Assessing the ability of image processing software to analyse spray quality on water-sensitive papers used as artificial targets

• Type: Article in International Scientific Journal
• Date: 2012

Authors

Mario Cunha (Author) FCUP; Claudia Carvalho (Author) Other; Andre R S Marcal (Author) FCUP

Journal

Title: Biosystems Engineering

Vol. 111

Pages: 11-23

ISSN: 1537-5110

Publisher: Elsevier

Indexing

ISI Web of Knowledge - 70 Citations

Scopus - 84 Citations

Scientific classification

FOS: Agrarian Sciences > Agricultural biotechnology

Other information

Authenticus ID: P-002-H6W

DOI: 10.1016/j.biosystemseng.2011.10.002

Abstract (EN): The performance of several commercial and experimental software packages (Gotas, StainMaster, ImageTool, StainAnalysis, AgroScan, DropletScan and Spray_imageI and II) that produce indicators of crop spraying quality based on the image processing of water-sensitive papers used as artificial targets were compared against known coverage, droplet size spectra and class size distribution verified through manual counting. A number of artificial targets used to test the software were obtained by controlled spray applications and given droplet density between 14 and 108 drops cm(-2) and a wide range of droplet size spectra. The results showed that artificial targets coupled with an appropriate image system can be an accurate technique to compute spray parameters. The between-methods differences were 6.7% for droplet density, 11.5% for volume median diameter, <3% for coverage (%) and <3% coverage density. For the 16 droplet class size distribution tested the between-methods differences were all <15%. However, most of the image analysis systems were not effective in accurately measuring coverage density when coverage rate is greater than about 17%. The Spray_imageII software estimated the coverage density with a mean absolute error of 2% and the absolute error is below 10%, even with about 43% of coverage rate. This software, when compared to the other programmes tested, provided the best accuracy for coverage and droplet size spectrum as well as for droplet class size distribution.

Language: English

Type (Professor's evaluation): Scientific

No. of pages: 13