Vision-spectroscopy cooperative sensing and cold atmospheric plasma inactivation for early warning and management of agricultural product diseases

Abstract Pathogen-induced disease causes heavy post-harvest losses and threatens food safety, yet routine inspection rarely identifies an infection before it becomes visible. This study evaluates multimodal sensing, threshold-based early warning and plasma-based decontamination for post-harvest disease management in strawberry (Fragaria × ananassa cv. ‘Hongyan’). A dual-branch network, pairing a ResNet-34 visual encoder with a one-dimensional convolutional spectral encoder through bidirectional cross-modal attention, fuses RGB imagery with near-infrared hyperspectral reflectance. Across five severity grades the fused model reached 93.5 ± 0.7% accuracy over fifteen cross-validation runs and 94.2% on a fruit-level held-out test set of 260 samples (95% CI 91.4 to 97.1), with a weighted F1 score of 0.938 and a Cohen’s Kappa of 0.927. Its 2.7-point margin over an unweighted late-fusion baseline did not reach significance after correction for eleven held-out comparisons (continuity-corrected McNemar $$\:{\chi\:}^{2}$$ = 4.00, Holm-adjusted $$\:p$$ = 0.140). Alarms preceded the expert-confirmed transition to a visible lesion by 52.0 ± 9.6 h in the 80-fruit monitoring cohort. Cold atmospheric plasma parameters were screened systematically: 1.5 W/cm 2 in Ar/O 2 for 180 s delivered more than a 3.5-log reduction for Botrytis cinerea , Colletotrichum gloeosporioides and Aspergillus niger ; the measured quality attributes did not differ significantly from untreated controls under the tested conditions. Leakage assays and a sigmoidal dose-response model support membrane damage associated with oxidative stress. The proposed progression from membrane peroxidation through enzyme disruption to DNA damage remains a mechanistic hypothesis. Sensing and treatment were evaluated separately; the present evidence is limited to one commodity, laboratory conditions and no integrated pilot-scale trial.

Authors

Publication Details

Journal
Scientific Reports
Published
2026-10-14
DOI
https://doi.org/10.1038/s41598-026-75069-z
Primary Topic
Smart Agriculture and AI
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Vision-spectroscopy cooperative sensing and cold atmospheric plasma inactivation for early warning and management of agricultural product diseases

Xin Hu, Wei Zhang, BaoJing Sun
Scientific Reports
Smart Agriculture and AI
article

Vision-spectroscopy cooperative sensing and cold atmospheric plasma inactivation for early warning and management of agricultural product diseases

Xin Hu, Wei Zhang, BaoJing Sun
article en

Abstract

Abstract Pathogen-induced disease causes heavy post-harvest losses and threatens food safety, yet routine inspection rarely identifies an infection before it becomes visible. This study evaluates multimodal sensing, threshold-based early warning and plasma-based decontamination for post-harvest disease management in strawberry (Fragaria × ananassa cv. ‘Hongyan’). A dual-branch network, pairing a ResNet-34 visual encoder with a one-dimensional convolutional spectral encoder through bidirectional cross-modal attention, fuses RGB imagery with near-infrared hyperspectral reflectance. Across five severity grades the fused model reached 93.5 ± 0.7% accuracy over fifteen cross-validation runs and 94.2% on a fruit-level held-out test set of 260 samples (95% CI 91.4 to 97.1), with a weighted F1 score of 0.938 and a Cohen’s Kappa of 0.927. Its 2.7-point margin over an unweighted late-fusion baseline did not reach significance after correction for eleven held-out comparisons (continuity-corrected McNemar $$\:{\chi\:}^{2}$$ = 4.00, Holm-adjusted $$\:p$$ = 0.140). Alarms preceded the expert-confirmed transition to a visible lesion by 52.0 ± 9.6 h in the 80-fruit monitoring cohort. Cold atmospheric plasma parameters were screened systematically: 1.5 W/cm 2 in Ar/O 2 for 180 s delivered more than a 3.5-log reduction for Botrytis cinerea , Colletotrichum gloeosporioides and Aspergillus niger ; the measured quality attributes did not differ significantly from untreated controls under the tested conditions. Leakage assays and a sigmoidal dose-response model support membrane damage associated with oxidative stress. The proposed progression from membrane peroxidation through enzyme disruption to DNA damage remains a mechanistic hypothesis. Sensing and treatment were evaluated separately; the present evidence is limited to one commodity, laboratory conditions and no integrated pilot-scale trial.

Scientific Reports
Zero hunger
Openalex Percentile: Top 15%
Smart Agriculture and AI
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.