Automatic Germination time estimation in Microscopic Images

This thesis propose an automatic algorithm for the estimation of germination time on microscopic images of Aspergillus Niger. The fungus is one of the most important producers of citric acid and catalytic enzymes where production is heavily determined by its present morphology. Study of the germination rate and how the fungi is affected by its growth factors allow for controlled regulations on optimal condition. By image processing on microscopic images the algorithm is capable of withdrawing necessary information for the estimated time of germination. A simple and efficient algorithm is achieved by combining tools for binary and grayscale morphology. Consecutive evaluation on visual outputs and the indicated consistency with biology concludes on sufficient accuracy of the germination time estimates. Area estimates at time of germination is returned and analysed. A one-way analyse of variance is conducted revealing equal mean for the experiments.