← 論文一覧に戻る

持続可能なコールドチェーンにおけるフロントエンド品質監視のための省エネ型エッジビジョンシステム：コーヒー果実の成熟度検出と組込み検証

An Energy-Efficient Edge Vision System for Front-End Quality Monitoring in Sustainable Cold Chains: Coffee Fruit Ripeness Detection and Embedded Validation (原題)

Xiu-Yu Cui, Shi-Rong Guo

Italian National Conference on Sensors📚 査読済 / ジャーナル2026-10-15#省エネ経営インパクト: コスト削減対象セクター: agriculture
DOI: 10.3390/s26206535
原典: https://doi.org/10.3390/s26206535
📄 PDF

🤖 gxceed AI 要約

日本語

本研究は、持続可能なコールドチェーンの上流品質監視に向け、RK3576 NPU上で動作する省エネ型エッジビジョンシステムを開発した。YOLO11を再構成し、細粒度表現・効率的な多スケール融合・簡素化回帰ヘッドを組み合わせ、640px入力で計算量を6.3から5.5 GFLOPsへ削減。mAP50は83.2%に向上し、スループットは24.2%増、320pxで50.3fps・10.06fps/Wを達成した。

English

This study develops an energy-efficient edge vision system on an RK3576 NPU for six-class coffee fruit ripeness and condition detection. A reconfigured YOLO11 (BCS) improves mAP50 to 83.2% while cutting complexity from 6.3 to 5.5 GFLOPs at 640px. Board measurements show 24.2% higher throughput and 50.3 fps at 10.06 fps/W at 320px, supporting low-power quality monitoring near harvest.

Unofficial AI-generated summary based on the public title and abstract. Not an official translation.

📝 gxceed 編集解説 — Why this matters

日本のGX文脈において

日本企業が関与するコーヒー・農産物サプライチェーンの上流品質管理や、省電力エッジAIによる一次産業のデジタル化に関わる。Scope 3上流の品質・廃棄削減の観点で参考になるが、開示制度との直接連関は弱い。

In the global GX context

Relevant to global sustainable supply-chain and food-loss reduction efforts, where front-end quality monitoring supports Scope 3 upstream traceability. The energy-efficient edge AI angle connects to low-power deployment in agriculture, though it does not directly address disclosure frameworks like CSRD or TCFD.

👥 読者別の含意

🔬研究者:省電力エッジAIと農業品質検出を組み合わせた実装・評価手法の参考になる。

🏢実務担当者:農産物サプライチェーンの上流品質監視や廃棄削減のための低コスト・省電力ソリューション検討に活用できる。

📄 Abstract（原文）

Reliable quality monitoring at the front end of sustainable cold chains requires direct observations of product condition in addition to environmental measurements. This study developed an energy-efficient edge vision system for six coffee fruit ripeness and visible-condition classes on an RK3576 neural processing unit (NPU). YOLO11 was reconfigured by combining a Branch-Integrated Feature Aggregation module for fine-grained representation, a YOLOv10-derived Cross-Stage Partial Network with Compact Inverted Block for efficient multi-scale fusion, and a Simplified Regression Head with reduced regression channels. On a public benchmark derived from 996 source images, YOLO11 and YOLO11BCS yielded mean average precision values at an intersection-over-union threshold of 0.5 (mAP50) of 81.5% and 83.2%, respectively, under the matched 320-pixel protocol. At a 640-pixel input, the computational complexity decreased from 6.3 to 5.5 giga floating-point operations. The largest class-level mAP50 difference was 9.7 percentage points for overripe fruit. Under controlled synthetic degradations, the model produced higher point estimates than YOLO11 for illumination variation and rain, whereas severe Gaussian noise and motion blur remained limiting. Board-input measurements showed throughput increases of 24.2% and 22.1% at 640- and 320-pixel inputs, respectively. The 320-pixel system reached 50.3 frames per second and 10.06 frames per second per watt. Within the supplied benchmark, the matched evaluation supports a comparative assessment of the models, while source-image correspondences across partitions define the scope of the reported absolute metrics. Within these boundaries, the results support low-power visual quality monitoring near the point of harvest.

🔗 Provenance — このレコードを発見したソース

🔔 こうした論文の新着を逃したくない方は キーワードアラート に登録（無料・3キーワードまで）。

gxceed は公開メタデータに基づく研究支援データセットです。要約・翻訳・解説は AI 支援で生成されています。 最終的な解釈・検証は利用者が原典資料に基づいて行うことを前提とします。