Abstract
• First PRISMA-ScR mapping of 79 HSI-ML asymptomatic detection studies (42 species, 74 pathogens) • Controlled-to-field accuracy gap quantified: 91.4% vs. 86.3% (5.1 pp, p = 0.0163 ) • 56.8% of studies omit temporal sampling documentation (CV = 139%) • SWIR underutilization (11.8%) reflects economic, not scientific, barriers • DBVS proposed as standardized temporal metric for cross-study comparability Plant disease management requires non-invasive detection methods capable of identifying infections before visible symptom manifestation, thereby enabling timely intervention. Hyperspectral imaging combined with machine learning and deep learning (HSI-ML) achieves 90.2% classification accuracy in controlled environments for asymptomatic plant detection; however, systematic characterization of methodological practices across this rapidly expanding field remains absent. This PRISMA-ScR compliant scoping review mapped 79 peer-reviewed studies (2010–2025) encompassing 42 plant species and 74 pathogenic agents using a Population-Concept-Context framework. Visible-near-infrared (VNIR) systems dominated deployment (61.8%, n = 49 ), while short-wave infrared (SWIR) systems remained substantially underutilized (11.8%, n = 9 ) due primarily to economic rather than scientific constraints. Among 67 unique algorithms identified, machine learning methods accounted for 30.7% (SVM, random forests, and PLS-DA predominant), whereas deep learning represented 28.4% (2D-CNN, 3D-CNN, and hybrid architectures). Critical methodological gaps emerged: 56.8% of studies omitted temporal sampling documentation (detection latency range: 1–56 days post-inoculation; coefficient of variation = 139%). Platform-stratified analysis revealed controlled environments achieved 91.4% ± 6.6% classification accuracy ( n = 48 ) versus 86.3% ± 9.1% for field/UAV deployments ( n = 26 ), representing a significant 5.1 percentage-point performance decrease ( p = 0.0163 ). Detection accuracy exhibited a weak negative correlation with detection timing ( ρ = − 0.33 , p = 0.067 ), though this association did not reach conventional statistical significance. Methodological heterogeneity—rather than algorithmic limitations—constitutes the primary barrier to field operationalization. Adoption of Days Before Visible Symptoms (DBVS) as a standardized temporal metric could resolve an estimated 40–50% of cross-study variance currently attributed to inconsistent asymptomatic-phase definitions.
Abstract from the paper’s OpenAlex record; open copy licensed cc-by-nc-nd. Read at the source ↗
The text the model read, with the source record’s inline markup (italics, sub- and superscripts) rendered and its character entities decoded; quote offsets refer to the raw text.
Claims 12
Claim · model-written summary
VNIR systems were the dominant sensor type in the reviewed asymptomatic detection studies (61.8%)
Verbatim
e0“encompassing 42 plant species and 74 pathogenic agents”
e1“Visible-near-infrared (VNIR) systems dominated deployment (61.8%, n = 49 )”
Claim · model-written summary
SWIR systems were substantially underutilized (11.8%), mainly for economic rather than scientific reasons
Verbatim
e0“encompassing 42 plant species and 74 pathogenic agents”
e1“short-wave infrared (SWIR) systems remained substantially underutilized (11.8%, n = 9 ) due primarily to economic rather than scientific constraints”
Claim · model-written summary
Machine learning methods accounted for 30.7% of the algorithms identified
Verbatim
e0“encompassing 42 plant species and 74 pathogenic agents”
e1“machine learning methods accounted for 30.7% (SVM, random forests, and PLS-DA predominant)”
Claim · model-written summary
Deep learning accounted for 28.4% of the algorithms identified
Verbatim
e0“encompassing 42 plant species and 74 pathogenic agents”
e1“deep learning represented 28.4% (2D-CNN, 3D-CNN, and hybrid architectures)”
Claim · model-written summary
Of the reviewed studies, 56.8% did not document temporal sampling
Verbatim
e0“encompassing 42 plant species and 74 pathogenic agents”
e1“56.8% of studies omitted temporal sampling documentation”
Claim · model-written summary
Detection latency varied widely across studies, with a coefficient of variation of 139%
Verbatim
e0“encompassing 42 plant species and 74 pathogenic agents”
e1“detection latency range: 1–56 days post-inoculation; coefficient of variation = 139%”
Claim · model-written summary
In controlled environments, HSI-ML classification accuracy averaged 91.4% ± 6.6%
Verbatim
e0“encompassing 42 plant species and 74 pathogenic agents”
e1“controlled environments achieved 91.4% ± 6.6% classification accuracy ( n = 48 )”
Claim · model-written summary
In field and UAV deployments, HSI-ML classification accuracy averaged 86.3% ± 9.1%
Verbatim
e0“encompassing 42 plant species and 74 pathogenic agents”
e2“classification accuracy ( n = 48 ) versus 86.3% ± 9.1% for field/UAV deployments ( n = 26 )”
Claim · model-written summary
Field and UAV deployments showed a significant 5.1 percentage-point decrease in classification accuracy compared with controlled environments
Verbatim
e0“encompassing 42 plant species and 74 pathogenic agents”
e1“controlled environments achieved 91.4% ± 6.6% classification accuracy ( n = 48 )”
e2“classification accuracy ( n = 48 ) versus 86.3% ± 9.1% for field/UAV deployments ( n = 26 )”
e3“representing a significant 5.1 percentage-point performance decrease ( p = 0.0163 )”
Claim · model-written summary
Detection accuracy showed a weak negative correlation with detection timing, not statistically significant
Verbatim
e0“encompassing 42 plant species and 74 pathogenic agents”
e1“Detection accuracy exhibited a weak negative correlation with detection timing ( ρ = − 0.33 , p = 0.067 ), though this association did not reach conventional statistical significance”
Claim · model-written summary
Adopting Days Before Visible Symptoms (DBVS) as a standardized temporal metric could resolve an estimated 40–50% of cross-study variance
Verbatim
e0“encompassing 42 plant species and 74 pathogenic agents”
e1“Adoption of Days Before Visible Symptoms (DBVS) as a standardized temporal metric could resolve an estimated 40–50% of cross-study variance”
Claim · model-written summary
Methodological heterogeneity, rather than algorithmic limitations, is the primary barrier to field operationalization of HSI-ML
Verbatim
e0“encompassing 42 plant species and 74 pathogenic agents”
e1“Methodological heterogeneity—rather than algorithmic limitations—constitutes the primary barrier to field operationalization”