Drones and artificial intelligence for hemp and miscanthus phenotyping

Ability drones lift data in high resolution and subsequent useintelligence artificial data analysis enabled the rapid expansion of these new technologies agriculture. Several scientific studies have highlighted the potential of these tools in estimation, mapping and characterization phenotypic traits crops quickly, accurately and non-destructively.

Specifically, this is the main focus of this studies OF phenotyping was a development algorithms artificial intelligence to guess some traits phenotypic genotypes hemp AND miscanthus using images multispectral acquired drone.

Drone phenotyping study objectives

Drone phenotyping study objectives

(Source: Giorgio Impolonia)

The images taken by drone and data collected in the field was used in the development algorithms artificial intelligence with a goal reckon phenotypic traits such as content humiditycontent leaf chlorophyll el’leaf surface index.

The algorithms were used on aerial photos in order to characterize the dynamics of phenotypic traits during the phase vegetative growth and aging. This analysis can be used to evaluate grow and Production different genotypes from hemp and miscanthus v land marginalor to choose the most suitable genotype for a certain environment.

Leaf area index (Lai) and leaf chlorophyll content (Lcc) maps of cannabis estimated from drone multispectral imagery

Leaf area index (Lai) and leaf chlorophyll content (Lcc) maps of cannabis estimated from drone multispectral imagery

(Source: Giorgio Impolonia)

Another relevant application aspect of this study concerns respect content humidity from biomass miscanthuswhich represents a qualitative characteristic very relevant with different impacts on collection, transport and storage. Through this estimation, it is possible Choose optimal moment collection and identify field with content lower humidity.

They were also approached two limits technicians related to the use of drones and artificial intelligence algorithms, Asinteroperability of sensors multispectral and portability of algorithms artificial intelligence.

L’interoperability of sensors it is a relevant aspect to be evaluated in situations where same crop is analyzed using sensors s different Properties ghostly. This can result in mistakes in feature estimation, so the developed algorithm sensor specific.

Senescence dynamics of different miscanthus hybrids based on the difference in estimated moisture content with the reference hybrid (M. x giganteus - Grc 9)

Senescence dynamics of different miscanthus hybrids based on the difference in estimated moisture content with the reference hybrid (M. x giganteus – Grc 9)

(Source: Giorgio Impolonia)

PUSH portability of Artificial intelligence algorithms were evaluated for their ability reckon strokes carefully environment AND season different compared to the one used to train the algorithm.

This study contributed to go into detail knowledge of use multispectral images obtained by drone for phenotyping applications and demonstrated that possible estimate i phenotypic traits of hemp and miscanthus with good accuracy. Especially through progresses specifications are possible overcome probleminteroperability by which we will make him algorithms artificial intelligence No employees of sensor used.

THE Result they also emphasized that number data a distribution of the feature values ​​used in training algorithms are some important features get one good portability of algorithms.

The future perspectives in the use of drones and artificial intelligence in crop phenotyping are very promising. Technologies are improving, i.e drones Hello spectral sensors they are increasingly accurate and powerful and therefore future generations drones could be collected data even more detailed.

with continuous development AND decrease of costs with drone and artificial intelligence technology, their use is likely technology you always become more accessibleenabling greater adoption of drone use for phenotyping crops on large scale.

Giorgio Impollonia, Precision agriculture category: data analysis and sharing

Giorgio Impollonia, category “Precision agriculture: data analysis and sharing”

(Source: Giorgio Impolonia)

Download it the complete works of Giorgio Impollonia

For any contacts giorgio.impollonia@unicatt.it

Edited by Giorgio Impollonia


AgroInnovation Award is a diploma award established by Image Line in collaboration with the Accademia dei Georgofili to promote the spread of innovative approaches, digital tools and the use of the Internet in agriculture.

AgroInnovation Award

Read winning papers of the sixth edition.

Leave a Comment