Pest Detection in Edible Crops at the Edge: An Implementation-Focused Review of Vision, Spectroscopy, and Sensors

TL;DR

This paper reviews pest detection methods for edible crops, comparing vision/AI, spectroscopy, and sensors based on performance, cost, and implementability to guide practical deployment.

Contribution

The paper introduces a modality-aware PCI rubric and decision maps to evaluate and compare pest detection systems for real-world deployment.

Findings

Vision/AI and sensor systems showed higher deployment-leaning PCI scores compared to spectroscopy.

Decision maps were developed to help practitioners choose suitable pest detection modalities based on deployment constraints.

Inter-rater agreement was substantial for sensors and spectroscopy but modest for vision/AI.

Abstract

What are the main findings?
We introduced a modality-aware PCI rubric (performance–cost–implementability) with inter-rater κ to compare vision/AI, spectroscopy, and indirect sensor systems for pest detection in edible crops.We derived compact decision maps that translate PCI evidence into field-ready choices under the constraints of power, cost, maintenance, connectivity, and required action granularity.

We introduced a modality-aware PCI rubric (performance–cost–implementability) with inter-rater κ to compare vision/AI, spectroscopy, and indirect sensor systems for pest detection in edible crops.

We derived compact decision maps that translate PCI evidence into field-ready choices under the constraints of power, cost, maintenance, connectivity, and required action granularity.

What is the implication of the main finding?
Practitioners can choose fit-for-purpose sensing modalities beyond accuracy-only benchmarks, improving the time to deployment.Reporting a minimum PCI metadata set enables reproducible, deployment-oriented comparisons across future studies.

Practitioners can choose fit-for-purpose sensing modalities beyond accuracy-only benchmarks, improving the time to deployment.

Reporting a minimum PCI metadata set enables reproducible, deployment-oriented comparisons across future studies.

Early pest detection in edible crops demands sensing solutions that can run at the edge under tight power, budget, and maintenance constraints. This review synthesizes peer-reviewed work (2015–2025) on three modality families—vision/AI, spectroscopy/imaging spectroscopy, and indirect sensors—restricted to edible crops and studies reporting some implementation or testing (n = 178; IEEE Xplore and Scopus). Each article was scored with a modality-aware performance–cost–implementability (PCI) rubric using category-specific weights, and the inter-reviewer reliability was quantified with weighted Cohen’s κ. We translated the evidence into compact decision maps for common deployment profiles (low-power rapid rollout; high-accuracy cost-flexible; and block-scale scouting). Across the corpus, vision/AI and well-engineered sensor systems more often reached deployment-leaning PCI (≥3.5: 32.0% and 33.3%, respectively) than spectroscopy (18.2%); the median PCI was 3.20 (AI), 3.17 (sensors), and 2.60 (spectroscopy). A Pareto analysis highlighted detector/attention models near (P,C,I)≈(4,5,4); sensor nodes spanning balanced (4,4,4) and ultra-lean (2,5,4) trade-offs; and the spectroscopy split between the early-warning strength (5,4,3) and portability (4,3,4). The inter-rater agreement was substantial for sensors and spectroscopy (pooled quadratic κ = 0.73–0.83; up to 0.93 by dimension) and modest for imaging/AI (PA vs. Author 2: κquadratic=0.30–0.44), supporting rubric stability with adjacency-dominated disagreements. The decision maps operationalize these findings, helping practitioners select a fit-for-purpose modality and encouraging a minimum PCI metadata set to enable reproducible, deployment-oriented comparisons.

Figures