CN103824307A - Method and device for determining invalid moving-object pixels - Google Patents

Method and device for determining invalid moving-object pixels Download PDF

Info

Publication number
CN103824307A
CN103824307A CN201210466476.5A CN201210466476A CN103824307A CN 103824307 A CN103824307 A CN 103824307A CN 201210466476 A CN201210466476 A CN 201210466476A CN 103824307 A CN103824307 A CN 103824307A
Authority
CN
China
Prior art keywords
pixel
illumination
current frame
frame image
fast
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN201210466476.5A
Other languages
Chinese (zh)
Other versions
CN103824307B (en
Inventor
潘华东
潘石柱
吴良健
孙莉
周洪涛
田建国
张兴明
傅利泉
朱江明
吴军
吴坚
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Zhejiang Dahua Technology Co Ltd
Original Assignee
Zhejiang Dahua Technology Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Zhejiang Dahua Technology Co Ltd filed Critical Zhejiang Dahua Technology Co Ltd
Priority to CN201210466476.5A priority Critical patent/CN103824307B/en
Publication of CN103824307A publication Critical patent/CN103824307A/en
Application granted granted Critical
Publication of CN103824307B publication Critical patent/CN103824307B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Landscapes

  • Image Analysis (AREA)

Abstract

The embodiment of the invention provides a method and device for determining invalid moving-object pixels so as to solve a problem that invalid moving-object pixels exist in foreground pixels determined by the prior moving object detection method so that moving-object detection is inaccurate. The method includes: determining pixels, in which texture information change happens, in an image of a current frame, illumination rapid-change pixels of pixels, in which texture information change does not happen, in to-be-detected moving-object pixels and illumination gradient pixels, which are not illumination rapid-change pixels and in which texture information change does not happen, in the to-be-detected moving object pixels; and determining that pixels, the similarity of which with a comparison pixel in a corresponding background image in the pixels, in which texture information change happens, is larger than a first threshold, in the to-be-detected moving object pixels, pixels, the similarity of which with a comparison pixel in a corresponding background image is larger than a forth threshold, in the illumination rapid-change pixels and pixels, the similarity of which with a comparison pixel in a corresponding background image, is larger than a fifth threshold, in the illumination gradient pixels are invalid moving-object pixels.

Description

A kind of method and apparatus of definite lost motion object pixel
Technical field
The present invention relates to intelligent video content analysis techniques field, relate in particular to a kind of method and apparatus of definite lost motion object pixel.
Background technology
Moving object detection is the basis of Intelligent Video Surveillance Technology, and it processes all extremely important to succeeding target classification, tracking and behavioural analysis.The detection of moving target is in order to remove background, from image sequence, interested moving region is extracted.
At present, there is the moving object detection of several different methods for image, as the frame differential method of image sequence, background subtraction method etc.
The frame differential method of image sequence is that the grey scale change by detecting between adjacent two two field pictures detects moving target.Specific practice is: by the f of k two field picture k(x, y) and k-1 two field picture f k-1(x, y) does calculus of differences, directly the difference of more continuous two width video frame images; Wherein f k(x, y) is the gray-scale value of the pixel of (x, y) position in k two field picture, wherein f k-1(x, y) is the gray-scale value of the pixel of (x, y) position in k-1 two field picture.By setting threshold T, obtain binaryzation difference image D (x, y):
D ( x , y ) = 1 , f k ( x , y ) - f k - 1 ( x , y ) > T 0 , f k ( x , y ) - f k - 1 ( x , y ) ≤ T
In binaryzation difference image D (x, y), pixel value represents that this pixel does not change in two frames of front and back, i.e. background pixel for " 0 "; Pixel value represents that variation has occurred this pixel in two frames of front and back, i.e. foreground pixel for " 1 ".
Background subtraction method is that current video image and the background image of storing in advance or obtain are in real time done to calculus of differences, determines according to comparative result whether pixel belongs to foreground pixel.There are some background subtraction methods to carry out real-time update to background image according to testing result, make the variation that background can adaptive environment, as classical ADAPTIVE MIXED Gaussian Background modeling.
But; in actual applications; the situation that before and after often there will be, the illumination of two two field pictures changes; adopt in the definite foreground pixel of the frame differential method of image sequence and background subtraction method and not only comprise real moving target pixel; also comprise the lost motion object pixel causing due to illumination variation, can affect like this accuracy and the reliability of moving object detection.
In sum, in the definite foreground pixel of existing moving target detecting method, not only comprise real moving target pixel, also comprise the lost motion object pixel causing due to illumination variation, this can cause moving object detection inaccurate.
Summary of the invention
The embodiment of the present invention provides a kind of method and apparatus of definite lost motion object pixel, detects inaccurate problem in order to solve existing moving target detecting method.
Based on the problems referred to above, the method for a kind of definite lost motion object pixel that the embodiment of the present invention provides, comprising:
For the pixel that texture information variation occurs in the moving target pixel to be detected in current frame image, determine the similarity of pixel that this generation texture information changes and corresponding compared pixels, compared pixels corresponding to pixel that texture information changes wherein occurs be to be arranged in background image and the pixel identical with the position of the current frame image at the pixel place of this generation texture information variation;
Determine that in moving target pixel to be detected, occurring in pixel that texture information changes to be greater than with the similarity of corresponding compared pixels the pixel that the described generation texture information of first threshold changes is lost motion object pixel.
The device of a kind of definite lost motion object pixel that the embodiment of the present invention provides, comprises determination module and confirms module;
Described determination module, for the pixel that texture information variation occurs in the moving target pixel to be detected in current frame image, determine the similarity of pixel that this generation texture information changes and corresponding compared pixels, compared pixels corresponding to pixel that texture information changes wherein occurs be to be arranged in background image and the pixel identical with the position of the current frame image at the pixel place of this generation texture information variation;
Described confirmation module is lost motion object pixel for determining that moving target pixel to be detected occurs in pixel that texture information changes to be greater than with the similarity of corresponding compared pixels the pixel that the described generation texture information of first threshold changes.
The beneficial effect of the embodiment of the present invention comprises:
The method and apparatus of definite lost motion object pixel that the embodiment of the present invention provides, by determining the moving target pixel to be detected in current frame image, be that the pixel that texture information changes occurs in foreground pixel, there is the similarity of the pixel compared pixels corresponding with it that texture information changes in judgement, if the pixel that this generation texture information changes and the similarity of compared pixels are greater than first threshold, confirm that the pixel that this generation texture information changes is lost motion object pixel, thereby determine the real moving target pixel in this moving target pixel to be detected, and then accuracy and the reliability of raising moving object detection.
Accompanying drawing explanation
One of process flow diagram of the method for definite lost motion object pixel that Fig. 1 provides for the embodiment of the present invention;
Fig. 2 provides the process flow diagram that whether texture information of fixation element changes really for the embodiment of the present invention;
Two of the process flow diagram of the method for definite lost motion object pixel that Fig. 3 provides for the embodiment of the present invention;
Three of the process flow diagram of the method for definite lost motion object pixel that Fig. 4 provides for the embodiment of the present invention;
Four of the process flow diagram of the method for definite lost motion object pixel that Fig. 5 provides for the embodiment of the present invention;
Five of the process flow diagram of the method for definite lost motion object pixel that Fig. 6 provides for the embodiment of the present invention;
The process flow diagram of the fast transshaping element of definite illumination that Fig. 7 provides for the embodiment of the present invention;
The process flow diagram of definite illumination gradual change pixel that Fig. 8 provides for the embodiment of the present invention;
Six of the process flow diagram of the method for definite lost motion object pixel that Fig. 9 provides for the embodiment of the present invention;
The process flow diagram of the illumination variation state of definite scene that Figure 10 provides for the embodiment of the present invention;
Figure 11 a-Figure 11 e is for using the method for definite lost motion object pixel of providing of the embodiment of the present invention to the figure obtaining in a certain two field picture A processing procedure;
The structure drawing of device of definite lost motion object pixel that Figure 12 provides for the embodiment of the present invention.
Embodiment
The embodiment of the present invention provides a kind of method and apparatus of definite lost motion object pixel, by confirming moving target pixel to be detected, causing the pixel of grey scale change in foreground pixel due to illumination variation, thereby improve accuracy and the reliability of moving object detection.
Below in conjunction with Figure of description, the embodiment of the method and apparatus of a kind of definite lost motion object pixel that the embodiment of the present invention is provided describes.
The method of a kind of definite lost motion object pixel that the embodiment of the present invention provides, as shown in Figure 1, specifically comprises the following steps:
S101, determine the pixel that texture information changes occurs in moving target pixel to be detected in current frame image;
S102, in the moving target pixel to be detected in current frame image, occur texture information change a pixel, determine the similarity of pixel that this generation texture information changes and corresponding compared pixels, compared pixels corresponding to pixel that texture information changes wherein occurs be to be arranged in background image and the pixel identical with the position of the current frame image at the pixel place of this generation texture information variation;
S103, determine that the pixel that the similarity that pixel that texture information the changes compared pixels corresponding with it occur in moving target pixel to be detected is greater than the described generation texture information variation of first threshold is lost motion object pixel.
In actual applications, can set up background model according to the image of input, obtain background image, background image unceasing study upgrades, and the target of not moving for a long time will be learned in background image.Then obtain moving target pixel to be detected by background image and current frame image, i.e. foreground pixel, had both comprised real moving target pixel in this moving target pixel to be detected, also comprised lost motion object pixel.
Determine that the pixel that texture information changes occurs in the moving target pixel to be detected in current frame image both can first determine the moving target pixel to be detected in current frame image, then determine the pixel that texture information variation occurs in moving target pixel to be detected, also can first determine in current frame image the pixel that texture information changes occurs, then determine the moving target pixel to be detected in the pixel that texture information variation occurs.
Particularly, judge whether pixel texture information occurs and change, and as shown in Figure 2, comprises the following steps:
S201, for a pixel, determine the compared pixels that this pixel is corresponding, compared pixels corresponding to this pixel is the N two field picture being arranged in before current frame image, and the pixel identical with the position of the current frame image at this pixel place;
S202, judge whether the regional average value poor ratio of the poor compared pixels with this pixel of this pixel in current frame image regional average value based on neighborhood based on neighborhood is not more than Second Threshold or is not less than the 3rd threshold value, wherein, the regional average value of pixel based on neighborhood is poor is the absolute value of the difference of the average gray of the gray-scale value of this pixel and the pixel of this neighborhood of pixels;
If so, carry out S203, otherwise, S204 carried out;
S203, determine that this pixel generation texture information changes;
S204, determine this pixel do not occur texture information change.
In concrete application, the regional average value of pixel based on neighborhood is poor can be done and differ from and take absolute value by calculating the average gray of the gray-scale value of this pixel and the pixel of its neighborhood, and specific formula for calculation is as follows:
D i , j = | P i , j - Σ m , n P m , n 9 | Formula 1
Wherein, D i, jfor the pixel of (i, j) position in the image regional average value based on neighborhood poor, P i, jfor the gray-scale value of the pixel of (i, j) position in this image, P m, nfor the gray-scale value of the pixel of (m, n) position in this image; Wherein, the pixel of (m, n) position is positioned at (i, j) neighborhood of the pixel of position, in the time getting the pixel of 3 × 3 neighborhoods of pixel of (i, j) position, (m, n) value is (i-1, j-1), (i, j-1), (i+1, j-1), (i-1, j), (i, j), (i+1, j), (i-1, j+1), (i, j+1), (i+1, j+1).Regional average value in the pixel that calculates (i, j) position in current frame image based on neighborhood is poor and compared pixels, before current frame image, in N two field picture, the regional average value of the pixel of (i, j) position based on neighborhood is poor
Figure BDA00002418397300053
afterwards, judgement
Figure BDA00002418397300054
whether be not more than Second Threshold or be not less than the 3rd threshold value; If so, the pixel generation texture information of (i, j) position changes in current frame image; Otherwise in current frame image there is not the variation of texture information in the pixel of (i, j) position.
The method of a kind of definite lost motion object pixel that further, the embodiment of the present invention provides can also be determined the fast transshaping element of illumination in the pixel that texture information variation does not occur in moving target pixel to be detected in current frame image and judge the lost motion object pixel in the fast transshaping element of illumination.As shown in Figure 3, the method also comprises:
S301, determine the fast transshaping element of illumination not occurring in the moving target pixel to be detected in current frame image in pixel that texture information changes;
S302, for a fast transshaping element of illumination, determine the similarity of the fast transshaping of this illumination element and corresponding compared pixels, wherein compared pixels corresponding to the fast transshaping element of illumination is to be arranged in background image and the pixel identical with the position of the current frame image at the plain place of the fast transshaping of this illumination;
S303, determine in the fast transshaping element of illumination that the fast transshaping element of illumination that is greater than the 4th threshold value with the similarity of corresponding compared pixels is lost motion object pixel.
Adopt said method, can further determine the fast transshaping element of illumination in the pixel that texture information variation does not occur in moving target pixel to be detected, and judge the lost motion object pixel in the fast transshaping element of illumination according to the similarity of the fast transshaping element of the illumination compared pixels corresponding with it, thereby judge more exactly the moving target of the relative background image of current frame image.
Further, the method of a kind of definite lost motion object pixel that the embodiment of the present invention provides can also be determined and in current frame image, in moving target pixel to be detected, not occur that texture information changes and be not the illumination gradual change pixel in the pixel of the fast transshaping element of illumination, and determines the lost motion object pixel in illumination gradual change pixel.As shown in Figure 4, the method also comprises:
S401, determine and in the moving target pixel to be detected in current frame image, do not occur that texture information changes and be not the illumination gradual change pixel in the pixel of the fast transshaping element of illumination;
S402, for an illumination gradual change pixel, determine the similarity of this illumination gradual change pixel and corresponding compared pixels, wherein compared pixels corresponding to illumination gradual change pixel is to be arranged in background image and the pixel identical with the position of the current frame image at this illumination gradual change pixel place;
S403, determine in illumination gradual change pixel that the illumination gradual change pixel that is greater than the 5th threshold value with the similarity of corresponding compared pixels is lost motion object pixel.
Adopt said method, can further determine the illumination gradual change pixel in the pixel that texture information variation does not occur in moving target pixel to be detected, and judge the lost motion object pixel in illumination gradual change pixel according to the similarity of the illumination gradual change pixel compared pixels corresponding with it, thereby judge more exactly the moving target of the relative background image of current frame image.
In the method for above-mentioned a kind of definite lost motion object pixel, adopt first to judge and in moving target pixel to be detected in current frame image, do not occur whether each pixel that texture information changes is the fast transshaping element of illumination, and then judge and texture information variation does not occur in the moving target pixel to be detected in current frame image and be not whether each pixel of the fast transshaping element of illumination is the order of illumination gradual change pixel; Can certainly adopt in the moving target pixel to be detected first judging in current frame image, not occur whether each pixel that texture information changes is illumination gradual change pixel, and then judge and texture information variation does not occur in the moving target pixel to be detected in current frame image and be not whether each pixel of illumination gradual change pixel is the order of the fast transshaping element of illumination.
First judge and in moving target pixel to be detected in current frame image, do not occur whether each pixel that texture information changes is illumination gradual change pixel when adopting, and then judge and in the moving target pixel to be detected in current frame image, do not occur that texture information changes and be not whether each pixel of illumination gradual change pixel is while being the order of the fast transshaping element of illumination, the method of a kind of definite lost motion object pixel that the embodiment of the present invention provides, as shown in Figure 5, also comprise:
S501, determine the illumination gradual change pixel not occurring in the moving target pixel to be detected in current frame image in pixel that texture information changes;
S502, for an illumination gradual change pixel, determine the similarity of this illumination gradual change pixel and corresponding compared pixels, wherein compared pixels corresponding to illumination gradual change pixel is to be arranged in background image and the pixel identical with the position of the current frame image at this illumination gradual change pixel place;
S503, determine in illumination gradual change pixel that the illumination gradual change pixel that is greater than the 5th threshold value with the similarity of corresponding compared pixels is lost motion object pixel.
Further, the method for a kind of definite lost motion object pixel that the embodiment of the present invention provides, as shown in Figure 6, also comprises:
S601, determine and in the moving target pixel to be detected in current frame image, do not occur that texture information changes and be not the fast transshaping element of illumination in the pixel of illumination gradual change pixel;
S602, for a fast transshaping element of illumination, determine the similarity of the fast transshaping of this illumination element and corresponding compared pixels, wherein compared pixels corresponding to the fast transshaping element of illumination is to be arranged in background image and the pixel identical with the position of the current frame image at the plain place of the fast transshaping of this illumination;
S603, determine in the fast transshaping element of illumination that the fast transshaping element of illumination that is greater than the 4th threshold value with the similarity of corresponding compared pixels is lost motion object pixel.
Particularly, as shown in Figure 7, judge that whether pixel is that the process of the fast transshaping element of illumination comprises the following steps:
S701, for a pixel in current frame image, determine this pixel respective pixel of the M in continuous N two field picture before current frame image, wherein respective pixel is to be arranged in any two field picture of continuous N two field picture before current frame image and the pixel identical with the position of the current frame image at this pixel place;
S702, judge whether the absolute value of the difference that is arranged in the gray-scale value of two respective pixel of arbitrary neighborhood two two field pictures in M respective pixel is all greater than the 6th threshold value; If so, carry out S703; Otherwise; Carry out S704;
S703, determine that this pixel is the fast transshaping element of illumination;
S704, determine that this pixel is not the fast transshaping element of illumination.
In the time of concrete application, for a pixel in current frame image, for example pixel of (i, j) position in current frame image, determines before current frame image in each two field picture in continuous N two field picture that the pixel of (i, j) position is all the respective pixel of this pixel; Calculate in M respective pixel, be arranged in the gray-scale value of two respective pixel of arbitrary neighborhood two two field pictures difference thoroughly deserve M-1 value, judge whether this M-1 value is all greater than the 6th threshold value; If so, this pixel that is arranged in current frame image is the fast transshaping element of illumination; Otherwise this pixel is not the fast transshaping element of illumination.
Particularly, as shown in Figure 8, judge that whether pixel is that the process of the pixel of illumination gradual change comprises the following steps:
S801, for a pixel in current frame image, determine the compared pixels that this pixel is corresponding, compared pixels corresponding to this pixel is to be arranged in background image and the pixel identical with the position of the current frame image at this pixel place;
S802, judge whether the absolute value of the difference of the average gray of the pixel of the neighborhood of the compared pixels that the average gray of pixel of the neighborhood of this pixel is corresponding with this pixel is greater than the 7th threshold value, if so, carries out S803, otherwise, S805 carried out;
S803, judge whether the absolute value of the difference of the Grad of the gray scale of the compared pixels that the Grad of gray scale of this pixel is corresponding with this pixel is less than the 8th threshold value, and wherein the Grad of pixel is the Grad of horizontal direction and the Grad sum of vertical direction of this pixel; If so, carry out S804, otherwise, S805 carried out;
S804, determine that this pixel is illumination gradual change pixel;
S805, determine that this pixel is not illumination gradual change pixel.
In the time of concrete application, for a pixel in current frame image, for example (i in current frame image, j) pixel of position, calculate the average gray of the pixel of the neighborhood of this pixel, if get the average gray of the pixel of these pixel 3 × 3 neighborhoods, calculate (i-1 in current frame image, j-1), (i, j-1), (i+1, j-1), (i-1, j), (i, j), (i+1, j), (i-1, j+1), (i, and (i+1 j+1), j+1) gray-scale value of 9 locational pixels only and, then this gray-scale value sum is obtained to (i in current frame image divided by 9, j) average gray of the pixel of the neighborhood of the pixel of position, then the similar process computation of basis goes out the compared pixels of this pixel, the i.e. average gray of the pixel of the neighborhood of the pixel of (i, j) position in background image.
The gradient of the gray scale of pixel is G x+ G y, wherein, G xfor the gradient of the horizontal direction of the gray scale of pixel, the gradient of the vertical direction of the gray scale that Gy is pixel.In current frame image, the value of the gradient of the gray scale of the pixel of (i, j) position is
Figure BDA00002418397300091
in background image, the value of the gradient of the gray scale of the pixel of (i, j) position is
Figure BDA00002418397300092
therefore, in current frame image, the absolute value of the difference of the value of the gradient of the value of the gradient of the gray scale of the pixel of (i, j) position and the gray scale of its compared pixels is:
| ( G x + G y ) ( i , j ) f - ( G x + G y ) ( i , j ) g | Formula 2
Preferably, as shown in Figure 9, the method for a kind of definite lost motion object pixel that the embodiment of the present invention provides also comprises the following steps:
S901, according to the number of the pixel of illumination gradual change in the moving target pixel to be detected in number and the current frame image of the fast transshaping element of illumination in the moving target pixel to be detected occurring in current frame image in number, the current frame image of pixel that texture information changes, determine the illumination variation state of current scene;
S902, according to the corresponding relation of illumination variation state and background image renewal speed, for current scene is set corresponding background image renewal speed.
Like this, can improve the ability that background image adapts to illumination variation state, thereby in the time of the moving target pixel to be detected of determining in subsequent frame image, reduce corresponding lost motion object pixel, thereby improve authenticity and the reliability of moving object detection.
Particularly, as shown in figure 10, determine that the illumination variation state of current scene comprises the following steps:
S1001, determine the number of illumination gradual change pixel in the moving target pixel to be detected in number and the current frame image of the fast transshaping element of illumination in the moving target pixel to be detected in number, the current frame image of the pixel that texture information in current frame image changes;
S1002, judge whether the number of the fast transshaping element of illumination in the moving target pixel to be detected in current frame image is greater than the 9th threshold value, if so, carries out S1003, otherwise, S1004 carried out;
S1003, determine that the illumination variation state of current scene is illumination fast state;
Whether the number whether S1004, the number that judges illumination gradual change pixel in the moving target pixel to be detected in current frame image are greater than the pixel that in the tenth threshold value and current frame image, texture information changes is greater than the 11 threshold value, if the number of this illumination gradual change pixel is greater than the number of the pixel of the tenth threshold value and the variation of this generation texture information and is greater than the 11 threshold value, carry out S1005, otherwise, carry out S1006;
S1005, determine that the illumination variation state of current scene is illuminance abrupt variation state;
S1006, judge whether the number of illumination gradual change pixel in the moving target pixel to be detected in current frame image is greater than the 12 threshold value, and whether the number of the pixel that in current frame image, texture information changes is less than the 13 threshold value, if the number that in the moving target pixel to be detected in current frame image, the number of illumination gradual change pixel is greater than the pixel that texture information in the 12 threshold value and current frame image changes is less than the 13 threshold value, carry out S1007, otherwise, carry out S1008;
S1007, determine that the illumination variation state of current scene is illumination gradual change state;
S1008, determine that the illumination variation state of current scene is illumination steady state (SS).
In the process of the illumination variation state of above-mentioned definite current scene, S1004 and S1006 execution sequence can exchange.
Further, after lost motion object pixel in the fast transshaping element of lost motion object pixel, illumination of determining in the pixel that in current frame image, the texture information in moving target pixel to be detected changes and the lost motion object pixel in illumination gradual change pixel, can the removal from moving target pixel to be detected by lost motion object pixel.Then, foreground image is carried out to morphologic filtering processing, wherein, moving target pixel to be detected is white pixel in foreground image, the pixel that non-moving target pixel to be detected is black, if the pixel of adularescent in the foreground image after processing through morphologic filtering, the relative background image of current frame image has moving target, otherwise the relative background image of current frame image does not have moving target.
In order to further illustrate the method for a kind of definite lost motion object pixel that the embodiment of the present invention provides, illustrate with the scene of some concrete application how the method realizes below.
First, set up background model according to the image of input, then learn to obtain background image by background, as shown in Figure 11 a.Then by current frame image and background image contrast, obtain foreground image, in this application scenarios, current frame image is the gray-scale map of a certain frame input picture A, as shown in Figure 11 b, the brightness of background image and current frame image has obvious difference, i.e. variation has occurred in illumination.The foreground image obtaining, as shown in Figure 11 c, each pixel in foreground image is divided into two classes, and a class pixel is white, and another kind of pixel is black.Being arranged in current frame image and the pixel identical with white location of pixels is moving target pixel to be detected, i.e. foreground pixel; Being arranged in current frame image and the pixel identical with the location of pixels of black is non-moving target pixel, i.e. background pixel.
Then according to the flow process shown in Fig. 2, the illumination variation of each pixel in current frame image is carried out to preliminary detection judgement.Here can be rule of thumb Second Threshold and the 3rd threshold value choosing value, in this application scenarios, Second Threshold be got 0.8, the three threshold value and is got 1.2,
Figure BDA00002418397300111
or
Figure BDA00002418397300112
current frame image in pixel be occur texture information change pixel, can label 2 be set for this pixel;
Figure BDA00002418397300113
Figure BDA00002418397300114
current frame image in pixel be do not occur texture information change pixel, can label 0 be set for this pixel;
And then determine the fast transshaping element of illumination in the pixel that in the moving target pixel to be detected in current frame image, label is 0 according to the flow process shown in Fig. 7.Rule of thumb the span of the 6th threshold value is [10,30], in this concrete scene, the 6th threshold value gets 20, M gets 6, the absolute value of the pixel that in moving target pixel even to be detected, label is 0 difference of the gray-scale value of the respective pixel in arbitrary neighborhood two two field pictures in 6 two field pictures before present frame is all greater than 20, and the pixel that this label is 0 is so the fast transshaping element of illumination, and the label of the pixel that can be 0 by this label is revised as 4; Otherwise the pixel that this label the is 0 fast transshaping element that is not illumination, does not revise this label and is the label of 0 pixel.
Then, then according to the flow process shown in Fig. 8 determine the illumination gradual change pixel in the pixel that in the moving target pixel to be detected in current frame image, label is 0.In this concrete application scenarios, the 7th threshold value is got 3, the eight threshold values and is got 10.If the pixel that in this moving target pixel to be detected, label is 0 is illumination gradual change pixel, the label of the pixel that is 0 by this label is revised as 1; If the pixel that in this moving target pixel to be detected, label is 0 is not illumination gradual change pixel, do not revise this label and be the label of 0 pixel.
In the time of concrete application, also can first determine the illumination gradual change pixel in the pixel that in moving target pixel to be detected, label is 0, and then determine the fast transshaping element of illumination in the pixel that in moving target pixel to be detected, label is 0.
And then there is according to the flow process judgement shown in Fig. 1 the pixel that texture information changes, be that label is whether 2 pixel is lost motion object pixel, if lost motion object pixel, will be arranged in the white pixel of foreground image and the pixel identical with this lost motion target pixel location is revised as black.In this concrete application scenarios, first threshold gets 75%.
Then, judge the fast transshaping element of illumination according to the flow process in Fig. 3 or Fig. 6 again, be that label is whether 4 pixel is lost motion object pixel, if lost motion object pixel, will be arranged in the white pixel of foreground image and the pixel identical with this lost motion target pixel location is revised as black.In this concrete application scenarios, the 4th threshold value gets 70%.
Then, judge illumination gradual change pixel according to the flow process in Fig. 4 or Fig. 5 again, be that label is whether 1 pixel is lost motion object pixel, if lost motion object pixel, will be arranged in the white pixel of foreground image and the pixel identical with this lost motion target pixel location is revised as black.In this concrete application scenarios, the 4th threshold value gets 80%.Finally obtain the foreground image as shown in Figure 11 d, then the foreground image shown in Figure 11 d is carried out to morphologic filtering, then according to being arranged in current frame image, and the pixel identical with the position of the white pixel in filtering foreground image afterwards determined the moving target of the relative background image of image A, as shown in Figure 11 e, in this image A, the moving target of determining is lived with square frame circle.
After obtaining the label of each pixel in current frame image, statistical labeling is 1 respectively, label is 2 and the number of the label pixel that is 4, then according to the flow process shown in Figure 10, judges the illumination variation state of current scene.In this concrete application scenarios, when the number of the label pixel that is 4 be greater than current frame image number of pixels 20% time, the illumination variation state of determining current scene is illumination fast state; When the number of the label pixel that is 4 be not more than current frame image number of pixels 20%, and the number of the pixel that label is 1 be greater than current frame image number of pixels 10%, label be the number of 2 the pixel number of pixels that is less than current frame image 20% time, the illumination variation state of determining current scene is illumination gradual change state; When the number of the label pixel that is 4 be not more than current frame image number of pixels 20%, and the number of the pixel that label is 1 be greater than current frame image number of pixels 10%, label be the number of 2 the pixel number of pixels that is greater than current frame image 20% time, the illumination variation state of determining current scene is illuminance abrupt variation state; Otherwise the illumination variation state of determining current scene is illumination steady state (SS).
Then after having determined the illumination variation state of current scene, upgrade background image according to the corresponding relation of the illumination variation state of scene and background image renewal speed.The renewal speed of the illumination variation state of scene from illumination steady state (SS), illumination gradual change state, illumination fast state to background image corresponding to illuminance abrupt variation state is from fastest to the slowest; Be the context update speed that the context update speed of the stable scene of illumination is greater than the scene of illumination gradual change, the context update speed of the scene of illumination gradual change is greater than the context update speed of the fast scene becoming of illumination, and the context update speed of the fast scene becoming of illumination is greater than the context update speed of the scene of illuminance abrupt variation.
Based on same inventive concept, the embodiment of the present invention also provides a kind of device of definite lost motion object pixel, because the principle that these devices are dealt with problems is similar to the method for aforementioned definite lost motion object pixel, therefore the enforcement of this device can be referring to the enforcement of preceding method, repeats part and repeat no more.
The embodiment of the present invention also provides a kind of device of definite lost motion object pixel, and as shown in figure 12, this device comprises determination module 121 and confirms module 122:
Determination module 121, the pixel changing for there is texture information in the moving target pixel to be detected for current frame image, determine the similarity of pixel that this generation texture information changes and corresponding compared pixels, compared pixels corresponding to pixel that texture information changes wherein occurs be to be arranged in background image and the pixel identical with the position of the current frame image at the pixel place of this generation texture information variation;
Confirming module 122, is lost motion object pixel for determining that moving target pixel to be detected occurs in pixel that texture information changes to be greater than with the similarity of corresponding compared pixels the pixel that the described generation texture information of first threshold changes.
Further, determination module 121, also for for a pixel, judge whether the regional average value poor ratio of the poor compared pixels with this pixel of this pixel in current frame image regional average value based on neighborhood based on neighborhood is not more than Second Threshold and is not less than the 3rd threshold value, wherein, the compared pixels of this pixel is the N two field picture being arranged in before current frame image, and the pixel identical with the position of the current frame image at this pixel place; The regional average value of pixel based on neighborhood is poor is the absolute value of the difference of the average gray of the gray-scale value of this pixel and the pixel of this neighborhood of pixels; If so, determine that this pixel generation texture information changes; Otherwise, determine that this pixel texture information does not occur and changes.
Further, determination module 121, does not also occur in the moving target pixel to be detected of current frame image for judging whether each pixel that texture information changes is the fast transshaping element of illumination; And for a fast transshaping element of illumination, determine the similarity of the fast transshaping of this illumination element and corresponding compared pixels, wherein compared pixels corresponding to the fast transshaping element of illumination is to be arranged in background image and the pixel identical with the position of the current frame image at the plain place of the fast transshaping of this illumination;
Confirm module 122, also for determining that the fast transshaping element of illumination that the fast transshaping element of illumination and the similarity of corresponding compared pixels are greater than the 4th threshold value is lost motion object pixel.
Further, determination module 121, does not also occur in the moving target pixel to be detected of current frame image that texture information changes and is not whether each pixel of the fast transshaping element of illumination is illumination gradual change pixel for judging; And for an illumination gradual change pixel, determine the similarity of this illumination gradual change pixel and corresponding compared pixels, wherein compared pixels corresponding to illumination gradual change pixel is to be arranged in background image and the pixel identical with the position of the current frame image at this illumination gradual change pixel place;
Confirm module 122, also for determining that illumination gradual change pixel is lost motion object pixel with the illumination gradual change pixel that the similarity of corresponding compared pixels is greater than the 5th threshold value.
Or determination module 121, does not also occur in the moving target pixel to be detected of current frame image for judging whether each pixel that texture information changes is illumination gradual change pixel; And for an illumination gradual change pixel, determine the similarity of this illumination gradual change pixel and corresponding compared pixels, wherein compared pixels corresponding to illumination gradual change pixel is to be arranged in background image and the pixel identical with the position of the current frame image at this illumination gradual change pixel place;
Confirm module 122, also for determining that illumination gradual change pixel is lost motion object pixel with the illumination gradual change pixel that the similarity of corresponding compared pixels is greater than the 5th threshold value.
Further, determination module 121, does not also occur in the moving target pixel to be detected of current frame image that texture information changes and is not whether each pixel of illumination gradual change pixel is the fast transshaping element of illumination for judging; And for a fast transshaping element of illumination, determine the similarity of the fast transshaping of this illumination element and corresponding compared pixels, wherein compared pixels corresponding to the fast transshaping element of illumination is to be arranged in background image and the pixel identical with the position of the current frame image at the plain place of the fast transshaping of this illumination;
Confirm module 122, also for determining that the fast transshaping element of illumination that the fast transshaping element of illumination and the similarity of corresponding compared pixels are greater than the 4th threshold value is lost motion object pixel.
Particularly, determination module 121, also for for a pixel, from the respective pixel of this pixel, select two respective pixel, the image at two respective pixel places of wherein selecting is two two field pictures of arbitrary neighborhood, and the respective pixel of this pixel is to be arranged in any two field picture of continuous N two field picture before current frame image and the pixel identical with the position of the current frame image at this pixel place; If the absolute value of the difference of the gray-scale value of two respective pixel of described selection is all greater than the 6th threshold value, determine that this pixel is the fast transshaping element of illumination; Otherwise, determine that this pixel is not the fast transshaping element of illumination.
Particularly, determination module 121, also for for a pixel, if the absolute value of the difference of the average gray of the pixel of the neighborhood of the compared pixels that the average gray of the pixel of the neighborhood of this pixel in current frame image is corresponding with this pixel is greater than the 7th threshold value, and the absolute value of the difference of the Grad of the gray scale of the compared pixels that the Grad of the gray scale of this pixel is corresponding with this pixel is less than the 8th threshold value, determine that this pixel is illumination gradual change pixel, wherein compared pixels corresponding to this pixel is to be arranged in background image, and the pixel identical with the position of the current frame image at this pixel place, the Grad of pixel is the Grad of horizontal direction and the Grad sum of vertical direction of the gray scale of this pixel, otherwise, determine that this pixel is not illumination gradual change pixel.
Preferably, confirm module 122, also, for there is the number of the pixel of illumination gradual change in the moving target pixel to be detected in number and the current frame image of the fast pixel becoming of illumination in the moving target pixel to be detected in number, the current frame image of pixel that texture information changes according to current frame image, determine the illumination variation state of current scene; And according to the corresponding relation of illumination variation state and background image renewal speed, for current scene is set corresponding background image renewal speed.
Particularly, confirm the illumination variation state of module 122 definite current scene, comprising:
If the number of the fast transshaping element of illumination is greater than the 9th threshold value in the moving target pixel to be detected in current frame image, the illumination variation state of current scene is illumination fast state;
If the number that in the moving target pixel to be detected in current frame image, the number of the fast transshaping element of illumination is not more than illumination gradual change pixel in the moving target pixel to be detected in the 9th threshold value and current frame image is greater than the number of the pixel that in the tenth threshold value and current frame image, texture information changes and is greater than the 11 threshold value, the illumination variation state of current scene is illuminance abrupt variation state;
If the number that in the moving target pixel to be detected in current frame image, the number of the fast transshaping element of illumination is not more than illumination gradual change pixel in the moving target pixel to be detected in the 9th threshold value and current frame image is greater than the number of the pixel that in the 12 threshold value and current frame image, texture information changes and is less than the 13 threshold value, the illumination variation state of current scene is illumination gradual change state; Otherwise, the illumination steady state (SS) of current scene.
Through the above description of the embodiments, those skilled in the art can be well understood to the embodiment of the present invention and can realize by hardware, and the mode that also can add necessary general hardware platform by software realizes.Based on such understanding, the technical scheme of the embodiment of the present invention can embody with the form of software product, it (can be CD-ROM that this software product can be stored in a non-volatile memory medium, USB flash disk, portable hard drive etc.) in, comprise that some instructions are in order to make a computer equipment (can be personal computer, server, or the network equipment etc.) carry out the method described in each embodiment of the present invention.
It will be appreciated by those skilled in the art that accompanying drawing is the schematic diagram of a preferred embodiment, the module in accompanying drawing or flow process might not be that enforcement the present invention is necessary.
It will be appreciated by those skilled in the art that the module in the device in embodiment can be distributed in the device of embodiment according to embodiment description, also can carry out respective change and be arranged in the one or more devices that are different from the present embodiment.The module of above-described embodiment can be merged into a module, also can further split into multiple submodules.
The invention described above embodiment sequence number, just to describing, does not represent the quality of embodiment.
Obviously, those skilled in the art can carry out various changes and modification and not depart from the spirit and scope of the present invention the present invention.Like this, if within of the present invention these are revised and modification belongs to the scope of the claims in the present invention and equivalent technologies thereof, the present invention is also intended to comprise these changes and modification interior.

Claims (20)

1. a method for definite lost motion object pixel, is characterized in that, comprising:
For the pixel that texture information variation occurs in the moving target pixel to be detected in current frame image, determine the similarity of pixel that this generation texture information changes and corresponding compared pixels, compared pixels corresponding to pixel that texture information changes wherein occurs be to be arranged in background image and the pixel identical with the position of the current frame image at the pixel place of this generation texture information variation;
Determine that in moving target pixel to be detected, occurring in pixel that texture information changes to be greater than with the similarity of corresponding compared pixels the pixel that the described generation texture information of first threshold changes is lost motion object pixel.
2. the method for claim 1, is characterized in that, judges according to the following step whether pixel texture information occurs and change:
For a pixel, judge whether the regional average value poor ratio of the poor compared pixels with this pixel of this pixel in current frame image regional average value based on neighborhood based on neighborhood is not more than Second Threshold or is not less than the 3rd threshold value, wherein, the compared pixels of this pixel is the N two field picture being arranged in before current frame image, and the pixel identical with the position of the current frame image at this pixel place; The regional average value of pixel based on neighborhood is poor is the absolute value of the difference of the average gray of the gray-scale value of this pixel and the pixel of this neighborhood of pixels;
If so, determine that this pixel generation texture information changes; Otherwise, determine that this pixel texture information does not occur and changes.
3. the method for claim 1, is characterized in that, the method also comprises:
Judge and in the moving target pixel to be detected in current frame image, do not occur whether each pixel that texture information changes is the fast transshaping element of illumination;
For a fast transshaping element of illumination, determine the similarity of the fast transshaping of this illumination element and corresponding compared pixels, wherein compared pixels corresponding to the fast transshaping element of illumination is to be arranged in background image and the pixel identical with the position of the current frame image at the plain place of the fast transshaping of this illumination;
Determine in the fast transshaping element of illumination that the fast transshaping element of illumination that is greater than the 4th threshold value with the similarity of corresponding compared pixels is lost motion object pixel.
4. method as claimed in claim 3, is characterized in that, the method also comprises:
Judge and in the moving target pixel to be detected in current frame image, do not occur that texture information changes and be not whether each pixel of the fast transshaping element of illumination is illumination gradual change pixel;
For an illumination gradual change pixel, determine the similarity of this illumination gradual change pixel and corresponding compared pixels, wherein compared pixels corresponding to illumination gradual change pixel is to be arranged in background image and the pixel identical with the position of the current frame image at this illumination gradual change pixel place;
Determine in illumination gradual change pixel that the illumination gradual change pixel that is greater than the 5th threshold value with the similarity of corresponding compared pixels is lost motion object pixel.
5. the method for claim 1, is characterized in that, the method also comprises:
Judge and in the moving target pixel to be detected in current frame image, do not occur whether each pixel that texture information changes is illumination gradual change pixel;
For an illumination gradual change pixel, determine the similarity of this illumination gradual change pixel and corresponding compared pixels, wherein compared pixels corresponding to illumination gradual change pixel is to be arranged in background image and the pixel identical with the position of the current frame image at this illumination gradual change pixel place;
Determine in illumination gradual change pixel that the illumination gradual change pixel that is greater than the 5th threshold value with the similarity of corresponding compared pixels is lost motion object pixel.
6. method as claimed in claim 5, is characterized in that, the method also comprises:
Judge and in the moving target pixel to be detected in current frame image, do not occur that texture information changes and be not whether each pixel of illumination gradual change pixel is the fast transshaping element of illumination;
For a fast transshaping element of illumination, determine the similarity of the fast transshaping of this illumination element and corresponding compared pixels, wherein compared pixels corresponding to the fast transshaping element of illumination is to be arranged in background image and the pixel identical with the position of the current frame image at the plain place of the fast transshaping of this illumination;
Determine in the fast transshaping element of illumination that the fast transshaping element of illumination that is greater than the 4th threshold value with the similarity of corresponding compared pixels is lost motion object pixel.
7. the method as described in claim 3 or 6, is characterized in that, judges according to the following step whether pixel is the fast transshaping element of illumination:
For a pixel, from the respective pixel of this pixel, select two respective pixel, the image at two respective pixel places of wherein selecting is two two field pictures of arbitrary neighborhood, and the respective pixel of this pixel is to be arranged in any two field picture of continuous N two field picture before current frame image and the pixel identical with the position of the current frame image at this pixel place;
If the absolute value of the difference of the gray-scale value of two respective pixel of described selection is all greater than the 6th threshold value, determine that this pixel is the fast transshaping element of illumination; Otherwise, determine that this pixel is not the fast transshaping element of illumination.
8. the method as described in claim 4 or 5, is characterized in that, judges according to the following step whether pixel is the pixel of illumination gradual change:
For a pixel, if the absolute value of the difference of the average gray of the pixel of the neighborhood of the compared pixels that the average gray of the pixel of the neighborhood of this pixel in current frame image is corresponding with this pixel is greater than the 7th threshold value, and the absolute value of the difference of the Grad of the gray scale of the compared pixels that the Grad of the gray scale of this pixel is corresponding with this pixel is less than the 8th threshold value, determine that this pixel is illumination gradual change pixel, wherein compared pixels corresponding to this pixel is to be arranged in background image, and the pixel identical with the position of the current frame image at this pixel place, the Grad of pixel is the Grad of horizontal direction and the Grad sum of vertical direction of the gray scale of this pixel,
Otherwise, determine that this pixel is not illumination gradual change pixel.
9. the method as described in claim 4 or 6, is characterized in that, the method also comprises:
According to the number of the pixel of illumination gradual change in the moving target pixel to be detected in number and the current frame image of the fast pixel becoming of illumination in the moving target pixel to be detected occurring in current frame image in number, the current frame image of pixel that texture information changes, determine the illumination variation state of current scene;
According to the corresponding relation of illumination variation state and background image renewal speed, for current scene is set corresponding background image renewal speed.
10. method as claimed in claim 9, is characterized in that, determines the illumination variation state of current scene, comprising:
If the number of the fast transshaping element of illumination is greater than the 9th threshold value in the moving target pixel to be detected in current frame image, the illumination variation state of current scene is illumination fast state;
If the number that in the moving target pixel to be detected in current frame image, the number of the fast transshaping element of illumination is not more than illumination gradual change pixel in the moving target pixel to be detected in the 9th threshold value and current frame image is greater than the number of the pixel that in the tenth threshold value and current frame image, texture information changes and is greater than the 11 threshold value, the illumination variation state of current scene is illuminance abrupt variation state;
If the number that in the moving target pixel to be detected in current frame image, the number of the fast transshaping element of illumination is not more than illumination gradual change pixel in the moving target pixel to be detected in the 9th threshold value and current frame image is greater than the number of the pixel that in the 12 threshold value and current frame image, texture information changes and is less than the 13 threshold value, the illumination variation state of current scene is illumination gradual change state;
Otherwise, the illumination steady state (SS) of current scene.
The device of 11. 1 kinds of definite lost motion object pixels, is characterized in that, described device comprises determination module and confirms module;
Described determination module, the pixel changing for there is texture information in the moving target pixel to be detected for current frame image, determine the similarity of pixel that this generation texture information changes and corresponding compared pixels, compared pixels corresponding to pixel that texture information changes wherein occurs be to be arranged in background image and the pixel identical with the position of the current frame image at the pixel place of this generation texture information variation;
Described confirmation module is lost motion object pixel for determining that moving target pixel to be detected occurs in pixel that texture information changes to be greater than with the similarity of corresponding compared pixels the pixel that the described generation texture information of first threshold changes.
12. devices as claimed in claim 11, is characterized in that;
Described determination module, also for for a pixel, judge whether the regional average value poor ratio of the poor compared pixels with this pixel of this pixel in current frame image regional average value based on neighborhood based on neighborhood is not more than Second Threshold or is not less than the 3rd threshold value, wherein, the compared pixels of this pixel is the N two field picture being arranged in before current frame image, and the pixel identical with the position of the current frame image at this pixel place; The regional average value of pixel based on neighborhood is poor is the absolute value of the difference of the average gray of the gray-scale value of this pixel and the pixel of this neighborhood of pixels; If so, determine that this pixel generation texture information changes; Otherwise, determine that this pixel texture information does not occur and changes.
13. devices as claimed in claim 11, is characterized in that;
Described determination module, does not also occur in the moving target pixel to be detected of current frame image for judging whether each pixel that texture information changes is the fast transshaping element of illumination; And for a fast transshaping element of illumination, determine the similarity of the fast transshaping of this illumination element and corresponding compared pixels, wherein compared pixels corresponding to the fast transshaping element of illumination is to be arranged in background image and the pixel identical with the position of the current frame image at the plain place of the fast transshaping of this illumination;
Described confirmation module, also for determining that the fast transshaping element of illumination that the fast transshaping element of illumination and the similarity of corresponding compared pixels are greater than the 4th threshold value is lost motion object pixel.
14. devices as claimed in claim 13, is characterized in that;
Described determination module, does not also occur in the moving target pixel to be detected of current frame image that texture information changes and is not whether each pixel of the fast transshaping element of illumination is illumination gradual change pixel for judging; And for an illumination gradual change pixel, determine the similarity of this illumination gradual change pixel and corresponding compared pixels, wherein compared pixels corresponding to illumination gradual change pixel is to be arranged in background image and the pixel identical with the position of the current frame image at this illumination gradual change pixel place;
Described confirmation module, also for determining that illumination gradual change pixel is lost motion object pixel with the illumination gradual change pixel that the similarity of corresponding compared pixels is greater than the 5th threshold value.
15. devices as claimed in claim 11, is characterized in that;
Described determination module, does not also occur in the moving target pixel to be detected of current frame image for judging whether each pixel that texture information changes is illumination gradual change pixel; And for an illumination gradual change pixel, determine the similarity of this illumination gradual change pixel and corresponding compared pixels, wherein compared pixels corresponding to illumination gradual change pixel is to be arranged in background image and the pixel identical with the position of the current frame image at this illumination gradual change pixel place;
Described confirmation module, also for determining that illumination gradual change pixel is lost motion object pixel with the illumination gradual change pixel that the similarity of corresponding compared pixels is greater than the 5th threshold value.
16. devices as claimed in claim 15, is characterized in that;
Described determination module, does not also occur in the moving target pixel to be detected of current frame image that texture information changes and is not whether each pixel of illumination gradual change pixel is the fast transshaping element of illumination for judging; And for a fast transshaping element of illumination, determine the similarity of the fast transshaping of this illumination element and corresponding compared pixels, wherein compared pixels corresponding to the fast transshaping element of illumination is to be arranged in background image and the pixel identical with the position of the current frame image at the plain place of the fast transshaping of this illumination;
Described confirmation module, also for determining that the fast transshaping element of illumination that the fast transshaping element of illumination and the similarity of corresponding compared pixels are greater than the 4th threshold value is lost motion object pixel.
17. devices as described in claim 13 or 16, is characterized in that;
Described determination module, also for for a pixel, from the respective pixel of this pixel, select two respective pixel, the image at two respective pixel places of wherein selecting is two two field pictures of arbitrary neighborhood, and the respective pixel of this pixel is to be arranged in any two field picture of continuous N two field picture before current frame image and the pixel identical with the position of the current frame image at this pixel place; If the absolute value of the difference of the gray-scale value of two respective pixel of described selection is all greater than the 6th threshold value, determine that this pixel is the fast transshaping element of illumination; Otherwise, determine that this pixel is not the fast transshaping element of illumination.
18. devices as described in claims 14 or 15, is characterized in that;
Described determination module, also for for a pixel, if the absolute value of the difference of the average gray of the pixel of the neighborhood of the compared pixels that the average gray of the pixel of the neighborhood of this pixel in current frame image is corresponding with this pixel is greater than the 7th threshold value, and the absolute value of the difference of the Grad of the gray scale of the compared pixels that the Grad of the gray scale of this pixel is corresponding with this pixel is less than the 8th threshold value, determine that this pixel is illumination gradual change pixel, wherein compared pixels corresponding to this pixel is to be arranged in background image, and the pixel identical with the position of the current frame image at this pixel place, the Grad of pixel is the Grad of horizontal direction and the Grad sum of vertical direction of the gray scale of this pixel, otherwise, determine that this pixel is not illumination gradual change pixel.
19. devices as described in claim 14 or 16, is characterized in that;
Described confirmation module, also, for there is the number of the pixel of illumination gradual change in the moving target pixel to be detected in number and the current frame image of the fast pixel becoming of illumination in the moving target pixel to be detected in number, the current frame image of pixel that texture information changes according to current frame image, determine the illumination variation state of current scene; And according to the corresponding relation of illumination variation state and background image renewal speed, for current scene is set corresponding background image renewal speed.
20. devices as claimed in claim 19, is characterized in that, described confirmation module is determined the illumination variation state of current scene, comprising:
If the number of the fast transshaping element of illumination is greater than the 9th threshold value in the moving target pixel to be detected in current frame image, the illumination variation state of current scene is illumination fast state;
If the number that in the moving target pixel to be detected in current frame image, the number of the fast transshaping element of illumination is not more than illumination gradual change pixel in the moving target pixel to be detected in the 9th threshold value and current frame image is greater than the number of the pixel that in the tenth threshold value and current frame image, texture information changes and is greater than the 11 threshold value, the illumination variation state of current scene is illuminance abrupt variation state;
If the number that in the moving target pixel to be detected in current frame image, the number of the fast transshaping element of illumination is not more than illumination gradual change pixel in the moving target pixel to be detected in the 9th threshold value and current frame image is greater than the number of the pixel that in the 12 threshold value and current frame image, texture information changes and is less than the 13 threshold value, the illumination variation state of current scene is illumination gradual change state;
Otherwise, the illumination steady state (SS) of current scene.
CN201210466476.5A 2012-11-16 2012-11-16 A kind of method and apparatus determining lost motion object pixel Active CN103824307B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201210466476.5A CN103824307B (en) 2012-11-16 2012-11-16 A kind of method and apparatus determining lost motion object pixel

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201210466476.5A CN103824307B (en) 2012-11-16 2012-11-16 A kind of method and apparatus determining lost motion object pixel

Publications (2)

Publication Number Publication Date
CN103824307A true CN103824307A (en) 2014-05-28
CN103824307B CN103824307B (en) 2016-04-06

Family

ID=50759348

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201210466476.5A Active CN103824307B (en) 2012-11-16 2012-11-16 A kind of method and apparatus determining lost motion object pixel

Country Status (1)

Country Link
CN (1) CN103824307B (en)

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105335951A (en) * 2014-06-06 2016-02-17 株式会社理光 Background model updating method and device
CN107331118A (en) * 2017-07-05 2017-11-07 浙江宇视科技有限公司 Fall detection method and device
CN107408305A (en) * 2015-03-27 2017-11-28 三菱电机株式会社 Camera device and method, operation device and method, program and recording medium
CN110490025A (en) * 2018-05-14 2019-11-22 杭州海康威视数字技术股份有限公司 A kind of object detection method, device, equipment and system
CN110705492A (en) * 2019-10-10 2020-01-17 北京北特圣迪科技发展有限公司 Stage mobile robot obstacle target detection method
CN111107361A (en) * 2018-10-29 2020-05-05 安讯士有限公司 Video processing apparatus and method for determining motion metadata of encoded video
CN118135065A (en) * 2024-05-07 2024-06-04 山东汉鑫科技股份有限公司 Tunnel dynamic gray scale map generation method, system, storage medium and electronic equipment

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20060133785A1 (en) * 2004-12-21 2006-06-22 Byoung-Chul Ko Apparatus and method for distinguishing between camera movement and object movement and extracting object in a video surveillance system
CN101916448A (en) * 2010-08-09 2010-12-15 云南清眸科技有限公司 Moving object detecting method based on Bayesian frame and LBP (Local Binary Pattern)
CN102147861A (en) * 2011-05-17 2011-08-10 北京邮电大学 Moving target detection method for carrying out Bayes judgment based on color-texture dual characteristic vectors

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20060133785A1 (en) * 2004-12-21 2006-06-22 Byoung-Chul Ko Apparatus and method for distinguishing between camera movement and object movement and extracting object in a video surveillance system
CN101916448A (en) * 2010-08-09 2010-12-15 云南清眸科技有限公司 Moving object detecting method based on Bayesian frame and LBP (Local Binary Pattern)
CN102147861A (en) * 2011-05-17 2011-08-10 北京邮电大学 Moving target detection method for carrying out Bayes judgment based on color-texture dual characteristic vectors

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105335951A (en) * 2014-06-06 2016-02-17 株式会社理光 Background model updating method and device
CN105335951B (en) * 2014-06-06 2018-06-15 株式会社理光 Background model update method and equipment
CN107408305A (en) * 2015-03-27 2017-11-28 三菱电机株式会社 Camera device and method, operation device and method, program and recording medium
CN107408305B (en) * 2015-03-27 2020-09-11 三菱电机株式会社 Image pickup apparatus and method, operation apparatus, and recording medium
CN107331118A (en) * 2017-07-05 2017-11-07 浙江宇视科技有限公司 Fall detection method and device
CN110490025A (en) * 2018-05-14 2019-11-22 杭州海康威视数字技术股份有限公司 A kind of object detection method, device, equipment and system
CN111107361A (en) * 2018-10-29 2020-05-05 安讯士有限公司 Video processing apparatus and method for determining motion metadata of encoded video
CN111107361B (en) * 2018-10-29 2021-09-28 安讯士有限公司 Video processing apparatus, method and medium for determining motion metadata of encoded video
CN110705492A (en) * 2019-10-10 2020-01-17 北京北特圣迪科技发展有限公司 Stage mobile robot obstacle target detection method
CN118135065A (en) * 2024-05-07 2024-06-04 山东汉鑫科技股份有限公司 Tunnel dynamic gray scale map generation method, system, storage medium and electronic equipment

Also Published As

Publication number Publication date
CN103824307B (en) 2016-04-06

Similar Documents

Publication Publication Date Title
CN103824307B (en) A kind of method and apparatus determining lost motion object pixel
US11693901B2 (en) Systems and methods for geolocation prediction
JP6893564B2 (en) Target identification methods, devices, storage media and electronics
WO2020098708A1 (en) Lane line detection method and apparatus, driving control method and apparatus, and electronic device
US11068713B1 (en) Video-based intelligent road traffic universal analysis
EP3182331A1 (en) Method and system to detect objects using block based histogram of oriented gradients
CN113468967A (en) Lane line detection method, device, equipment and medium based on attention mechanism
US20200065664A1 (en) System and method of measuring the robustness of a deep neural network
CN110222726A (en) Image processing method, device and electronic equipment
CN111222509B (en) Target detection method and device and electronic equipment
CN112329762A (en) Image processing method, model training method, device, computer device and medium
CN111753729B (en) False face detection method and device, electronic equipment and storage medium
Kocur et al. Traffic camera calibration via vehicle vanishing point detection
Yang et al. YOLOv8-Lite: A Lightweight Object Detection Model for Real-time Autonomous Driving Systems
CN109242882B (en) Visual tracking method, device, medium and equipment
CN108520532B (en) Method and device for identifying motion direction of object in video
Rishika et al. Real-time vehicle detection and tracking using YOLO-based deep sort model: A computer vision application for traffic surveillance
US20230177666A1 (en) Degradation detection device, degradation detection system, degradation detection method, and program
CN113807354A (en) Image semantic segmentation method, device, equipment and storage medium
CN111950517A (en) Target detection method, model training method, electronic device and storage medium
CN116188954A (en) Online evaluation method and system for target detection algorithm, electronic equipment and storage medium
CN111210513A (en) Information processing method, device, equipment and storage medium
CN115984640A (en) Target detection method, system and storage medium based on combined distillation technology
CN113160217B (en) Method, device, equipment and storage medium for detecting circuit foreign matters
KR20220155882A (en) Data processing method and apparatus using a neural network

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant