Identifying sensors for better IPM in cotton: NEC1901
Technical report
Title | Identifying sensors for better IPM in cotton: NEC1901 |
---|---|
Report Type | Technical report |
Research Report Category | Industry |
Authors | McCarthy, Alison, Long, Derek and Grundy, Paul |
Institution of Origin | University of Southern Queensland |
Number of Pages | 57 |
Year | 2022 |
Publisher | Cotton Research and Development Corporation |
Place of Publication | Australia |
Abstract | Insect pests in cotton can cause significant economic loss if undetected or managed incorrectly. Pest populations and plant symptoms are typically monitored manually by agronomists. Accurate pest detection can be difficult with many pests such as Silverleaf Whitefly (SLW) and aphids located on the undersides of leaves and can have irregular distribution throughout a field area. Accurate detection and spatial/temporal tracking of pests is critical to an effective cotton integrated pest management (IPM) strategy that relies on timely decision making. Machine vision has potential to detect cotton pests and symptoms using handheld cameras and software through automatic image analysis. This project developed an automated SLW and aphid counting App for agronomists and growers to improve cotton IPM. This involved working closely with agronomists in workshops to identify required functions of the App. Images were collected using smartphones over four cotton seasons and glasshouse trials and machine vision algorithms developed that counted healthy and parasitised SLW and aphids with 75% accuracy. The algorithms were implemented in UniSQ-developed Apps, as an alpha version for image collection by QDAF, and then as a beta version (25+ users) for image collection and visualisation with the CRDC/CSIRO/QDAF’s Decision Support Tool (DST) for spray decisions. Clevvi Marketing were identified as a commercial partner for App development through CRDC’s Expression of Interest process and developed a version of the App for public release. The project also investigated using imagery and automated analysis to detect honeydew on cotton plants. This involved conducting glasshouse and commercial field trials with known SLW levels and assessing collected imagery. |
Keywords | Machine vision, insects, image analysis |
ANZSRC Field of Research 2020 | 300202. Agricultural land management |
460304. Computer vision | |
Public Notes | Files associated with this item cannot be displayed due to copyright restrictions. |
Byline Affiliations | University of Southern Queensland |
Department of Agriculture and Fisheries, Queensland |
https://research.usq.edu.au/item/z22q1/identifying-sensors-for-better-ipm-in-cotton-nec1901
66
total views3
total downloads3
views this month0
downloads this month