Skip to main content
ARS Home » Research » Publications at this Location » Publication #126769
Title: DESCRIBING CELL AND TISSUE GEOMETRY IN LEAF CROSS-SECTIONS USING IMAGE CLASSIFICATION TECHNIQUES

Author
item Pachepsky, Ludmila
item FAUSTINELLI, PAOLA - IFFIVE-INTA, ARGENTINA
item DARDANELLI, JULIO - IFFIVE-INTA, ARGENTINA
item COLLINO, DANIEL - IFFIVE-INTA, ARGENTINA
item FERRYRA, ANDRES - UNIVERSITY OF FLORIDA
item LIU, LAN
item Reddy, Vangimalla

Submitted to: Abstract of Agronomy Meetings
Publication Type: Abstract Only
Publication Acceptance Date: 3/1/2002
Publication Date: N/A
Citation: N/A

Interpretive Summary:

Technical Abstract: New image-analysis technique is presented for measuring the sizes of mesophyll, water storage, epidermal cells, and the area of the intercellular space in leaf cross-sections of peanut, cotton, and soybean. Cross-section microphotographs were digitized, and the resulting RGB color images were processed with image processing software to produce artificial "bands" by histogram equalization and color adjustment. All images were then imported into the GIS package Idrisi. The original image was processed with a maximum likelihood classifier using the processed images to develop signatures for the different tissues. The sizes of the cells, the tissues and the intercellular spaces were then measured using a pixel-counting function. The data obtained were similar to the results of conventional measurements with software packages like SigmaScan, where all the features must be completely digitized all across the image. The classification-based technique requires only about 10% of the image for classification training sites, and it takes about 80% less time than conventional measurements, with an error of less than 5%.