The use of image processing techniques for the automated detection of blue-green algae

The determination of water quality in freshwater lakes and reservoirs is an important task, which must be carried out on a regular basis. Information about long term water quality must be provided by the existence of particular organisms, for example blue-green algae. Currently the detection of these algae is done in a very time consuming manual way, involving highly trained biologists, for example those employed by the National Rivers Authority. This thesis is a first investigation in the automatic detection of blue-green algae using image processing techniques.<br/><br/>Samples of seven species of blue-green algae and two species of green algae were examined under a microscope and transferred to a computer. The micro­scope pictures were then stored as digital images.<br/><br/>In order to locate the organisms Image Segmentation routines were applied. In particular, a newly developed LoG Thresholding Operator proved to be effective for the segmentation of biological organisms. Image Enhancement improved the quality and appearance of the segmented species in the images. In the identification process the biological key, which describes some important features of each species, needed to be implemented. With the aid of shape algorithms and textural algorithms both occluding and non-occluding organisms were analyzed and meaningful features were extracted.<br/><br/>The obtained features were then used to classify the organisms into different species. Both, discriminant analysis and neural networks were used for classification purposes. A detection rate of approximately 90% was achieved.<br/><br/>The approach has produced promising results and it is hoped that further investigations will be encouraged.<br/>

Name dc:type.qualificationname

Doctoral Thesis

Level dc:type.qualificationlevel

Student thesis

Year dc:date.issued

1994

Author dc:creator

• Thiel, Stefan

Language dc:language

eng

Identifier

oai:pure.atira.dk:studenttheses/fd73551d-72d8-46a1-a0e8-e3c08b51f03e

OAI identifier oai:identifier

oai:pure.atira.dk:studenttheses/fd73551d-72d8-46a1-a0e8-e3c08b51f03e