CN105427342A - Method and system for detecting and tracking underwater small-target sonar image target - Google Patents
Method and system for detecting and tracking underwater small-target sonar image target Download PDFInfo
- Publication number
- CN105427342A CN105427342A CN201510789128.5A CN201510789128A CN105427342A CN 105427342 A CN105427342 A CN 105427342A CN 201510789128 A CN201510789128 A CN 201510789128A CN 105427342 A CN105427342 A CN 105427342A
- Authority
- CN
- China
- Prior art keywords
- target
- image
- template
- current
- area
- 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
Links
Landscapes
- Measurement Of Velocity Or Position Using Acoustic Or Ultrasonic Waves (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a method for detecting and tracking an underwater small-target sonar image target and a system corresponding to the method. The method comprises the following steps: obtaining a normalized template with the target as a central point; obtaining a current image containing the target; calculating and judging whether the image containing the target is similar to the template according to a similarity coefficient or a correlation coefficient, and if so, judging that the central point of the current image containing the target is a measurement point of a current target track; obtaining a starting track, obtaining a prediction point of the current target track by Kalman filtering, carrying out data correlation on the measurement point and the prediction point, and determining a target track point of the current image; and outputting target point track information according to the starting track and the target track point of the current image. According to the method and the system for detecting and tracking the underwater small-target sonar image target disclosed by the invention, the method combining template matching with the similarity coefficient is used for greatly saving the operation time and meeting the real-time requirements of the detection system.
Description
Technical field
The invention of this reality relates to image procossing and Underwater Target Detection field, particularly relates to one Small object sonar image target detection tracking method and system under water.
Background technology
Current China is also in the quite weak stage, especially for the defence of Small object, as frogman, frogman's vehicle and small-sized AUV etc. for the underwater defence of the targets such as harbour, seashore and naval vessel.Sonar is in imaging process, and due to bottom reverberation, the reason such as the shoal of fish and reef, can produce a large amount of Noise and Interference sources in the picture, and the existence in Noise and Interference source makes natively faint Small object under water more be difficult to detect.
The object of sonar image target detection will extract target area exactly from Reverberation at the bottom of Complex Sea, and it is the committed step of graphical analysis.Only on the basis that sonar image target accurately detects, feature extraction and parameter measurement could be carried out to submarine target, make higher level sonar image analysis and identification become possibility.But due to the complicacy of sound field environment and the non-linear of sonar set imaging under water, the sonar image under water collected has that contrast is low, image quality is poor, by features such as noise pollution are serious.Because frogman waits the target of Small object under water little, signal to noise ratio (S/N ratio) is very low, is often submerged in ground unrest, and object brightness changes greatly in different distance, most filtering method and detection method effect all not fully up to expectations.
Traditional image detecting method based on marginal information or statistical information is difficult to the object detection results obtaining high precision, strong robustness.Adopt template matches under low signal-to-noise ratio, well can detect target, but because matched filter needs to carry out associative operation, and associative operation belongs to exhaustive search method, the time of at substantial is used for computing, cause the overlong time of sonar image target detection, the requirement of system real time cannot be met.
Summary of the invention
The object of this invention is to provide one Small object sonar image target detection tracking method and system under water, to solve the problem.
This is Small object sonar image target detection tracking method under water, comprising:
Obtain the normalized template of goal-orientation point;
Obtain the current image comprising target;
Calculate and currently comprise the image of target and the likeness coefficient of template or related coefficient;
Judge that whether the described current image comprising target is similar to template according to described likeness coefficient or related coefficient,
If the described current image comprising target is similar to template, then judge the described current measurement point of image center as current goal track comprising target, and continue to calculate other and comprise the image of target and the likeness coefficient of template or related coefficient,
If described current comprise target image and template dissimilar, then directly calculate other and comprise the image of target and the likeness coefficient of template or related coefficient;
Obtain initial flight path;
Kalman filtering is utilized to obtain the future position of current goal track;
Determine the target trajectory point of present image;
Target trajectory point according to described initial flight path and present image exports impact point trace information.
The invention also discloses one Small object sonar image target detection tracker under water, comprising:
Template acquisition module, described template acquisition module is for obtaining the normalized template of goal-orientation point;
Image segmentation module, described image segmentation module is for obtaining the current image comprising target;
Similarity determination module, described similarity determination module is connected with image segmentation module with described template acquisition module respectively, currently the image of target and the likeness coefficient of template or related coefficient is comprised for calculating, and judge that whether the described current image comprising target is similar to template according to described likeness coefficient or related coefficient
If the described current image comprising target is similar to template, then judge the described current measurement point of image center as current goal track comprising target, and continue to calculate other and comprise the image of target and the likeness coefficient of template or related coefficient,
If described current comprise target image and template dissimilar, then directly calculate other and comprise the image of target and the likeness coefficient of template or related coefficient;
Initial flight path acquisition module, described initial flight path acquisition module is for obtaining initial flight path;
Future position acquisition module, the future position of described future position acquisition module for utilizing Kalman filtering to obtain current goal track;
Target trajectory point acquisition module, described target trajectory point acquisition module is connected with future position acquisition module with described similarity determination module respectively, for determining the target trajectory point of present image;
Output module, described output module is connected with initial flight path acquisition module with described target trajectory point acquisition module respectively, exports impact point trace information for the target trajectory point according to described initial flight path and present image.
The sonar image of Small object under water target detection tracking method disclosed in this invention, first carries out statistical study to the target in sonar image, obtains a normalized template put centered by described target.Then multiple current image comprising target is obtained, recycle each current image (may be order target area) comprising target in described template and image to ask for and currently comprise the image of target and the likeness coefficient (or related coefficient) of template, according to the size detection of likeness coefficient (or related coefficient), whether the current pixel comprised in the image of target is possible impact point, and will may be that the pixel of impact point is as measurement point in the current image comprising target.Future position and the described measurement point of the current goal track obtained utilizing Kalman filtering carry out data correlation, simultaneously using the weighting coefficient of described likeness coefficient (or related coefficient) as distance metric function during data correlation, thus determine the target trajectory point of present image.Finally obtain the movement locus of target according to the initial flight path of employing 2 extrapolation methods acquisitions and the target trajectory point of present image, and export.The method that the sonar image of Small object under water target detection tracking method disclosed in this invention adopts template matches and likeness coefficient to combine, greatly can save operation time, meet the requirement of real-time of system.
Accompanying drawing explanation
In order to be illustrated more clearly in the embodiment of the present invention or technical scheme of the prior art, be briefly described to the accompanying drawing used required in the embodiment of the present invention or description of the prior art below, apparently, accompanying drawing in the following describes is only some embodiments of the present invention, for those skilled in the art, under the prerequisite not paying creative work, other accompanying drawing can also be obtained according to these accompanying drawings.
Fig. 1 is the sonar image of the Small object under water target detection tracking method process flow diagram disclosed in the embodiment of the present invention;
Fig. 2 is the sonar image of the Small object under water target detection tracker schematic diagram disclosed in the embodiment of the present invention.
Embodiment
For making the object of the embodiment of the present invention, technical scheme and advantage clearly, below in conjunction with the accompanying drawing in the embodiment of the present invention, technical scheme in the embodiment of the present invention is clearly and completely described, obviously, described embodiment is the present invention's part embodiment, instead of whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art, not making the every other embodiment obtained under creative work prerequisite, belong to the scope of protection of the invention.
This clearly demarcated embodiment discloses one Small object sonar image target detection tracking method under water, and as shown in Figure 1, the method comprises:
Obtain the normalized template put centered by described target.
Obtain the current image comprising target.
Calculate and currently comprise the image of target and the likeness coefficient of template or related coefficient.
Judge that whether the described current image comprising target is similar to template according to described likeness coefficient or related coefficient, if the described current image comprising target is similar to template, then judge the described current measurement point of image center as current goal track comprising target, and continue to calculate other and comprise the image of target and the likeness coefficient of template or related coefficient, if described current comprise target image and template dissimilar, then directly calculate other and comprise the image of target and the likeness coefficient of template or related coefficient.
Obtain initial flight path.
Kalman filtering is utilized to obtain the future position of current goal track.
Measurement point and future position are carried out data correlation, determines the target trajectory point of present image.
Target trajectory point according to described initial flight path and present image exports impact point trace information.
The sonar image of Small object under water target detection tracking method disclosed in this invention, first carries out statistical study to the target in sonar image, obtains a normalized template put centered by described target.Then multiple current image comprising target is obtained, recycle each current image (may be order target area) comprising target in described template and image to ask for and currently comprise the image of target and the likeness coefficient (or related coefficient) of template, according to the size detection of likeness coefficient (or related coefficient), whether the current pixel comprised in the image of target is possible impact point, and will may be that the pixel of impact point is as measurement point in the current image comprising target.Future position and the described measurement point of the current goal track obtained utilizing Kalman filtering carry out data correlation, simultaneously using the weighting coefficient of described likeness coefficient (or related coefficient) as distance metric function during data correlation, thus determine the target trajectory point of present image.Finally obtain the movement locus of target according to the initial flight path of employing 2 extrapolation methods acquisitions and the target trajectory point of present image, and export.The method that the sonar image of Small object under water target detection tracking method disclosed in this invention adopts template matches and likeness coefficient to combine, greatly can save operation time, meet the requirement of real-time of system.
Wherein, obtaining the normalized template put centered by described target is that detailed process comprises by utilizing object brightness information acquisition in different distance:
Constructing a size is the rectangular window of n × n, and sonar image is divided into multiple region by described rectangular window, is convenient to follow-up template and sets up.
Utilize the order target area A={area in different distance in described window extraction sonar image
1, area
2..., area
t, wherein, t is the number of target area.
To target area A={area
1, area
2..., area
tin arbitrary pixel area
i(θ, d) normalizes to [0,1], and normalized result is:
Area '
i(θ, d)=area
i(θ, d)/[max (area
i)-min (area
i)], wherein, i ∈ [1, t], max (area
i) represent and ask region area
imax pixel value, min (area
i) represent and ask region area
iminimum pixel value, θ is the angle of target relative to sonar, and as the horizontal ordinate of sonar image, d is the distance between target and sonar, as the ordinate of sonar image.
Target area A ' after further acquisition normalization=area '
1, area '
2..., area '
t.
Constructing a size is the template T of n × n, and by the target area A ' after normalization=area '
1, area '
2..., area '
tbe located at the center of template T, determine the pixel value of template T:
Described template T is as the reference data of current sonar image similarity.
The present embodiment utilizes the method asking for local mean value to obtain segmentation threshold, splits, obtain the current image comprising target to current sonar image, avoids and adopts global threshold to cause remote target to be submerged in background.The current process comprising the image of target of described acquisition, comprising:
By wide be w, high be divided into m interval according to the distance from image base (position at corresponding sonar place) for the pixel in the current sonar image f (θ, d) of h, then between each distance regions be:
wherein, k=0,1,2..., m;
To calculate in current sonar image f (θ, d) each pixel to the distance of image base:
dis(θ,d)=d;
To divide residing for described pixel between distance regions according to pixel each in current sonar image to the distance of image base;
Calculate the average μ and variances sigma that obtain pixel in each distance regions
2;
Then segmentation threshold th is:
Th=μ+p σ
2, wherein, p is scale-up factor p=dis (θ, d)/max{dis (θ, d) }, p ∈ (0,1);
Utilize described segmentation threshold th to split between the distance regions corresponding to described segmentation threshold th, obtain the current image f ' (θ, d) comprising target.
Owing to all having a segmentation threshold between each distance regions between m distance regions, the current image f ' (θ, d) of target that comprises therefore obtained is for m.
Calculate the current process comprising the image of target and the likeness coefficient of template or related coefficient, comprising:
Binary conversion treatment is carried out to the described current image comprising target, and the current central point comprising each non-zero region in the image of target after obtaining binary conversion treatment.And this process comprises:
Utilize described segmentation threshold th to carry out binary conversion treatment to the described current image f ' (θ, d) comprising target, obtain the image f ' after binaryzation
b(θ, d):
Image f ' after mark binaryzation
bnon-zero region B in (θ, d), then have m non-zero region, i.e. B={b accordingly
1, b
2..., b
m, wherein, b
ibe the image f ' after i-th binaryzation
bnon-zero region in (θ, d), 1≤i≤m.
Calculate the image f ' after binary conversion treatment
bthe central point c of non-zero region in (θ, d)
i(θ, d):
wherein, M is the image f ' after binary conversion treatment
bthe number of pixels of i-th non-zero region in (θ, d), j is integer, and 1≤j≤M, central point c
i(θ, d) is the image f ' after a jth binaryzation
bthe central point of the non-zero region in (θ, d).
The central point pixel value of described non-zero region is assigned to the central point of described template, and is described template assignment in proportion.Assignment is mapped the position of the value of template with place, region, because different regions is in different distances, due to the attenuation of signal, even the pixel value of same target in different distance is also different, the image f ' therefore after obtaining binary conversion treatment
b(θ, d) in non-zero region central point after, the central point pixel value of described non-zero region is assigned to the central point of described template, and be described template assignment in proportion, template is made to have the brightness similar to the region in this distance like this in different distances, to carry out subsequent calculations.Because the pixel value of pixel each in template is between [0,1], wherein the pixel value of central point is 1, then the pixel value of other pixels can passing ratio relation obtain.
Calculate according to the assigned result of described template and currently comprise the image of target and the likeness coefficient of template or related coefficient.This process comprises:
Calculate and currently comprise the image of target and the likeness coefficient R of template
fT:
Or, calculate and currently comprise the image of target and the related coefficient γ (θ, d) of template:
Judge to comprise the process whether described present image is similar to template according to described likeness coefficient or related coefficient:
Relatively likeness coefficient R
fTwith the similar threshold value th0 preset, if R
fT>th0, then judge that the current image comprising target is similar to template, now, then judge the described current measurement point of image center as current goal track comprising target; If R
fT≤ th0, then judge the current image and the template dissmilarity that comprise target, giving up this current image comprising target need not.
Or, described related coefficient γ (θ, d) is taken absolute value | γ (θ, d) |;
The relatively absolute value of related coefficient | γ (θ, d) | with the similar threshold value th0 preset, if | γ (θ, d) | >th0, then judge that the current image comprising target is similar to template, now, then the described current measurement point of image center as current goal track comprising target is judged; If | γ (θ, d) |≤th0, then judge the current image and the template dissmilarity that comprise target, giving up this current image comprising target need not.
In the present invention, the target point that pixel value is maximum in the described current image comprising target in theory, but when described current comprise that in the image of target, target area is very little time, get regional center point identical with area pixel value maximum point effect, and the point that multiple pixel value may be had in this region maximum, so it is more reasonable as the measurement point of current goal track to get regional center point in the present embodiment.
Each flight path has a beginning, and initial flight path is exactly this beginning.Described initial flight path adopts employing 2 extrapolation methods to obtain, and described employing 2 extrapolation methods obtain initial flight path be exactly with the front two-point measurement value extrapolation of successful association thirdly, and the measurement point that thirdly can obtain as predicted value and the 3rd two field picture associates with second point, if do not obtain effective measurement point, directly using the calculating thirdly carrying out the 4th two field picture as tracing point obtained of extrapolating.Predicted value afterwards all obtains by Kalman filtering.
In the present embodiment, the process of the initial flight path of described acquisition, comprising:
According to twice surving coordinate point (θ before current i-th target
1, d
1) and (θ
2, d
2), adopt the coordinate of 2 current i-th targets of extrapolated forecasting method:
Described measurement point and future position are carried out data correlation, determine the process of the target trajectory point of present image, comprising:
Calculate the Euclidean distance Dis between described measurement point and future position:
In association door, using measured value minimum for Euclidean distance Dis as the target trajectory point in present image.
Or,
In association door, using measured value minimum for Euclidean distance Dis as the target trajectory point in present image.
It should be noted that, in a two field picture, comprise multiple suspicious target area (i.e. the current image f ' (θ, d) comprising target), the corresponding likeness coefficient (or related coefficient) in each region and Euclidean distance.Such as, have an objective track, obtain multiple suspicious region according to present image, so these suspicious regions all calculate corresponding likeness coefficient (or related coefficient) and Euclidean distance to target trajectory.Again calculate Euclidean distance according to likeness coefficient (or related coefficient), and choose minimum current of corresponding Euclidean distance and comprise the track of the impact point in the image f ' (θ, d) of target as current goal.
Described association door determines according to the kinetic characteristic of target, and within the limited time, the displacement that the target of certain particular type has it maximum, this maximum displacement associates door exactly.There is no the necessity calculated outside association door, operand can be reduced greatly in described associations.
Finally export impact point trace information, to analyze further according to the target trajectory point of described initial flight path and present image.
The sonar image of Small object under water target detection tracking method disclosed in this invention adopts the method for template matches to carry out target detection to sonar image, well target can be detected under low signal-to-noise ratio.And the method adopting template matches and likeness coefficient to combine further, greatly can save operation time, meet the requirement of real-time of detection system.
Another embodiment of the present invention also discloses one Small object sonar image target detection tracker under water, as shown in Figure 2, comprising:
Template acquisition module, described template acquisition module is for obtaining the normalized template of goal-orientation point;
Image segmentation module, described image segmentation module is for obtaining the current image comprising target;
Similarity determination module, described similarity determination module is connected with image segmentation module with described template acquisition module respectively, currently the image of target and the likeness coefficient of template or related coefficient is comprised for calculating, and judge that whether the described current image comprising target is similar to template according to described likeness coefficient or related coefficient
If the described current image comprising target is similar to template, then judge the described current measurement point of image center as current goal track comprising target, and continue to calculate other and comprise the image of target and the likeness coefficient of template or related coefficient,
If described current comprise target image and template dissimilar, then directly calculate other and comprise the image of target and the likeness coefficient of template or related coefficient;
Initial flight path acquisition module, described initial flight path acquisition module is for obtaining initial flight path;
Future position acquisition module, the future position of described future position acquisition module for utilizing Kalman filtering to obtain current goal track;
Target trajectory point acquisition module, described target trajectory point acquisition module is connected with future position acquisition module with described similarity determination module respectively, for measurement point and future position are carried out data correlation, determines the target trajectory point of present image;
Output module, described output module is connected with initial flight path acquisition module with described target trajectory point acquisition module respectively, and the target trajectory point according to described initial flight path and present image exports impact point trace information.
Wherein, described template acquisition module, comprising:
Window construction unit, described window construction unit is the rectangular window of n × n for constructing a size, and sonar image is divided into multiple region by described rectangular window, is convenient to follow-up template and sets up.
Image extraction unit, described image extraction unit is connected with described window construction unit, for utilizing the order target area A={area in described window extraction sonar image in different distance
1, area
2..., area
t, wherein, t is the number of target area.
Normalization unit, described normalization unit is connected with described image extraction unit, for target area A={area
1, area
2..., area
tin arbitrary pixel area
i(θ, d) normalizes to [0,1], and normalized result is:
Area '
i(θ, d)=area
i(θ, d)/[max (area
i)-min (area
i)], wherein, i ∈ [1, t], max (area
i) represent and ask region area
imax pixel value, min (area
i) represent and ask region area
iminimum pixel value, θ is the angle of target relative to sonar, and as the horizontal ordinate of sonar image, d is the distance between target and sonar, as the ordinate of sonar image.And obtain further the target area A ' after normalization=area '
1, area '
2..., area '
t.
Template construction unit, described template construction unit is connected with described normalization unit, is the template T of n × n for constructing a size, and by the target area A ' after normalization=area '
1, area '
2..., area '
tbe located at the center of template T, determine the pixel value of template T:
Described template T is as the reference data of current sonar image similarity.
Described image segmentation module, comprising:
Interal separation unit, described interal separation unit be used for by wide be w, high be the current sonar image f (θ of h, d) pixel in is divided into m interval according to the distance from image base (position at corresponding sonar place), then between each distance regions be:
wherein, k=0,1,2..., m.
Pixel concludes unit, and described pixel is concluded unit and is connected with described interal separation unit, for calculating in current sonar image f (θ, d) each pixel to the distance of image base:
dis(θ,d)=d;
And to divide residing for described pixel between distance regions according to pixel each in current sonar image to the distance of image base.
Threshold computation unit, described threshold computation unit and described pixel are concluded unit and are connected, for calculating the average μ and variances sigma that obtain pixel in each distance regions
2;
Then segmentation threshold th is:
Th=μ+p σ
2, wherein, p is scale-up factor, p=dis (θ, d)/max{dis (θ, d) }, p ∈ (0,1).
Segmentation image acquisition unit, described segmentation image acquisition unit is connected with described threshold computation unit, for utilizing described segmentation threshold th to split between the distance regions corresponding to described segmentation threshold th, obtain the current image f ' (θ, d) comprising target.Owing to all having a segmentation threshold between each distance regions between m distance regions, the current image f ' (θ, d) of target that comprises therefore obtained is for m.
Described similarity determination module, comprising:
Binarization unit, described binarization unit is used for carrying out binary conversion treatment to the described current image comprising target, and the current central point comprising each non-zero region in the image of target after obtaining binary conversion treatment.This process comprises:
Utilize described segmentation threshold th to carry out binary conversion treatment to the described current image f ' (θ, d) comprising target, obtain the image f ' after binaryzation
b(θ, d):
Image f ' after mark binaryzation
bnon-zero region B in (θ, d), then have m non-zero region, i.e. B={b accordingly
1, b
2..., b
m, wherein, b
ibe the image f ' after i-th binaryzation
bnon-zero region in (θ, d), 1≤i≤m;
Calculate the image f ' after binary conversion treatment
bthe central point c of non-zero region in (θ, d)
i(θ, d):
wherein, M is the image f ' after binary conversion treatment
bthe number of pixels of i-th non-zero region in (θ, d), j is integer, and 1≤j≤M, central point c
i(θ, d) is the image f ' after a jth binaryzation
bthe central point of the non-zero region in (θ, d).
Template assignment unit, described template assignment unit is connected with described binarization unit, for the central point pixel value of described non-zero region being assigned to the central point of described template, and is described template assignment in proportion.Assignment is mapped the position of the value of template with place, region, because different regions is in different distances, due to the attenuation of signal, even the pixel value of same target in different distance is also different, the image f ' therefore after obtaining binary conversion treatment
b(θ, d) in non-zero region central point after, the central point pixel value of described non-zero region is assigned to the central point of described template, and be described template assignment in proportion, template is made to have the brightness similar to the region in this distance like this in different distances, to carry out subsequent calculations.Because the pixel value of pixel each in template is between [0,1], wherein the pixel value of central point is 1, then the pixel value of other pixels can passing ratio relation obtain.
Coefficient calculation unit, described coefficient calculation unit is connected with described template assignment unit, currently comprises the image of target and the likeness coefficient of template or related coefficient for calculating according to the assigned result of described template.This process comprises:
Calculate and currently comprise the image of target and the likeness coefficient R of template
fT:
Or, calculate and currently comprise the image of target and the related coefficient γ (θ, d) of template:
Similarity judging unit, described similarity judging unit is connected with coefficient calculation unit, and for judging that whether described present image is similar to template according to described likeness coefficient or related coefficient, this process comprises:
Relatively likeness coefficient R
fTwith the similar threshold value th0 preset, if R
fT>th0, then judge that the current image comprising target is similar to template, now, then judge the described current measurement point of image center as current goal track comprising target; If R
fT≤ th0, then judge the current image and the template dissmilarity that comprise target, giving up this current image comprising target need not.
Or, described related coefficient γ (θ, d) is taken absolute value | γ (θ, d) |;
The relatively absolute value of related coefficient | γ (θ, d) | with the similar threshold value th0 preset, if | γ (θ, d) | >th0, then judge that the current image comprising target is similar to template, now, then the described current measurement point of image center as current goal track comprising target is judged; If | γ (θ, d) |≤th0, then judge the current image and the template dissmilarity that comprise target, giving up this current image comprising target need not.
Described initial flight path acquisition module obtains the process of initial flight path, comprising:
According to twice surving coordinate point (θ before current i-th target
1, d
1) and (θ
2, d
2), adopt the coordinate of 2 current i-th targets of extrapolated forecasting method:
Measurement point and future position are carried out data correlation by described target trajectory point acquisition module, determine the process of the target trajectory point of present image, comprising:
Calculate the Euclidean distance Dis between described measurement point and future position:
In association door, using measured value minimum for Euclidean distance Dis as the target trajectory point in present image.
Or,
In association door, using measured value minimum for Euclidean distance Dis as the target trajectory point in present image.
It should be noted that, in a two field picture, comprise multiple suspicious target area (i.e. the current image f ' (θ, d) comprising target), the corresponding likeness coefficient (or related coefficient) in each region and Euclidean distance.Such as, have an objective track, obtain multiple suspicious region according to present image, so these suspicious regions all calculate corresponding likeness coefficient (or related coefficient) and Euclidean distance to target trajectory.Again calculate Euclidean distance according to likeness coefficient (or related coefficient), and choose minimum current of corresponding Euclidean distance and comprise the track of the impact point in the image f ' (θ, d) of target as current goal.
Described association door determines according to the kinetic characteristic of target, and within the limited time, the displacement that the target of certain particular type has it maximum, this maximum displacement associates door exactly.There is no the necessity calculated outside association door, operand can be reduced greatly in described associations.
Output module finally exports impact point trace information, to analyze further according to the target trajectory point of described initial flight path and present image.
The sonar image of Small object under water target detection tracker disclosed in this invention adopts the method for template matches to carry out target detection to sonar image, well target can be detected under low signal-to-noise ratio.And the method adopting template matches and likeness coefficient to combine further, greatly can save operation time, meet the requirement of real-time of detection system.
Above the sonar image of Small object under water target detection tracking method provided by the present invention is described in detail.Apply specific case herein to set forth principle of the present invention and embodiment, the explanation of above embodiment just understands method of the present invention and core concept thereof for helping.It should be pointed out that for those skilled in the art, under the premise without departing from the principles of the invention, can also carry out some improvement and modification to the present invention, these improve and modify and also fall in the protection domain of the claims in the present invention.
Claims (10)
1. a Small object sonar image target detection tracking method under water, is characterized in that, comprising:
Obtain the normalized template of goal-orientation point;
Obtain the current image comprising target;
Calculate and currently comprise the image of target and the likeness coefficient of template or related coefficient;
Judge that whether the described current image comprising target is similar to template according to described likeness coefficient or related coefficient,
If the described current image comprising target is similar to template, then judge the described current measurement point of image center as current goal track comprising target, and continue to calculate other and comprise the image of target and the likeness coefficient of template or related coefficient,
If described current comprise target image and template dissimilar, then directly calculate other and comprise the image of target and the likeness coefficient of template or related coefficient;
Obtain initial flight path;
Kalman filtering is utilized to obtain the future position of current goal track;
Measurement point and future position are carried out data correlation, determines the target trajectory point of present image;
Target trajectory point according to described initial flight path and present image exports impact point trace information.
2. Small object sonar image target detection tracking method under water according to claim 1, is characterized in that, obtain the process of the normalized template of goal-orientation point, comprising:
Constructing a size is the rectangular window of n × n;
Utilize the order target area A={area in different distance in described window extraction sonar image
1, area
2..., area
t, wherein, t is the number of target area;
To target area A={area
1, area
2..., area
tin arbitrary pixel area
i(θ, d) normalizes to [0,1], and normalized result is:
Area '
i(θ, d)=area
i(θ, d)/[max (area
i)-min (area
i)], wherein, i ∈ [1, t], max (area
i) represent and ask region area
imax pixel value, min (area
i) represent and ask region area
iminimum pixel value, θ is the angle of target relative to sonar, and d is the distance between target and sonar;
Target area A ' after acquisition normalization=area '
1, area '
2..., area '
t;
Constructing a size is the template T of n × n, and by the target area A ' after normalization=area '
1, area '
2..., area '
tbe located at the center of template T, determine the pixel value of template T:
3. Small object sonar image target detection tracking method under water according to claim 2, is characterized in that, obtain the current process comprising the image of target, comprising:
Pixel in the wide w of being, the high current sonar image being h is divided into m interval according to the distance from image base, then between each distance regions is:
wherein, k=0,1,2..., m;
To calculate in current sonar image each pixel to the distance of image base:
dis(θ,d)=d;
To divide residing for described pixel between distance regions according to pixel each in current sonar image to the distance of image base;
Calculate the average μ and variances sigma that obtain pixel in each distance regions
2;
Then segmentation threshold th is:
Th=μ+p σ
2, wherein, p is scale-up factor, p ∈ (0,1);
Utilize described segmentation threshold th to split between the distance regions corresponding to described segmentation threshold th, obtain the current image f ' (θ, d) comprising target.
4. Small object sonar image target detection tracking method under water according to claim 3, it is characterized in that, the current process comprising the image of target and the likeness coefficient of template or related coefficient of described calculating, comprising:
Binary conversion treatment is carried out to the described current image comprising target, and the current central point comprising each non-zero region in the image of target after obtaining binary conversion treatment;
The central point pixel value of described non-zero region is assigned to the central point of described template, and is described template assignment in proportion;
Calculate according to the assigned result of described template and currently comprise the image of target and the likeness coefficient of template or related coefficient.
5. Small object sonar image target detection tracking method under water according to claim 4, it is characterized in that, described binary conversion treatment is carried out to the current image comprising target, and the current process comprising the central point of each non-zero region in the image of target after obtaining binary conversion treatment, comprising:
Utilize described segmentation threshold th to carry out binary conversion treatment to the described current image f ' (θ, d) comprising target, obtain the image f ' after binaryzation
b(θ, d):
Calculate the image f ' after binary conversion treatment
bthe central point c of non-zero region in (θ, d)
i(θ, d):
wherein, M is the image f ' after binary conversion treatment
bthe number of pixels of i-th non-zero region in (θ, d), j is integer, and 1≤j≤M.
6. Small object sonar image target detection tracking method under water according to claim 5, it is characterized in that, the described assigned result according to described template calculates the current process comprising the image of target and the likeness coefficient of template or related coefficient, comprising:
Calculate and currently comprise the image of target and the likeness coefficient R of template
fT:
Or, calculate and currently comprise the image of target and the related coefficient γ (θ, d) of template:
7. Small object sonar image target detection tracking method under water according to claim 6, is characterized in that, describedly judges to comprise the process whether described present image is similar to template according to described likeness coefficient or related coefficient:
Relatively likeness coefficient R
fTwith the similar threshold value th0 preset, if R
fT>th0, then judge that the current image comprising target is similar to template;
Or, described related coefficient γ (θ, d) is taken absolute value | γ (θ, d) |;
The relatively absolute value of related coefficient | γ (θ, d) | with the similar threshold value th0 preset, if | γ (θ, d) | >th0, then judge that the current image comprising target is similar to template.
8. Small object sonar image target detection tracking method under water according to claim 7, is characterized in that, obtain the process of initial flight path, comprising:
According to twice surving coordinate point (θ before current i-th target
1, d
1) and (θ
2, d
2), adopt the coordinate of 2 current i-th targets of extrapolated forecasting method:
9. Small object sonar image target detection tracking method under water according to claim 8, is characterized in that, measurement point and future position are carried out data correlation, determine the process of the target trajectory point of present image, comprising:
Calculate the Euclidean distance Dis between described measurement point and future position:
In association door, using measured value minimum for Euclidean distance Dis as the target trajectory point in present image;
Or,
In association door, using measured value minimum for Euclidean distance Dis as the target trajectory point in present image.
10. a Small object sonar image target detection tracker under water, is characterized in that, comprising:
Template acquisition module, described template acquisition module is for obtaining the normalized template of goal-orientation point;
Image segmentation module, described image segmentation module is for obtaining the current image comprising target;
Similarity determination module, described similarity determination module is connected with image segmentation module with described template acquisition module respectively, currently the image of target and the likeness coefficient of template or related coefficient is comprised for calculating, and judge that whether the described current image comprising target is similar to template according to described likeness coefficient or related coefficient
If the described current image comprising target is similar to template, then judge the described current measurement point of image center as current goal track comprising target, and continue to calculate other and comprise the image of target and the likeness coefficient of template or related coefficient,
If described current comprise target image and template dissimilar, then directly calculate other and comprise the image of target and the likeness coefficient of template or related coefficient;
Initial flight path acquisition module, described initial flight path acquisition module is for obtaining initial flight path;
Future position acquisition module, the future position of described future position acquisition module for utilizing Kalman filtering to obtain current goal track;
Target trajectory point acquisition module, described target trajectory point acquisition module is connected with future position acquisition module with described similarity determination module respectively, for measurement point and future position are carried out data correlation, determines the target trajectory point of present image; ;
Output module, described output module is connected with initial flight path acquisition module with described target trajectory point acquisition module respectively, exports impact point trace information for the target trajectory point according to described initial flight path and present image.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510789128.5A CN105427342B (en) | 2015-11-17 | 2015-11-17 | A kind of underwater Small object sonar image target detection tracking method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510789128.5A CN105427342B (en) | 2015-11-17 | 2015-11-17 | A kind of underwater Small object sonar image target detection tracking method and system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN105427342A true CN105427342A (en) | 2016-03-23 |
CN105427342B CN105427342B (en) | 2018-06-08 |
Family
ID=55505522
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510789128.5A Active CN105427342B (en) | 2015-11-17 | 2015-11-17 | A kind of underwater Small object sonar image target detection tracking method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105427342B (en) |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106114095A (en) * | 2016-06-27 | 2016-11-16 | 中国石油大学(华东) | A kind of amphibious sniffing robot |
CN106338733A (en) * | 2016-09-09 | 2017-01-18 | 河海大学常州校区 | Forward-looking sonar object tracking method based on frog-eye visual characteristic |
CN106910204A (en) * | 2016-12-30 | 2017-06-30 | 中国人民解放军空军预警学院监控***工程研究所 | A kind of method and system to the automatic Tracking Recognition of sea ship |
CN107290731A (en) * | 2017-06-21 | 2017-10-24 | 哈尔滨工业大学 | A kind of radar track initial mode based on image area template matches |
CN107526087A (en) * | 2016-06-21 | 2017-12-29 | 北京臻迪科技股份有限公司 | A kind of method and system for obtaining underwater 3D faultage images |
US20180120472A1 (en) * | 2016-10-28 | 2018-05-03 | Korea Institute Of Geoscience And Mineral Resources | Apparatus and method for localizing underwater anomalous body |
CN109242019A (en) * | 2018-09-01 | 2019-01-18 | 哈尔滨工程大学 | A kind of water surface optics Small object quickly detects and tracking |
CN109559321A (en) * | 2018-11-28 | 2019-04-02 | 清华大学 | A kind of sonar image dividing method and equipment |
CN111426693A (en) * | 2020-04-26 | 2020-07-17 | 湖南恒岳重钢钢结构工程有限公司 | Quality defect detection system and detection method thereof |
CN113052872A (en) * | 2021-03-12 | 2021-06-29 | 浙江大学 | Underwater moving object tracking method based on sonar image |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101673403A (en) * | 2009-10-10 | 2010-03-17 | 安防制造(中国)有限公司 | Target following method in complex interference scene |
CN102215377A (en) * | 2010-04-07 | 2011-10-12 | 北京智安邦科技有限公司 | Device and method for PTZ (planning, tilting and zooming) single-target automatic tracking |
CN102915545A (en) * | 2012-09-20 | 2013-02-06 | 华东师范大学 | OpenCV(open source computer vision library)-based video target tracking algorithm |
KR101279561B1 (en) * | 2012-01-19 | 2013-06-28 | 광운대학교 산학협력단 | A fast and accurate face detection and tracking method by using depth information |
CN103996208A (en) * | 2014-05-21 | 2014-08-20 | 国家电网公司 | Method for conducting automatic tracking of PTZ single target in video image |
CN105046258A (en) * | 2015-09-08 | 2015-11-11 | 中国电子科技集团公司第三研究所 | Target detection method and target detection device for small target detection sonar images |
-
2015
- 2015-11-17 CN CN201510789128.5A patent/CN105427342B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101673403A (en) * | 2009-10-10 | 2010-03-17 | 安防制造(中国)有限公司 | Target following method in complex interference scene |
CN102215377A (en) * | 2010-04-07 | 2011-10-12 | 北京智安邦科技有限公司 | Device and method for PTZ (planning, tilting and zooming) single-target automatic tracking |
KR101279561B1 (en) * | 2012-01-19 | 2013-06-28 | 광운대학교 산학협력단 | A fast and accurate face detection and tracking method by using depth information |
CN102915545A (en) * | 2012-09-20 | 2013-02-06 | 华东师范大学 | OpenCV(open source computer vision library)-based video target tracking algorithm |
CN103996208A (en) * | 2014-05-21 | 2014-08-20 | 国家电网公司 | Method for conducting automatic tracking of PTZ single target in video image |
CN105046258A (en) * | 2015-09-08 | 2015-11-11 | 中国电子科技集团公司第三研究所 | Target detection method and target detection device for small target detection sonar images |
Cited By (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107526087A (en) * | 2016-06-21 | 2017-12-29 | 北京臻迪科技股份有限公司 | A kind of method and system for obtaining underwater 3D faultage images |
CN106114095A (en) * | 2016-06-27 | 2016-11-16 | 中国石油大学(华东) | A kind of amphibious sniffing robot |
CN106114095B (en) * | 2016-06-27 | 2018-08-03 | 中国石油大学(华东) | A kind of amphibious sniffing robot |
CN106338733B (en) * | 2016-09-09 | 2018-11-09 | 河海大学常州校区 | Forward-Looking Sonar method for tracking target based on frogeye visual characteristic |
CN106338733A (en) * | 2016-09-09 | 2017-01-18 | 河海大学常州校区 | Forward-looking sonar object tracking method based on frog-eye visual characteristic |
US20180120472A1 (en) * | 2016-10-28 | 2018-05-03 | Korea Institute Of Geoscience And Mineral Resources | Apparatus and method for localizing underwater anomalous body |
CN106910204A (en) * | 2016-12-30 | 2017-06-30 | 中国人民解放军空军预警学院监控***工程研究所 | A kind of method and system to the automatic Tracking Recognition of sea ship |
CN106910204B (en) * | 2016-12-30 | 2018-04-27 | 中国人民解放军空军预警学院监控***工程研究所 | A kind of method and system to the automatic Tracking Recognition of sea ship |
CN107290731A (en) * | 2017-06-21 | 2017-10-24 | 哈尔滨工业大学 | A kind of radar track initial mode based on image area template matches |
CN107290731B (en) * | 2017-06-21 | 2019-12-10 | 哈尔滨工业大学 | Radar track starting method based on image domain template matching |
CN109242019A (en) * | 2018-09-01 | 2019-01-18 | 哈尔滨工程大学 | A kind of water surface optics Small object quickly detects and tracking |
CN109559321A (en) * | 2018-11-28 | 2019-04-02 | 清华大学 | A kind of sonar image dividing method and equipment |
CN111426693A (en) * | 2020-04-26 | 2020-07-17 | 湖南恒岳重钢钢结构工程有限公司 | Quality defect detection system and detection method thereof |
CN113052872A (en) * | 2021-03-12 | 2021-06-29 | 浙江大学 | Underwater moving object tracking method based on sonar image |
CN113052872B (en) * | 2021-03-12 | 2022-06-10 | 浙江大学 | Underwater moving object tracking method based on sonar image |
Also Published As
Publication number | Publication date |
---|---|
CN105427342B (en) | 2018-06-08 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105427342A (en) | Method and system for detecting and tracking underwater small-target sonar image target | |
Chen et al. | Ship detection from coastal surveillance videos via an ensemble Canny-Gaussian-morphology framework | |
CN108805904B (en) | Moving ship detection and tracking method based on satellite sequence image | |
He et al. | Inshore ship detection in remote sensing images via weighted pose voting | |
CN102542289B (en) | Pedestrian volume statistical method based on plurality of Gaussian counting models | |
CN101214851B (en) | Intelligent all-weather actively safety early warning system and early warning method thereof for ship running | |
CN102598057B (en) | Method and system for automatic object detection and subsequent object tracking in accordance with the object shape | |
CN104463877B (en) | A kind of water front method for registering based on radar image Yu electronic chart information | |
CN108664930A (en) | A kind of intelligent multi-target detection tracking | |
CN103514448A (en) | Method and system for navicular identification | |
Lipschutz et al. | New methods for horizon line detection in infrared and visible sea images | |
CN109188430B (en) | Target extraction method based on ground monitoring radar system | |
CN111323756B (en) | Marine radar target detection method and device based on deep learning | |
CN108573280B (en) | Method for unmanned ship to autonomously pass through bridge | |
CN116109936B (en) | Target detection and identification method based on optical remote sensing | |
CN109063669B (en) | Bridge area ship navigation situation analysis method and device based on image recognition | |
CN111323757B (en) | Target detection method and device for marine radar | |
Li et al. | Real-time underwater target detection for AUV using side scan sonar images based on deep learning | |
CN104268574A (en) | SAR image change detecting method based on genetic kernel fuzzy clustering | |
Liu et al. | A multi-scale feature pyramid SAR ship detection network with robust background interference | |
CN103996207A (en) | Object tracking method | |
Zou et al. | Sonar Image Target Detection for Underwater Communication System Based on Deep Neural Network. | |
Tang et al. | A novel approach for automatic recognition of LPI radar waveforms based on CNN and attention mechanisms | |
CN104463896A (en) | Image corner point detection method and system based on kernel similar region distribution characteristics | |
Wan et al. | Orientation Detector for Small Ship Targets in SAR Images Based on Semantic Flow Feature Alignment and Gaussian Label Matching |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |