Diatom characteristic description and classification method based on contour drawing

Abstract

The invention discloses a diatom characteristic description and classification method based on a contour drawing, which relates to a characteristic representation method of a diatom microscopic image. The invention provides a diatom characteristic description and classification method based on a contour drawing, which automatically describes the contour of a diatom microscopic image, can be applied directly to a diatom recognition system and can provide a high-efficient diatom automatic recognition method for scientific research and cultivation. The method comprises the following steps: a diatom body contour drawing is obtained by the pretreatment and contour segmentation of the diatom microscopic image; the contour characteristics are analyzed and obtained: a contour describing function is obtained by showing the contour with a function to the target area, of which an angle is an independent variable and a value is the radius length from the center of gravity to the contour; standardization is realized by the radius value of the contour describing function divided by the max radius value; a characteristic matrix is formed by extracting the drawing characteristics of the drawing; and a classifier is adopted to classify the characteristic matrix and then classification result is obtained and output.

Description

technical field [0001] The present invention relates to a feature representation method of microscopic images of diatoms, in particular to a method for describing the unique contour features of diatoms by means of a marker map, and extracting the eigenvalues ​​of the marker map functions to form a feature description matrix, and finally according to the The feature matrix realizes the automatic identification and automatic retrieval of diatoms. Background technique [0002] Diatoms are a class of single-celled organisms widely distributed in various habitats and play a vital role in the balance of ecosystems. They are an important part of the primary productivity of the ocean and account for nearly 20% of the global net carbon sequestration production. (up to 20Pg per year), exceeding the sum of the carbon sequestration of all rainforests on land ([1] Mann DG. The species concept in diatoms. Phycologia, 1999, 38: 437-495). In addition, diatoms are widely used in many fields...

AI Technical Summary

Problems solved by technology

The outline of diatoms has a large span, and traditional simple shape descriptors such as "circularity" and "rectangularity" cannot describe the outline characteristics of diatoms well.

However, other boundary descriptors such as Fourier descriptors and statistical moment descriptors cannot intuitively describe diatom boundaries, and cannot effectively train classifiers when the training samples are too small, and cannot be directly applied to the field of automatic image retrieval of algae, resulting in Automatic recognition and retrieval of diatom images is almost impossible to achieve

Images