This paper proposes an image analysis scheme to automatically detect the morphological feature of a fungi, Penicillium. While the previous method of detection is rather difficult as it is time consuming and only an experienced mycologist can evaluate the fungi cultures, the proposed method aims to ease the works of mycologists and shorten the time of detection. The image of fungi usually suffers from a low contrast and presence of noise which makes it difficult to detect the morphological features. Therefore, an application of sharpening filter together with adaptive histogram equalization was investigated in the image enhancement stage to overcome the low quality Penicillium image. To eliminate the noise, a combined operation of Canny edge detection algorithm, morphological operation and largest connected pixel region was applied on the image. To reduce the time consumption, only the morphological features of the fungi is segmented in this study. The resulted image from each process of enhancement and segmentation is then evaluated using PSNR. The results shows Laplacian filter outperform AHE with a value of 54. 105dB.
