A direct statistical method for binary segmentation of the abundance of bird populations in static images

Abstract

In order to overcome the technical problems of the existing static picture bird population abundance statistical method with low processing speed and accuracy, the present invention provides a binary segmentation direct statistical method for the static picture bird population abundance; this method performs relevant pseudo-frame difference calculations on each color value of each pixel point by establishing a translation sub-matrix of a still image, then weights the absolute values of different color differences after each pixel frame difference as the basic data of each frame difference element, then binarizes the basic data of each frame difference element using a given threshold, performs closed area marking on the basis of the binarization, and finally judges the shape and size of each of the already marked closed areas one by one, distinguishes the ground, trees, cloud background and miscellaneous points according to the size, and counts the number of single birds or multiple overlapping birds according to the shape, and counts the number of birds in the picture.

Description

Technical Field

[0001] The present invention relates to an image processing method, and particularly to a method for statistically calculating the abundance of bird populations based on static bird pictures, belonging to the field of image processing. Background Art

[0002] The ecological environment has gradually become one of the important indicators for evaluating the achievements of the government. How to achieve harmonious coexistence with nature has become an urgent problem to be solved in society. The statistical calculation of bird abundance and population identification are of great significance for biology, environmental protection, and the sustainable development of the country, and are even more important reference bases for ecological environment assessment. As a gregarious animal, birds often suffer from inaccurate counting or even inability to count with the naked eye due to human visual errors during the process of static high-density abundance statistics. If the counting method cannot be improved, it will consume a large amount of manpower, material resources, and time. At the same time, for endangered rare birds, their habitats can be effectively protected by analyzing their behavioral characteristics.

[0003] Currently, bird researchers and protectors often take static images, perform preprocessing such as image binarization and enhancement, and then perform processing such as segmentation and extraction. Not only is the processing speed low, but the extraction accuracy is also low. Summary of the Invention

[0004] In order to overcome the technical problems of low processing speed and accuracy in the existing method for statistically calculating the abundance of bird populations in static pictures, the present invention provides a direct binary segmentation statistical method for the abundance of bird populations in static pictures. This method calculates the relevant pseudo-frame differences for each color value of each pixel point by establishing a translation sub-matrix of the static image, then weights the absolute values of different color differences after each pixel frame difference as the basic data for each frame difference element, and then binarizes the basic data of each frame difference element using a given threshold. On the basis of binarization, closed area marking is performed, and finally, the marked closed areas are judged one by one according to their shapes and sizes, the ground, trees, cloud backgrounds, and miscellaneous points are distinguished by size, and single birds or overlapping counts of multiple birds are distinguished by shape, and the number of birds in the picture is statistically calculated.

[0005] The technical solution adopted by the present invention to solve its technical problems is: a direct binary segmentation statistical method for the abundance of bird populations in static pictures, which is characterized by including the following steps:

[0006] (1) Store the obtained static bird picture as a corresponding planar image matrix , with subscripts and being the number of rows and columns of the planar image matrix respectively. The following description of the subscript meanings of the planar image matrix is the same;

[0007] (2) Establish a sub-matrix of ​, where is the rear column, is an arbitrary non-zero positive integer selected according to the actual image, which is equivalent to the left shift amount before the pseudo-frame difference processing of the image, simulating the moving image targeted by the general frame difference method;

[0008] (3) Perform binarization processing with a threshold in the following manner. Define:

[0009]

[0010] Among them, is the rd row, th column element, the rd row, th

[0011] ;

[0012]

[0013] Among them, , , , are positive numbers;

[0014] (5) Mark the closed regions of the points in . If the number of points in the left, upper left, lower left, right, upper right, lower right, upper, and lower neighborhood directions of the already marked closed region is less than points is zero, then mark them as the same closed region, is a positive integer, otherwise mark them as different closed regions;

[0015] (6) Count the number of birds in the non-eliminated regions of the picture: Judge the shape and size of the already marked closed regions one by one. If the number of pixel points in a certain region , then it is determined as background noise; if the number of pixel points in a certain region , then it is determined as the background of the ground, trees, and clouds; is the number of pixel points in the marked closed region, are all positive integers.

[0016] The beneficial effects of the present invention are as follows: By performing pseudo-frame difference operations on static images and weighting the absolute values of different color differences after frame difference for each pixel as the basic data of each frame difference element, using a given threshold to binarize the basic data of each frame difference element, marking closed regions on the basis of binarization, and judging the shape and size of each closed region one by one, the number of birds in the picture can be counted, solving the technical problem of the low processing speed and accuracy of the existing statistical method for the abundance of bird populations in static pictures.

[0017] The following specifically describes the present invention in detail in conjunction with the specific embodiments. Specific Embodiments

[0018] (1) Store the obtained static bird picture as a corresponding planar image matrix , with subscripts and being the number of rows and columns of the planar image matrix respectively. The following descriptions of the subscript meanings of the planar image matrix are the same;

[0019] (2) Establish a sub-matrix , where is the last columns of , equivalent to shifting the image one bit to the left before performing pseudo-frame difference processing, simulating the moving images targeted by the general frame difference method;

[0020] (3) Perform binarization processing with a threshold in the following manner. Define:

[0021]

[0022] Among them, is the th row, th column element, and th chromaticity value of , is the th row, th column element, and th chromaticity value of , where

[0023] ;

[0024]

[0025] Among them, , , , are positive numbers;

[0026] (5) For Mark the enclosed areas for the points. If the neighborhood of the already marked enclosed area is less than points is zero, then mark it as the same enclosed area. is a positive integer; otherwise, mark it as a different enclosed area.

[0027] (6) Count the number of birds in the non-excluded areas of the picture: Judge the shape and size of the already marked enclosed areas one by one. If the number of pixel points in a certain area , then it is determined as background noise points; if the number of pixel points in a certain area , then it is determined as the background of the ground, trees, and clouds. is the number of pixel points of the unmarked enclosed area, both are positive integers.

Claims

1. A direct statistical method for binary segmentation of the abundance of bird populations in static pictures, characterized by the following steps: (1) Store the obtained static bird pictures as corresponding planar image matrices , the subscripts and are the number of rows and columns of the planar image matrix respectively, and the following descriptions of the subscript meanings of the planar image matrix are the same; (2) Establish 's sub - matrix , where is 's last columns, is an arbitrary non - zero positive integer selected according to the actual image, equivalent to the left - shift amount before performing pseudo - frame difference processing on the image, simulating the moving image targeted by the general frame - difference method; (3) Perform binarization processing with a threshold in the following manner, and define: Among them, for No. Row, No. Column element, Chroma values, for No. Row, No. Column element, Chroma values, where ; Among them, , , , are positive numbers; (5) Close the area of the points of . If the number of points in the left, upper left, lower left, right, upper right, lower right, upper, and lower neighborhoods of the already marked closed area is less than points is zero, then mark them as the same closed area, is a positive integer, otherwise mark them as different closed areas; (6) Count the number of birds in the non-excluded area of the picture: For the marked closed areas, judge their shapes and sizes one by one. If the number of pixel points in a certain area , it is determined as background noise; if the number of pixel points in a certain area , it is determined as the background of the ground, trees, and clouds; is the number of pixel points in the marked closed area, are all positive integers.

Images