RU2446471C1 - Method for detecting moving objects and determining parameters thereof - Google Patents

Abstract

FIELD: information technology.

SUBSTANCE: phase decompensation (phase-energy characteristic) arises during movement of an object due to the displacement section. The phase value at which decompensation arises is the maximum phase; it is characterised by a clear-cut maximum or minimum, after zero values, corresponds to the edge of the object furthest from the middle of the row (column); in determining the direction of displacement, using information on coordinates of the object, by comparing maximum phase difference values of phase-energy (hereinafter maximum phase) characteristics for a sequence of three frames; in determining contrast of the object relative the background, using defined values of coordinates and direction of movement of the object; in determining displacement of the object (in pixels), using information on coordinates and direction of movement of the object, by comparing spectral density values corresponding to the maximum phase.

EFFECT: high efficiency of detecting moving objects and determining motion parameters thereof in low luminance conditions, including low-contrast and small-sized objects on the complex inhomogeneous background of a series of digital images.

4 dwg

Description

The invention relates to the field of television and computer technology and can be used in the construction of intelligent systems of technical vision for the automatic detection and registration of motion parameters of objects.

A known method for detecting the movement of objects in a digital image, based on the step-by-step determination of motion vectors from a series of images, the frame area is divided into equal square blocks, the size of which depends on the size of the image in pixels, then for the first frame the calculation of the sum of the quadratic deviations of each block. When calculating the standard deviation of the blocks of the second frame, a certain neighborhood of the block is crawled in search of the maximum match to the image of the block on the previous frame within this neighborhood, the smaller the value of the standard deviation, the more similar the blocks, then motion vectors are formed on the same blocks (see patent EP 1631073 B1, March 26, 2008, “Apparatus and methods for the detection of abnormal motion in a video stream”). This method requires significant computing resources, in this regard, it can be used exclusively for compressing a video sequence (reducing redundancy) and is highly sensitive to changes in image brightness, has a low resistance to noise damage.

The closest in technical essence to the claimed method is the method of the invention is a method for analyzing the interframe difference (K. Fu, R. Gonzalez, K. Lee. Robotics / M .: Mir, 1989, p. 423-426). It consists in determining the changes between two image frames, one of which having stationary components, contains a reference object, and in the subsequent threshold processing of the obtained difference, as a result of which the image pixels of a moving object are extracted. Using the received segmented image, you can determine the location and motion parameters of the object.

The disadvantages of this method is the difficulty to detect (fix) the movement of the object and determine the parameters of its movement in cases of changing illumination within the boundaries of the object, a high noise level, for example, in low light, small size or low contrast of a moving object.

The technical result of the proposed method is to increase the detection efficiency of moving objects and determine their motion parameters in low light conditions, including low-contrast and small-sized objects on a complex inhomogeneous background of a sequence of digital images.

The specified technical result is achieved by measuring the spectral density of the video sequence of two frames and phase-energy characteristics of the sequence of three frames, determining the interframe difference of the spectral density of the video signal and the pairwise inter-frame difference of phase-energy characteristics of the video sequence taken for three frames, on the basis of which the parameters of the object’s movement are determined, namely, the coordinates according to one interframe difference of phase-energy characteristics; the direction of movement of the object compared to at least two inter-frame differences of phase-energy characteristics; the contrast of the object relative to the background compared to the values of interframe differences of the phase-energy characteristics and spectral density; the displacement of the object in pixels by the difference in spectral density of the values corresponding to the maximum phases.

The essence of the proposed method is as follows:

in the detection of moving objects, the determination of their coordinates by the interframe difference of the phase-energy characteristics. When the object moves due to the displacement site, phase decompensation (phase-energy characteristic) occurs. The value of the phase at which decompensation occurs - the maximum phase, it is characterized by a pronounced maximum or minimum, after zero values, corresponds to the boundary of the object farthest from the middle of the row (column);

in determining the direction of movement, using information about the coordinates of the object, comparing the values of the maximum phases of the difference of the phase-energy (hereinafter the maximum phase) characteristics for a sequence of three frames, in accordance, if the object is on the left side of the image and the value of the maximum phase decreases, then the object moves to the right, the maximum phase value increased, the object moves to the left, if the object is on the right side of the image and the maximum phase value decreases, moves to the left, according to figure 1;

in determining the contrast of the object relative to the background, using certain values about the coordinates and direction of movement of the object, in accordance with figure 2;

in determining the movement of an object (in pixels), using information about the coordinates and direction of movement of the object, by analyzing the spectral density values corresponding to the maximum phase.

The method can be implemented by the following sequence of actions:

- three frames of a video sequence are entered into the database (memory), and rows and columns are sampled (see A. Andreev, Automated television surveillance systems. Part I / SPb: SPbGUITMO, 2005, pp. 14-17)

- further, the spectral density for the first and second frames is measured row by row (in columns) (see patent RU 2370780 dated 10.20.2009, Bogoslovsky A.V., Bogoslovsky E.A., Yudakov D.S. Method for measuring the parameters of the energy spectrum of a two-dimensional signal );

- measure the phase-energy characteristics for three frames, which contain information about both spatial and frequency properties of the image (see Bogoslovsky A.V., Zhigulina I.V. Use of phase-frequency spatial characteristics to evaluate motion. - Successes of modern radio electronics, 2009, No. 11 p. 61-63);

- measure the difference in spectral density;

- measure the interframe difference in phase energy characteristics in pairs;

- based on the inter-frame difference of the phase-energy characteristics, the presence of a moving object in the frame and its coordinates by the maximum phase are determined (see Bogoslovsky A.V., Zhigulina I.V., Kopylov O.E., Yakovlev V.A. an image based on interframe differences in frequency characteristics - Radio Engineering, 2010, No. 5, pp. 55-59);

- determine the direction of motion by changing the values of the maximum phase of the pairwise difference taken for three frames of the video sequence, in accordance with figure 1;

- determine the contrast of the object relative to the background by signs in the maximum phase of the difference of the phase-energy characteristics and spectral density in accordance with figure 2;

- determine the displacement of the object (in pixels) by measuring the magnitude of the front of the "burst" of the difference in spectral density corresponding to the maximum phase, in accordance with figure 3;

- the conclusion about the correct detection of the object of interest is carried out by a person (operator).

The method can be implemented, for example, using a device whose structural diagram is presented in figure 4, where are indicated:

1 - input image forming system;

2 - buffer device;

3-1 - selection by rows;

3-2 - selection by columns;

4-1 - line analysis unit;

4-2 - column analysis block;

5 - operator

1) a sequence of images of a real scene containing the desired moving object arrives at block 1;

2) then a digital sequence of images, in the form of pixel brightness values from the output of block 1, is fed to the input of block 2, where the frame is recorded in the form of rows and columns of three frames of a video sequence;

3) from the output of block 2, the frames of the video sequence simultaneously arrive at the inputs of blocks 3 and 4, where, respectively, samples are formed by rows and columns;

4) from the outputs of blocks 3-1 and the sample in rows goes to blocks 4-1, from the output of block 3-2, the selection in the image columns goes to the input of block 4-2;

5) in block 4-1, three consecutive frames are memorized line by line, measurement of the phase-energy characteristic and spectral density for each line, determination of the inter-frame difference, based on which the object’s location and motion characteristics are determined (can be implemented, for example, using a processor for digital processing signals);

6) in block 4-2, similar operations occur as in block 4-1; only for columns in a video sequence;

7) block 5, the operator decides to detect objects.

The use of this search method will increase the likelihood of correct detection of the movement of objects in a sequence of digital images with a complex inhomogeneous background, in conditions of high noise levels.

The analysis of the prior art by the applicants, including a search by patent and scientific and technical sources of information and identification of sources containing information about analogues of the invention, allowed us to establish that the applicants did not find analogues that are characterized by features identical to all the essential features of the invention.

Therefore, the claimed invention "A method for detecting moving objects and determining their parameters" meets the criterion of "novelty."

Claims (1)

A method for detecting moving objects and determining their parameters, based on the formation of an image video signal, storing a sequence of frames and obtaining inter-frame differences, characterized in that the spectral densities of the video sequence of two frames and phase-energy characteristics of the sequence of three frames are determined, the inter-frame difference of the spectral density of the video signal and the pairwise inter-frame difference are determined phase-energy characteristics taken for three frames in sequence STI, on the basis of which the parameters of the object’s motion are determined, namely: coordinates along one interframe difference of phase energy characteristics; the direction of movement of the object compared to at least two inter-frame differences of phase-energy characteristics; the contrast of the object relative to the background compared to the values of interframe differences of the phase-energy characteristics and spectral density; the displacement of the object in pixels by the difference in spectral density of the values corresponding to the maximum phases.

Images