Dunn, Mark (2007) Applications of vision sensing in agriculture. [Thesis (PhD/Research)]
PDF (Introductory Pages)
PDF (Whole Thesis)
[Abstract]: Machine vision systems in agricultural applications are becoming commonplace as technology becomes both affordable and robust. Applications such as fruit and vegetable grading were amongst the earliest applications, but the field has diversified into areas such as yield monitoring, weed identification and spraying, and tractor guidance.
Machine vision systems generally consist of a number of steps that are similar between applications. These steps include image pre-processing, analysis, and post-
processing. This leads the way towards a generalisation of the systems to an almost ‘colour by number’ methodology where the platform may be consistent between many applications, and only algorithms specific to the application differ.
Shape analysis is an important part of many machine vision applications. Many methods exist for determining existence of particular objects, such as Hough Transforms and statistical matching. A method of describing the outline of objects, called s-ψ (s-psi) offers advantages over other methods in that it reduces a two dimensional object to a series of one dimensional numbers. This graph, or chain, of numbers may be directly manipulated to perform such tasks as determining the convex hull, or template matching.
A machine vision system to automate yield monitoring macadamia harvesting is proposed as a partial solution to the labour shortage problems facing researchers
undertaking macadamia varietal trials in Australia.
A novel method for objectively measuring citrus texture is to measure the shape of a light terminator as the fruit is spun in front of a video camera. A system to accomplish this task is described.
S-psi template matching is used to identify animals to species level in another case study. The system implemented has the capability to identify animals, record video and also open or shut a gate remotely, allowing control over limited resources.
Statistics for this ePrint Item
|Item Type:||Thesis (PhD/Research)|
|Item Status:||Live Archive|
|Additional Information:||Doctor of Philosophy (PhD) thesis. Please contact author for associated CD.|
|Depositing User:||epEditor USQ|
|Faculty / Department / School:||Historic - Faculty of Engineering and Surveying - Department of Mechanical and Mechatronic Engineering|
|Date Deposited:||28 Apr 2008 00:31|
|Last Modified:||02 Jul 2013 23:01|
|Uncontrolled Keywords:||machine vision systems; machine vision; guidance; agriculture; agricultural applciations|
|Fields of Research :||08 Information and Computing Sciences > 0801 Artificial Intelligence and Image Processing > 080104 Computer Vision
09 Engineering > 0906 Electrical and Electronic Engineering > 090605 Photodetectors, Optical Sensors and Solar Cells
09 Engineering > 0999 Other Engineering > 099901 Agricultural Engineering
Actions (login required)
|Archive Repository Staff Only|