Fungus Detection Through Optical Sensor System Using Two Different Kinds of Feature Vectors for the Classification

This revolutionary era emphasizes the need of novel, simple, automatic fungal detection system to control the devastation caused by fungal species. Fungus is highly notorious for human health, food, and archives. We have developed a device and algorithms to automatically detect fungus from air by employing optical sensor system. Several air samples were collected and examined under high magnification microscopic camera in order to take images of samples through the designed system. System collects three different types of images (dark-field, bright-field, and auto fluorescence) with the help of different light sources. This provided detailed temporal variation of fungus spores to effectively analyze fungus images. Two different algorithms were developed for fungus spore detection: hand crafted features and histogram of oriented gradients as features. Then, support vector machine was used for the purpose of classification. Furthermore, we compared the fungus spores count obtained with our algorithm and manual counting method. The performance of the proposed system on new images obtained by the optical sensor system was evaluated and all results of precision, recall, and accuracy are also presented in the paper. Convincing results were obtained in favor of the proposed approach for precision recall ratios. The obtained results proved the possibility of developing a device for early detection of several types of fungus spores. The system successfully detected the number of fungus spores present in the air and could estimate all possible threats.