Predictive framework of plant height in commercial cotton fields using a remote sensing and machine learning approach

Andréa, Maria Carolina da Silva; Nascimento, João Pedro F. de Oliveira; Mota, Fabrícia Conceição Menez; Oliveira, Rodrigo de Souza

doi:10.1016/j.atech.2022.100154

Cited by 8 publications

(8 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Despite using a linear model for ground elevation adjustment, their reported RMSE error for maize crop height remains around 14.17 cm. Meanwhile, our findings are in agreement with the study by Da Silva Andrea et al [19] which tested different machine learning algorithms combined with different manual feature selections using satellite multispectral imagery for cotton height prediction during the entire growing season and obtained MAE errors ranging from 8 to 25 cm. Moreover, Osco et al [18] reported RMSE errors between 17 cm and 30 cm for maize plant height predictions using drone multispectral images, combined with machine learning methodologies and manual feature extraction.…”

Section: Discussionsupporting

confidence: 92%

“…To be specific, Vegetation Indices (VIs), derived from mathematical formulations of canopy spectral reflectances, have a strong relation to specific characteristics in crops. Recently, some studies have used these specific VIs through machine learning methods to develop models for crop height prediction including wheat [17], maize [18], cotton [19], potatoes [20], and sunflower [21]. Unlike the traditional method that relies on preset models or equations, machine learning uses specific features as model input to calibrate and identify the most optimal parameters with the ground truth.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

G-DMD: A Gated Recurrent Unit-Based Digital Elevation Model for Crop Height Measurement from Multispectral Drone Images

Wang,

Oishi,

Birch

et al. 2023

Machines

View full text Add to dashboard Cite

Crop height is a vital indicator of growth conditions. Traditional drone image-based crop height measurement methods primarily rely on calculating the difference between the Digital Elevation Model (DEM) and the Digital Terrain Model (DTM). The calculation often needs more ground information, which remains labour-intensive and time-consuming. Moreover, the variations of terrains can further compromise the reliability of these ground models. In response to these challenges, we introduce G-DMD, a novel method based on Gated Recurrent Units (GRUs) using DEM and multispectral drone images to calculate the crop height. Our method enables the model to recognize the relation between crop height, elevation, and growth stages, eliminating reliance on DTM and thereby mitigating the effects of varied terrains. We also introduce a data preparation process to handle the unique DEM and multispectral image. Upon evaluation using a cotton dataset, our G-DMD method demonstrates a notable increase in accuracy for both maximum and average cotton height measurements, achieving a 34% and 72% reduction in Root Mean Square Error (RMSE) when compared with the traditional method. Compared to other combinations of model inputs, using DEM and multispectral drone images together as inputs results in the lowest error for estimating maximum cotton height. This approach demonstrates the potential of integrating deep learning techniques with drone-based remote sensing to achieve a more accurate, labour-efficient, and streamlined crop height assessment across varied terrains.

show abstract

Section: Discussionsupporting

confidence: 92%

Section: Introductionmentioning

confidence: 99%

G-DMD: A Gated Recurrent Unit-Based Digital Elevation Model for Crop Height Measurement from Multispectral Drone Images

Wang,

Oishi,

Birch

et al. 2023

Machines

View full text Add to dashboard Cite

show abstract

“…High temporal resolution data collection facilitates a better understanding of end of season traits that are the culmination of a growing season's worth of genotype-by-environment interactions (de Jesus Colwell et al, 2021). 5 Such temporal measurements have been completed on canopy cover and biomass in soybean (Moreira et al, 2019;Herrero-Huerta and Rainey, 2019;Li et al, 2022;Freitas Moreira et al, 2021), plant height, canopy cover, growth dynamics, and yield prediction in barley (Herzig et al, 2021), plant height and canopy cover in banana (Musa paradisiaca L.) (Aeberli et al, 2023), nutrient status, canopy cover, canopy volume, and plant height in potatoes (Liu et al, 2021;de Jesus Colwell et al, 2021), plant height in cotton (da Silva Andrea et al, 2023), and plant height in eggplant, tomato, cabbage (Moeckel et al, 2018). In maize, multi-temporal measurements of multiple traits have been used to explain the effect of phosphorus on plant growth and final yield (Pedersen et al, 2021;Herrmann et al, 2020), phenotypic variation due to genotypic background differences (Pugh et al, 2018;Han et al, 2019;Alper Adak, Anderson, et al, 2023;Adak, Murray, Božinović, et al, 2021;Adak et al, 2024), phenotypic variation due to environmental factors such as drought (Machado et al, 2002) or high-speed wind events (Tirado et al, 2021), and canopy cover (Jin et al, 2020;Lu et al, 2021).…”

Section: Introductionmentioning

confidence: 99%

Temporally resolved growth patterns reveal novel information about the polygenic nature of complex quantitative traits

Sweet,

Tirado,

Cooper

et al. 2024

Preprint

View full text Add to dashboard Cite

Plant height can be an indicator of plant health across environments and used to identify superior genotypes or evaluate abiotic stress factors. Typically plant height is measured at a single time point when plants have reached terminal height for the season. Evaluating plant height using unoccupied aerial vehicles (UAVs) is faster, allowing for measurements throughout the growing season, which facilitates a better understanding of plant-environment interactions and the genetic basis of this complex trait. To assess variation throughout development, plant height data was collected weekly for a panel of ∼500 diverse maize inbred lines over four growing seasons. The variation in plant height throughout the season was significantly explained by genotype, year, and genotype-by-year interactions to varying extents throughout development. Genome-wide association studies revealed significant SNPs associated with plant height and growth rate at different parts of the growing season specific to certain phases of vegetative growth that would not be identified by terminal height associations alone. When plant height growth rates were compared to growth rates estimated from canopy cover, greater Fréchet distance stability was observed in plant height growth curves than for canopy cover. This indicated canopy cover may be more useful for understanding environmental modulation of overall plant growth and plant height better for understanding genotypic modulation of overall plant growth. This study demonstrated that substantial information can be gained from high temporal resolution data to understand how plants differentially interact with the environment and can enhance our understanding of the genetic basis of complex polygenic traits.

show abstract

“…While this method was effective in a soybean field, collecting measurements 4 rows at a time is time-consuming and as maize gets much taller throughout the seasons measurements would be increasingly difficult to use this approach. In commercial cotton fields, plant height was estimated using satellite imagery containing multi-spectral bands and machine learning, with height estimates over large areas, averaging 15 measurements across 175 acre fields (da Silva Andrea et al, 2023). While this study was able to show some success in estimating plant height with R 2 values as high as 0.87, the study areas are too large for useful sub-field resolution information needed for precision farming.…”

Section: Introductionmentioning

confidence: 99%

Plant height defined growth curves during vegetative development have the potential to predict end of season maize yield and assist with mid-season management decisions

Sweet,

Cooper,

Hirsch

et al. 2024

Preprint

View full text Add to dashboard Cite

Precision farming has been developing with the intention of identifying within field variability to adjust management strategies and maximize end of season yield and profitability and minimize negative environmental impacts. The development of quick, easy, and low cost methods to quantify field level variation is essential to successful implementation of precision agriculture at scale. Temporal plant height and growth rates collected with unoccupied aerial vehicles mounted with red, green, blue sensors have the potential to predict end of season grain yield, which could facilitate mid-season management decisions. Image-based plant height data was collected weekly from commercial maize fields in three growing seasons to assess variation within fields and the relationship with grain yield variation. Plant height, growth rate, and grain yield had variable relationships depending on the time point and growth environment. Models developed using temporal traits predicted grain yield variation within a commercial field up to r = 0.7, though insufficient water affected the prediction accuracy in one field due to the limited representation of drought environments in the model development. In the future, with more data from stress environments, such as drought, this method has potential for high accuracy grain yield prediction across a range of environmental conditions. This study demonstrates the potential of using unoccupied aerial vehicles to derive vegetative growth patterns and model within field variations, and has application in making mid-season management decisions.

show abstract

Predictive framework of plant height in commercial cotton fields using a remote sensing and machine learning approach

Cited by 8 publications

References 34 publications

G-DMD: A Gated Recurrent Unit-Based Digital Elevation Model for Crop Height Measurement from Multispectral Drone Images

G-DMD: A Gated Recurrent Unit-Based Digital Elevation Model for Crop Height Measurement from Multispectral Drone Images

Temporally resolved growth patterns reveal novel information about the polygenic nature of complex quantitative traits

Plant height defined growth curves during vegetative development have the potential to predict end of season maize yield and assist with mid-season management decisions

Contact Info

Product

Resources

About