Articles
IMAGE PROCESSING OF PLANTS FOR EVALUATION OF GROWTH IN RELATION TO ENVIRONMENT CONTROL
Article number
87_30
Pages
283 – 290
Language
Abstract
The methods of image processing of plants were developed, in order to obtain the information about the plant growth to be used in an online computer system for optimalizing the environment of cultivation.
A plant image was taken through the vidicon camera and digitized in binary notation, by setting a constant slice level of the brightness of the plant to denote the plant image as ‘1’ and background as ‘0’. To eliminate the variation among the plant images, which was caused by camera angle, nine images taken from different angles to a plant were summed and expressed as a compounded matrix.
The quantitative growth was evaluated by the function of sum(P) of matrix elements, and the difference in feature of growth was estimated by statistical pattern recognition in coordinate system of P and the vertical position of "centroid" in the compounded matrix.
In practical cultivation, the variation of plant image, caused by camera angle, became negligible because many plants were taken together in an image.
So, the evaluation of the growth in plant population was succeeded by using function of P in a digitized image taken from a constant angle.
On the other hand, background noises behind the plant was eliminated by using the infrared region(850–1100nm) for sensing the image in practical cultivation, and clear digitized image of plants was extracted enough to use for image processing.
By these methods, evaluation of plant growth was made possible by the mathematical model represented as several functions obtained from the image processing of plants.
A plant image was taken through the vidicon camera and digitized in binary notation, by setting a constant slice level of the brightness of the plant to denote the plant image as ‘1’ and background as ‘0’. To eliminate the variation among the plant images, which was caused by camera angle, nine images taken from different angles to a plant were summed and expressed as a compounded matrix.
The quantitative growth was evaluated by the function of sum(P) of matrix elements, and the difference in feature of growth was estimated by statistical pattern recognition in coordinate system of P and the vertical position of "centroid" in the compounded matrix.
In practical cultivation, the variation of plant image, caused by camera angle, became negligible because many plants were taken together in an image.
So, the evaluation of the growth in plant population was succeeded by using function of P in a digitized image taken from a constant angle.
On the other hand, background noises behind the plant was eliminated by using the infrared region(850–1100nm) for sensing the image in practical cultivation, and clear digitized image of plants was extracted enough to use for image processing.
By these methods, evaluation of plant growth was made possible by the mathematical model represented as several functions obtained from the image processing of plants.
Authors
T. Matsui, H. Eguchi
Keywords
Online Articles (45)