Work

Field Validation of Infrared Thermography as a Non-Destructive Screening Tool for Detecting Termite Infestation and Wood Decay in Mango Trees

Year: 2026
Type: conference-paper
Abstract: Early detection of termite infestation and wood decay is important for maintaining tree health and plantation productivity. This study evaluated infrared thermography (IRT) as a non-destructive screening tool for identifying field-observed signs of deterioration in mango (Mangifera indica L.) trees. We imaged 100 mature trees in a 4.29-ha commercial plantation in Butuan City, Philippines, with a handheld thermal camera (120 × 90 pixels) from January to March 2026. We assessed diagnostic performance using a stratified random split: 70 trees for model development and 30 for internal holdout validation. The training AUC was 0.999, and the training-derived Youden threshold was ΔT ≥ 1.00°C. Applied to the validation subset, this threshold yielded an AUC of 1.000 and 100% sensitivity, specificity, positive predictive value, negative predictive value, and accuracy. A full-dataset linear regression described the association between ΔT and the recorded Severity Index (Severity Index = 0.688 + 1.304ΔT; R 2 = 0.972). These findings suggest that IRT may help prioritize trees for field inspection when combined with direct field assessment and spatial mapping. However, the validation subset was small, came from the same plantation, and used field-observed signs rather than independent diagnostic confirmation. The Severity Index also includes thermal indicators, creating potential circularity in its association with ΔT. The threshold and regression therefore require cautious interpretation and independent validation.
Language: English
Open Access status: diamond
OpenAlex ID: W7225753026
FWCI:
Cites: 15
Cited by:
Related to: 10