Skip to main content
ARS Home » Midwest Area » Wooster, Ohio » Application Technology Research » Research » Publications at this Location » Publication #165114

Title: Dynamic Segmentation for Automatic Spray Deposits Analysis on Uneven Leaf Surfaces

Author
item RAMALINGAM, N - OSU
item LING, P - OSU
item Derksen, Richard

Submitted to: Transactions of the ASAE
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 2/1/2003
Publication Date: 8/1/2003
Citation: Ramalingam, N., Ling, P.P., Derksen, R.C. 2003. Dynamic Segmentation for Automatic Spray Deposits Analysis on Uneven Leaf Surfaces. Transactions of the ASAE. 46(3):893-900.

Interpretive Summary: Spray deposit quality on target surfaces is as important of a factor in the efficacy of a pesticide application as the total amount of pesticide on the target. Uneven leaf surfaces, varying intensities of fluorescent tracers, background-to-object contrast and other factors contribute to the difficulties of assessing foliar spray coverage. Four different dynamic thresholding algorithms were implemented and evaluated for their ability to increase the accuracy and decrease the time required to measure foliar spray coverage. All four dynamic thresholding algorithms were implemented and found able to segment the images of spray deposits with varied contrasts of fluorescent tracer. The areas of spray deposits measured automatically by the dynamic thresholding algorithms were compared to those obtained from the human measurement. There were no significant differences in the areas calculated by the dynamic thresholding algorithms and by the human analysis. Among the four methods investigated, the minmax thresholding technique had the best general performance and was least dependent on image characteristics. A two-pass approach incorporating the minmax technique and an image-specific thresholding algorithm produced less error than all one-pass approaches. Incorporating the one-pass minmax algorithm or the two-pass thresholding algorithms will produce results as good as subjective, human-orientated, thresholding approaches and will significantly reduce the time required to process and evaluate differences in sprayer performance resulting in better equipment and pest management recommendations for growers.

Technical Abstract: Assessing foliar spray coverage can aid in determining factors leading to the loss of pesticide efficacy. Spray coverage evaluation on leaf surfaces is usually a very subjective process, made more difficult by variations in leaf surface morphology and lighting conditions. The objectives of this study were to implement and evaluate four different dynamic thresholding algorithms that could increase the efficiency of spray coverage measurements on leaf surfaces and reduce reliance on subjective operator judgments. The algorithms were evaluated for accuracy to effectively segment the images that had non-uniform contrast between the droplets and the leaf background and with varying intensities of the fluorescent tracer over uneven leaf surfaces. The analysis included segmentation of the droplets, estimation of droplet area and the total number of different size and shape droplet deposits on a leaf surface. Guidelines for selecting a suitable dynamic thresholding technique for given image characteristics are proposed. A two-pass approach was evaluated wherein the images were first thresholded by the best general purpose technique to compute the blob characteristics, based on which an image specific thresholding algorithm was applied as the second pass for improved accuracy. The experimental results identify techniques that will aid in the objective evaluation of spray coverage on leaf surfaces. Implementing these techniques will provide manufacturers and growers with more confidence in reported results and will increase the speed in which coverage evaluations can be made and reported.