AuToDiDAC: Automated Tool for Disease Detection and Assessment for Cacao Black Pod Rot

Abstract

Pest control strategies for crop diseases highly depend on visual inspection to assess the severity of the infection, which usually lead to inconsistencies: either over or under assessment. These inconsistencies could be attributed to the limitations of humans to perceive small differences. A more precise disease assessment is needed for better pest management decision, which will result to a more efficient utilization and allocation of resources for farm inputs. This translates to a better income for cacao farmers. This paper introduces a mobile application named AuToDiDAC or Automated Tool for Disease Detection and Assessment for Cacao Black Pod Rot (BPR). AuToDiDAC automatically detects, separates, and assesses the infection level of BPR in cacao through image processing and machine learning techniques. It gives the farmers the capacity to objectively monitor and report the infection level of the BPR compared to the common visual rating for plant disease level of infection. Pixel-level accuracy test of the tool showed an average of 84% accuracy on an independent test set of ten cacao pod images.

Plant phenotyping relevance

カカオ果実の病徴を画像処理・機械学習で検出、分離し、感染レベルを評価する手法と精度検証が研究の中心であるため、植物病害状態のフェノタイピング手法として含める。

Code and data availability

公開本文の所在を確認できませんでした。非公開または購読が必要な可能性があります。

No evidence-backed public reproduction asset is currently recorded.