A Computational Approach for Recognition of Electronic Microscope Plant Pollen Images

In this paper, a computer application program to process electronic microscope images of plant pollen is presented. We use contrast enhancement to separate pure object pollen from original image. Hough edge detector is used to extract complete and inside-detail-kept edges from pure object image for recognition. Pre-process such as smoothing, grayscale equipoise, etc., is performed to remove noise and enhance image. In the recognition process, factors of mathematics and statistics are calculated. Those factors are basis of the final judgment to determine whether the sample image is the same species as one in the templates library. The whole program is built in C++. It makes a low time and memory space cost per recognizing task.