Detection of fungus through an optical sensor system using the histogram of oriented gradients

Fungus is an important component in our ecosystem. It performs an important task of decomposition. But on the other hand, it is the main risk for human health, archives, food logistics and millions of euros lost per annum just due to different kinds of fungus. The main aim of this research is to develop an automated system for the detection of fungus spores in air. We have developed a novel system for fungus detection through an optical sensor system. First of all, our system will collect air samples. Then, the handling system moved them under a microscopic camera and get images of the sample. Since images have noise regions and the spore boundaries were unclear, pre-processing techniques have been applied and different filters have been used. Then, the histogram of oriented gradient (HOG) features were extracted and the feature vector was formed. Following this, a trained Support Vector Machine (SVM) was used for the purpose of classification.