/Classifications

Space-borne (30m)

Puletti, N., Camarretta, N. and Corona, P., 2016

Oil Spill Detection


Airborne (4-8m)

AISA hyperspectral image


Objective:

The main objective of study is to discriminate


the crop and soil at the level of the plant
canopy
Study Area & drone HSI:

Imaging Sensor : FIREFLEYE 185

Flight Altitude : 40m

Spectral sampling : 4 nm

Spectral range : 450 – 950 nm

Spectral bands : 139


Methodology:
Spectral Unmixing is a solution or not: mapping of microplastic?

Ideal Case Mixing


Scenario (Linear)

Macroscopic View
(25 % of each Component A,B,C,D)

Bioucas-Dias, José M., et al. 2013


Results: Abundance of cabbage plant and soil
CLS SUnSAL FCLS CLSUnSAL FAN_BL GBM MLMM PPPNM HAPKE
Abundance

Image
Cabbage

Field

Image
Abundance
Soil

Field
Results: Abundance of cabbage plant and soil
Areal abundance (100%) of cabbage plant and soil Areal abundance (0%) of cabbage plant and soil
using two different endmember sources at using two different endmember sources at
geographical location geographical location
Field spectral library Image spectral library Field spectral library Image spectral library
Crop Soil Crop Soil Crop Soil Crop Soil
CLS 58.532 132.095 98.335 89.332 CLS 88.11 0.38 9.558 3.965
SUnSAL 58.533 132.094 98.335 89.331 SUnSAL 88.107 0.38 9.557 3.965
FCLS 85.123 84.633 91.14 96.185 FCLS 14.877 15.367 8.86 3.815
CLSUnSAL 85.123 84.633 91.14 96.185 CLSUnSAL 14.877 15.367 8.86 3.815
FAN_BL 69.459 86.309 81.738 97.403 FAN_BL 30.541 13.691 18.262 2.597
GBM 73.801 86.342 89.541 96.178 GBM 26.199 13.658 10.459 3.822
PPNM 25.815 99.927 86.881 90.851 PPNM 74.185 0.073 13.119 9.149
MLMM 58.77 99.099 92.047 96.422 MLMM 41.23 0.901 7.953 3.578
HAPKE 5.351 99.906 77.308 84.612 HAPKE 94.649 0.094 22.692 15.388
Results: RMSE and Validation
C S FC CL FA G M P H
0.2

Image
RMSE

Field

C: CLS
S: SUnSAL
FC: FCLS
CL : CLSUnSAL
FA: FAN_BL
G: GBM
M: MLMM
P: PPNM
Source of endmember: Field Image H: HAPKE
Reference value: Mean SRE
Conclusion & Future Scope:
➢ Imaging spectral library is superior to that of the field spectral library for the cabbage
crop, achieving an accuracy rate of 98.3%

➢ Field spectral library has the highest performance for the soil, with an accuracy rate of
99.9%

➢ To understand soil nutrients and plant phenology independently for the effective
utilization of soil to increase the productivity of the crop with very less external input

➢ Soil and crop discrimination with the integration of LiDAR at plant level understand
the health, growth, and infection at subpixel level
References:
Sishodia, R. P., Ray, R. L., & Singh, S. K. (2020). Applications of remote sensing in precision agriculture: A
review. Remote Sensing, 12(19), 3136.

Shafi, U., Mumtaz, R., García-Nieto, J., Hassan, S. A., Zaidi, S. A. R., & Iqbal, N. (2019). Precision agriculture techniques
and practices: From considerations to applications. Sensors, 19(17), 3796.

A. Bégué et al., ‘Remote Sensing and Cropping Practices: A Review’, Remote Sens., vol. 10, no. 2, p. 99, Jan. 2018

S. D. Suchi, A. Menon, A. Malik, J. Hu, and J. Gao, ‘Crop Identification Based on Remote Sensing Data using Machine
Learning Approaches for Fresno County, California’, in 2021 IEEE Seventh International Conference on Big Data
Computing Service and Applications (BigDataService), Oxford, United Kingdom: IEEE, Aug. 2021, pp. 115–124.

L. Hashemi-Beni, A. Gebrehiwot, A. Karimoddini, A. Shahbazi, and F. Dorbu, ‘Deep Convolutional Neural Networks for
Weeds and Crops Discrimination From UAS Imagery’, Front. Remote Sens., vol. 3, p. 755939, Feb. 2022.

N. Keshava, ‘A Survey of Spectral Unmixing Algorithms’, Linc. Lab. J., vol. 14, no. 1, pp. 55–78, 2003.

R. Heylen, V. Andrejchenko, Z. Zahiri, M. Parente, and P. Scheunders, ‘Nonlinear Hyperspectral Unmixing With
Graphical Models’, IEEE Trans. Geosci. Remote Sens., vol. 57, no. 7, pp. 4844–4856, Jul. 2019.
Acknowledgements

➢ Council of Scientific and Industrial Research (CSIR), Govt. of India

➢ Department of Science and Technology (DST), Govt. of India

➢ Department of Biotechnology (DBT), Govt. of India

➢ whispers 2023
Thank you…!

You might also like