CN107341793A - A kind of target surface image processing method and device - Google Patents
A kind of target surface image processing method and device Download PDFInfo
- Publication number
- CN107341793A CN107341793A CN201710488510.1A CN201710488510A CN107341793A CN 107341793 A CN107341793 A CN 107341793A CN 201710488510 A CN201710488510 A CN 201710488510A CN 107341793 A CN107341793 A CN 107341793A
- Authority
- CN
- China
- Prior art keywords
- current frame
- frame image
- image
- gray value
- shell hole
- 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.)
- Pending
Links
- 238000003672 processing method Methods 0.000 title claims abstract description 10
- 238000012545 processing Methods 0.000 claims abstract description 50
- 238000012937 correction Methods 0.000 claims abstract description 38
- 239000000284 extract Substances 0.000 claims abstract description 16
- 238000000034 method Methods 0.000 claims description 23
- 238000004422 calculation algorithm Methods 0.000 claims description 21
- 239000011435 rock Substances 0.000 abstract description 7
- 238000005286 illumination Methods 0.000 abstract description 5
- 238000001514 detection method Methods 0.000 description 12
- 238000000605 extraction Methods 0.000 description 9
- 238000004458 analytical method Methods 0.000 description 8
- 238000010586 diagram Methods 0.000 description 8
- 238000004590 computer program Methods 0.000 description 7
- 230000006870 function Effects 0.000 description 5
- 230000000694 effects Effects 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 239000011159 matrix material Substances 0.000 description 4
- 238000012360 testing method Methods 0.000 description 4
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 230000003044 adaptive effect Effects 0.000 description 2
- 238000003860 storage Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000001149 cognitive effect Effects 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000003708 edge detection Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000005096 rolling process Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000009738 saturating Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
- G06T7/001—Industrial image inspection using an image reference approach
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/80—Geometric correction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a kind of target surface image processing method and device, including:Template two field picture is obtained, and extracts target surface edge contour;Determine some first angle points on the target surface edge contour of the template two field picture;Current frame image is obtained, and extracts target surface edge contour;Determine some second angle points on the target surface edge contour of the current frame image;Determine that the second angle point corresponding with each first angle point forms characteristic point pair according to characteristic matching degree on each first angle point periphery;According to each characteristic point to carrying out the current frame image after being corrected to current frame image.The processing that the present invention is carried out based on frame of video, can avoid influence of the change of illumination condition to indication of shots precision;The higher distortion correction of precision can be carried out to shaking motion video simultaneously, solve the problems, such as that target surface rocks.
Description
Technical field
The present invention relates to technical field of image processing, more particularly to a kind of target surface image processing method and device.
Background technology
The automatic target-indicating algorithm for being currently based on image recognition mainly has two ways, the identification of shell hole based on single-frame images and
How shell hole identification method based on video sequence, solve the target caused by the impact of environmental factor or bullet in identification process
The problem of face is rocked, and accurately extraction shell hole and the identification to repeated hole (shell hole to partially overlap) are always from frame of video
One important research topic.
In traditional technical scheme based on frame of video, after target surface region is obtained, directly to front and rear video
Frame is made the difference, and then carries out connected domain analysis to error image, and shell hole is extracted by Feature Selections such as area, circularity.It is for example, a kind of
Shell hole recognition methods based on image technique includes:Pretreatment stage:Adjacent two field pictures are gathered as picture to be identified, and point
Indescribably take includes the area-of-interest of target position just in two frames picture to be identified, different thresholds is used for regions of different colours
Value, frame difference processing is carried out to the area-of-interest in two frames picture to be identified, obtains picture after frame difference;Cognitive phase:Utilize side
Profile information after edge detection algorithm searching frame difference in picture;The confidence level stage:Calculate each profile absolute area confidence level,
Length-width ratio confidence level and area dutycycle confidence level, absolute area confidence level, length-width ratio confidence level and area dutycycle
Confidence level sum highest profile as identifies obtained shell hole.In the program, be by frame it is poor, extraction profile, Feature Selection
Carry out the extraction of shell hole.
In the technical scheme based on single-frame images, target surface region is obtained first, then image is entered according to shell hole feature
Row analysis, such as according to shell hole area grayscale value is relatively low, feature localized region is contrasted with surrounding contrast's degree is more apparent etc.
Degree analysis;Then wherein shell hole is extracted, newly-increased shell hole is obtained further according to history testing result.
The deficiencies in the prior art are:
For the Bullet holes detection scheme based on frame of video, although do not rock in target surface, in the absence of repeated hole in the case of detect
Effect it is preferable, but due to not corrected to the situation of rocking or correction accuracy is inadequate, and spy is not done to repeated hole situation
Different analysis, it is therefore, relatively low in the accuracy rate of both in particular cases detections to shell hole;
For the Bullet holes detection scheme based on single-frame images, due to being that directly image is analyzed, can solve target surface rolling
The dynamic influence brought, but this mode is more sensitive to illumination variation, it is more apparent in outdoor illumination variation and be likely to occur light
It is not strong according to adaptability in the case of uneven.
The content of the invention
The invention provides a kind of target surface image processing method and device, and target surface image is examined to solve rocking for target surface
Survey the influence of effect.
The embodiments of the invention provide a kind of target surface image processing method, including:
Template two field picture is obtained, and extracts target surface edge contour;
Determine some first angle points on the target surface edge contour of the template two field picture;
Current frame image is obtained, and extracts target surface edge contour;
Determine some second angle points on the target surface edge contour of the current frame image;
Determine that the second angle point corresponding with each first angle point forms spy according to characteristic matching degree on each first angle point periphery
Sign point pair;
According to each characteristic point to carrying out the current frame image after being corrected to current frame image.
It is preferred that the template two field picture is when detecting shell hole, the image of shell hole is detected.
It is preferred that the current video two field picture is the newest two field picture collected.
It is root it is preferred that according to each characteristic point to carrying out the current frame image after being corrected to current frame image
According to each characteristic point to the current frame image after being corrected after being fitted to current frame image.
It is preferred that when according to each characteristic point to being fitted to current frame image, it is fitted using RANSAC algorithms.
It is preferred that further comprise:
The shell hole on image is determined on current frame image;
After carrying out Connected area disposal$ to current frame image, determined whether according to connected region area for repeated hole.
It is preferred that the shell hole on image is determined on current frame image, including:
After being corrected to current frame image, determine current frame image and previous frame correction after current frame image between
Gray value;
The shell hole on current frame image is determined according to the difference of the grey scale change degree of gray value and average gray.
It is preferred that determined whether according to connected region area for repeated hole, including:
History shell hole position is determined, and determines that connected region area is less than the shell hole of predetermined threshold value;
Determine whether the shell hole is repeated hole according to the shell hole position and history shell hole position.
It is preferred that further comprise:
After being corrected to current frame image, the gray value differences between current frame image after being corrected according to previous frame
Carry out interference processing.
It is preferred that the gray value differences between current frame image after being corrected according to previous frame carry out interference processing, including:
It is determined that on current frame image after correction pixel position, and the first gray value on the position;
Determine the region that this is positioned adjacent on the current frame image after previous frame correction, and second gray scale in the region
Value;
First gray value and the second gray value are carried out making poor processing.
It is preferred that the second gray value be in the region with the immediate gray value of the first gray value.
The embodiments of the invention provide a kind of target surface image processing apparatus, including:
Angle point determining module, for obtaining template two field picture, and extract target surface edge contour;Determine the template two field picture
Target surface edge contour on some first angle points;Current frame image is obtained, and extracts target surface edge contour;Determine described current
Some second angle points on the target surface edge contour of two field picture;
Characteristic point is to module, for corresponding with each first angle point according to the determination of characteristic matching degree on each first angle point periphery
The second angle point formed characteristic point pair;
Rectification module, for according to each characteristic point to carrying out the present frame figure after being corrected to current frame image
Picture.
It is preferred that angle point determining module is further used for detect the image of shell hole as described in when detecting shell hole
Template two field picture.
It is preferred that the newest two field picture that angle point determining module is further used for collecting is as the current video
Two field picture.
It is preferred that rectification module is further used for according to each characteristic point to being corrected after being fitted to current frame image
Current frame image afterwards.
It is preferred that rectification module is further used for when according to each characteristic point to being fitted to current frame image, use
RANSAC algorithms are fitted.
It is preferred that further comprise:
Shell hole determining module, for determining the shell hole on image on current frame image;
Repeated hole determining module, after carrying out Connected area disposal$ to current frame image, it is according to the determination of connected region area
No is repeated hole.
It is preferred that shell hole determining module is further used for when determining the shell hole on image on current frame image, including:
After being corrected to current frame image, determine current frame image and previous frame correction after current frame image between
Gray value;
The shell hole on current frame image is determined according to the difference of the grey scale change degree of gray value and average gray.
It is preferred that repeated hole determining module be further used for according to connected region area determine whether for repeated hole when, including:
History shell hole position is determined, and determines that connected region area is less than the shell hole of predetermined threshold value;
Determine whether the shell hole is repeated hole according to the shell hole position and history shell hole position.
It is preferred that further comprise:
Processing module is disturbed, for after being corrected to current frame image, according to the present frame after being corrected with previous frame
Gray value differences between image carry out interference processing.
It is preferred that interference processing module is further used for according to the ash between the current frame image after previous frame correction
When angle value difference carries out interference processing, including:
It is determined that on current frame image after correction pixel position, and the first gray value on the position;
Determine the region that this is positioned adjacent on the current frame image after previous frame correction, and second gray scale in the region
Value;
First gray value and the second gray value are carried out making poor processing.
It is preferred that interference processing module be further used for using in the region with the immediate gray value of the first gray value as
Second gray value.
The present invention has the beneficial effect that:
In technical scheme provided in an embodiment of the present invention, extraction template two field picture and current frame image target surface edge contour
Afterwards, it is determined that some angle points on the two profile;And characteristic point pair is formed, then according to each characteristic point to current frame image
Carry out the current frame image after being corrected.By being then based on the processing of frame of video progress, therefore avoid illumination condition
Influence of the change to indication of shots precision;Simultaneously because employing according to feature angle point to correcting, therefore enter to shaking motion video
The higher distortion correction of precision of having gone, solves the problems, such as that target surface rocks.
Further, due to after being corrected to current frame image, always according to the present frame figure after being corrected with previous frame
Gray value differences as between carry out interference processing, thus be excluded that the interference that is likely to occur.
Further, determine whether always according to connected region area for repeated hole, due to individually being analyzed repeated hole problem,
Solve the repeated hole test problems under certain registration.
Brief description of the drawings
Accompanying drawing described herein is used for providing a further understanding of the present invention, forms the part of the present invention, this hair
Bright schematic description and description is used to explain the present invention, does not form inappropriate limitation of the present invention.In the accompanying drawings:
Fig. 1 is target surface image processing method implementation process diagram in the embodiment of the present invention;
Fig. 2 is target surface image processing apparatus structural representation in the embodiment of the present invention.
Embodiment
For technical scheme provided in an embodiment of the present invention by for the rocking of target surface, repeated hole problem, target is that solve target surface
Rock with the influence to Bullet holes detection effect such as repeated hole, improve the accuracy rate of automatic target-indicating.Below in conjunction with the accompanying drawings to the present invention's
Embodiment illustrates.
Fig. 1 is target surface image processing method implementation process diagram, as illustrated, can include:
Step 101, template two field picture is obtained, and extract target surface edge contour;
Step 102, determine some first angle points on the target surface edge contour of the template two field picture;
Step 103, current frame image is obtained, and extract target surface edge contour;
Step 104, determine some second angle points on the target surface edge contour of the current frame image;
Step 105, each first angle point periphery according to characteristic matching degree determine it is corresponding with each first angle point second jiao
Point forms characteristic point pair;
Step 106, according to each characteristic point to being corrected to current frame image.
It is to complete to distort to target surface image caused by rocking using the feature angle point of target surface image border in scheme
Correction.
Implementation on angle point can be as follows:
If a point is selected from simultaneously very big blank wall, then it is difficult to trace into this point in subsequent frame of video
, because being just as a little on wall or similar, if on the contrary, selection is unique or close to unique point,
The probability for so finding this point again is just very big.So what point be considered it is unique, it is considered that one
The point for having obvious derivative on two orthogonal directions is more likely to be unique.Many traceable characteristic points are all referred to as
Angle point.
Intuitively say, angle point is one kind containing enough information and be able to can extracted from present frame and next frame
Point.
The angle point most generally used defines to be put forward by Harris, and the basis of definition is the second order of gradation of image intensity
Jacobian matrix.Harris Corner Detections process can be as follows:
Harris Corner Detection basic thoughts:The characteristics of image from the wicket of image local, if window is to any
The movement in direction results in the significant change of gradation of image, then has angle point in window.
In scheme, exactly because the reason for going to detect angle point angle point has unique feature and is not easy to change in the picture
Become.
Implementation explanation is carried out with reference to specific scheme.
Template two field picture is obtained in step 101,102, and extracts target surface edge contour, determines the target of the template two field picture
In the implementation of some first angle points on the edge contour of face, template two field picture can be obtained, then carries out adaptive threshold two-value
The feature extraction target surface edge wheels such as the processing such as change, closing operation of mathematical morphology, connected domain analysis, area, length-width ratio further according to target surface
It is wide.Then the Corner Detection of sub-pixel precision is carried out, the gray feature of extraction angle point region, preserves testing result.
Wherein, adaptive threshold binaryzation, closing operation of mathematical morphology, connected domain analysis, according to the area of target surface, length-width ratio etc.
Feature extraction target surface edge contour it will be understood by those skilled in the art that easily know, meanwhile, using other technologies means come
It is also that can reach scheme purpose to obtain target surface edge contour.
It is to determine the angle point on image outline to extract profile, so first extracting profile, angle is then extracted on profile
Point, and the extraction of angle point may refer to preceding description and use Harris Corner Detection schemes, certainly, using other technologies scheme
It is feasible.The present invention is to determine angle point according to angle character in implementing, and the acquisition modes of angle point have many kinds, but right
For chest silhouette target, so preferable scheme of effect.
In implementation, template two field picture is the image that shell hole is detected when detecting shell hole.
Specifically, template frame can be initial frame, when then detecting shell hole every time, it is updated, shell hole will be detected
Image as template two field picture.Wherein, initial frame just refers to that camera starts the first two field picture of detection.
The gray value being related in implementation is the value of each pixel on image, is required in each step of image procossing
Use basic parameter.
Current frame image is obtained in step 103,104, and extracts target surface edge contour, determines the target of the current frame image
In the implementation of some second angle points on the edge contour of face, step 101,102 pairs of masterplate frame embodiments are may be referred to.
Specifically, handling current frame of video (namely present frame), above-mentioned image preprocessing flow is equally also carried out
Image border is obtained, then carries out Corner Detection in the neighborhood of several pixel distances of template frame corner location, its purpose
It is the correction then being had an X-rayed in order to extract the angle point at edge, solves to rock problem before and after target surface.
The angle point detected will use in step 105, that is, calculating characteristic matching degree, find out matching degree highest
Corner location.The characteristic point pair being so commonly available on the edge contour of 20 or so matchings.
In implementation, the current video two field picture is the newest two field picture collected.
Specifically, current video frame refers to collect that newest two field picture, just as the picture that people is currently seen.
In specific implementation, in order to reduce treating capacity, can only template frame corner location several pixel distances neighbour
Corner Detection is carried out in domain, in concrete practice, is determined according to mass data and experimental result, generally 20 pixel distances.
It is to present frame figure according to each characteristic point according to each characteristic point to being corrected to current frame image in implementation
As the current frame image after being corrected after being fitted.
Specific implementation can be as follows:
Corner Detection is carried out in the neighborhood of several pixel distances of template frame corner location, calculates characteristic matching journey
Degree, when finding out matching degree highest corner location, is in each angle neighborhood of a point of template frame, looks for present frame most phase
As angle point.It is based on the corner location on masterplate frame, according to the corner location in current frame of video, to current frame of video
Corrected.
Find out matching degree highest corner location obtain 20 or so matching edge contour on characteristic point pair when,
It is after the angle point of 20 or so is extracted on template frame, has the angle of a matching degree highest present frame in each angle point
Point, the point pair of the point pair of 20 angle points, characteristic point pair namely angle point has been achieved in that it.
When finding out matching degree highest corner location, matching degree highest is the corner location of present frame.
In implementation, when according to each characteristic point to being fitted to current frame image, RANSAC algorithms can be used to carry out
Fitting.
Specifically, using RANSAC (Random Sample Consensus, random sampling uniformity) algorithm to matching
Fitting of the characteristic point to progress perspective matrix, the minimum perspective matrix of mean square error is obtained, then current frame image is carried out saturating
Depending on conversion, the image after being corrected, the position of the target in the image obtained after correction is identical with the position of target in template frame, just
The processing such as frame difference is carried out later.
Specifically, general pattern linear distortion can be corrected by perspective transform.
In specific implementation, in addition to RANSAC algorithms, existing conventional meanses can also be used:Least-squares algorithm,
RANSAC algorithms are advantageous in that, can reject noise spot, and fitting result is more accurate.
Wherein, when being fitted using RANSAC algorithms, the input of RANSAC algorithms is that one group of observation data (often contains
Have larger noise or Null Spot), a parameterized model and some believable parameters for being used to explain observation data.
RANSAC reaches target by one group of random subset being chosen in data.The subset being selected is assumed to be intra-office point,
And it can be verified with following manner:
There is the intra-office point that a model is adapted to hypothesis, i.e., all unknown parameters can calculate from the intra-office point of hypothesis
Go out.
Gone to test all other data with obtained model, if some point is suitable for the model of estimation, it is believed that it
It is intra-office point.
If enough points are classified as the intra-office point of hypothesis, then the model of estimation is just reasonable enough.
Then, go to reevaluate model (for example using least square method) with the intra-office point of all hypothesis, because it is only
By initial hypothesis intra-office point estimation.
Finally, by estimating the error rate of intra-office point and model come assessment models.
Said process is repeatedly executed fixed number, otherwise model caused by every time is because intra-office point is given up very little
Abandon, otherwise it is selected because of more preferable than existing model.
It is to complete to distort to target surface image caused by rocking using the feature angle point of target surface image border in scheme
Correction, further, can also be then complete using neighborhood difference algorithm in force by correction accuracy control in two pixels
The work of correction rear video frame is poor in pairs, eliminates image and rocks the influence brought.
That is, it can further include in implementing:
After being corrected to current frame image, the gray value differences between current frame image after being corrected according to previous frame
Carry out interference processing.
In specific implementation, the gray value differences between current frame image after being corrected according to previous frame carry out interference processing,
It can include:
It is determined that on current frame image after correction pixel position, and the first gray value on the position;
Determine the region that this is positioned adjacent on the current frame image after previous frame correction, and second gray scale in the region
Value;
First gray value and the second gray value are carried out making poor processing.
In specific implementation, the second gray value be in the region with the immediate gray value of the first gray value.
Specifically, in implementing, the part processing mode is referred to as " neighborhood difference algorithm " processing, it is right in neighborhood difference processing
It is poor that present frame after correction and previous frame image are made, and makees poor mode and uses neighborhood difference algorithm, i.e., to each pixel on image,
It is poor with the immediate value work of its gray value to be found in 3*3 the or 5*5 neighborhoods of another image, obtains error image, neighborhood is poor
Algorithm can eliminate the influence of 1 to 2 pixel-shifts of image after correction, remove disturbing factor, facilitate shell hole in error image
Extraction.
Wherein, previous frame image is also with the image after the correction of masterplate frame.
Neighborhood difference algorithm is that picture noise is big after the frame difference caused by direct frame difference in order to solve, and perspective transform correction
The problem of precision is inadequate, be during specific implementation when make difference processing to each pixel in image be not it is direct with it is another
The point of image correspondence position subtracts each other, but finds one and its immediate point in a certain neighborhood of correspondence position on another image
Subtracted each other.
For example, the gray value that some is put on image is 255, correspondence position value is 200 on another image, but the ash of surrounding
Angle value is 255, then the result that direct frame difference obtains is 55, and the result that neighborhood difference obtains is 0, therefore reduces the dry of noise
Disturb.
For the influence of repeated hole, can also further comprise in implementation after interference processing:
The shell hole on image is determined on current frame image;
After carrying out Connected area disposal$ to current frame image, determined whether according to connected region area for repeated hole.
In specific implementation, the shell hole on image is determined on current frame image, can be included:
After being corrected to current frame image, determine current frame image and previous frame correction after current frame image between
Gray value;
The shell hole on current frame image is determined according to the difference of the grey scale change degree of gray value and average gray.
In specific implementation, determine whether for repeated hole, to include according to connected region area:
History shell hole position is determined, and determines that connected region area is less than the shell hole of predetermined threshold value;
Determine whether the shell hole is repeated hole according to the shell hole position and history shell hole position.
Specifically, when being extracted to shell hole, connected domain analysis is carried out to error image, according to area, length, width
Etc. the preliminary screening that feature carries out shell hole region, secondary judgement then is carried out to doubtful shell hole region.
By calculating grey scale change degree and the present frame suspicious region of present frame and former frame suspicious region in implementation
Carry out determining whether shell hole with the two features of the difference of the average gray of peripheral region.During specific implementation, when gray scale becomes
Change degree and the difference of average gray the two values are possible to be shell hole when being more than a certain threshold value.
It may doubtful be repeated hole if connected region area is smaller, must have history shell hole near repeated hole, can be by inquiring about ratio
Determine whether repeated hole compared with the position of history shell hole and doubtful repeated hole.Specifically, repeated hole is that is there is two holes to overlap, but
It is not to be completely superposed, then in the shell hole of the coincidence detected, its side will also have a history shell hole.
Based on same inventive concept, a kind of target surface image processing apparatus is additionally provided in the embodiment of the present invention, due to the dress
Put that the principle solved the problems, such as is similar to a kind of target surface image processing method, therefore the implementation of the device may refer to the reality of method
Apply, repeat part and repeat no more.
Fig. 2 is target surface image processing apparatus structural representation, as illustrated, can include:
Angle point determining module 201, for obtaining template two field picture, and extract target surface edge contour;Determine the template frame
Some first angle points on the target surface edge contour of image;Current frame image is obtained, and extracts target surface edge contour;It is it is determined that described
Some second angle points on the target surface edge contour of current frame image;
Characteristic point is to module 202, for being determined and each first angle point according to characteristic matching degree on each first angle point periphery
Corresponding second angle point forms characteristic point pair;
Rectification module 203, for according to each characteristic point to carrying out the present frame after being corrected to current frame image
Image.
In implementation, angle point determining module is further used for detect the image of shell hole as described in when detecting shell hole
Template two field picture.
In implementation, the newest two field picture that angle point determining module is further used for collecting is as the current video
Two field picture.
In implementation, rectification module is further used for according to each characteristic point to being corrected after being fitted to current frame image
Current frame image afterwards.
In implementation, rectification module is further used for when according to each characteristic point to being fitted to current frame image, uses
RANSAC algorithms are fitted.
In implementation, further comprise:
Shell hole determining module 204, for determining the shell hole on image on current frame image;
Repeated hole determining module 205, after carrying out Connected area disposal$ to current frame image, determined according to connected region area
Whether it is repeated hole.
In implementation, shell hole determining module is further used for when determining the shell hole on image on current frame image, including:
After being corrected to current frame image, determine current frame image and previous frame correction after current frame image between
Gray value;
The shell hole on current frame image is determined according to the difference of the grey scale change degree of gray value and average gray.
In implementation, repeated hole determining module be further used for according to connected region area determine whether for repeated hole when, including:
History shell hole position is determined, and determines that connected region area is less than the shell hole of predetermined threshold value;
Determine whether the shell hole is repeated hole according to the shell hole position and history shell hole position.
In implementation, further comprise:
Processing module 206 is disturbed, for after being corrected to current frame image, according to current after being corrected with previous frame
Gray value differences between two field picture carry out interference processing.
In implementation, interference processing module is further used for according to the ash between the current frame image after previous frame correction
When angle value difference carries out interference processing, including:
It is determined that on current frame image after correction pixel position, and the first gray value on the position;
Determine the region that this is positioned adjacent on the current frame image after previous frame correction, and second gray scale in the region
Value;
First gray value and the second gray value are carried out making poor processing.
In implementation, interference processing module be further used for using in the region with the immediate gray value of the first gray value as
Second gray value.
For convenience of description, each several part of apparatus described above is divided into various modules with function or unit describes respectively.
Certainly, each module or the function of unit can be realized in same or multiple softwares or hardware when implementing of the invention.
In summary, in technical scheme provided in an embodiment of the present invention, there is provided made using target surface image border angle point
It is characterized a little, the mode that perspective matrix is fitted using Feature Points Matching RANSAC algorithms corrects fault image;And use neighborhood
Difference algorithm eliminates the influence that minor shifts bring, using connected domain analysis, characteristic filter come determine whether shell hole and whether
For repeated hole etc..
Due to using the processing mode based on frame of video, influence of the change of illumination condition to indication of shots precision is avoided;It is right
Shake motion video and carried out the higher distortion correction of precision, solve the problems, such as that target surface rocks;Neighborhood is carried out to the image after correction
Difference processing, eliminates the interference being likely to occur;Repeated hole problem is individually analyzed, solves the repeated hole inspection under certain registration
Survey problem.
It should be understood by those skilled in the art that, embodiments of the invention can be provided as method, system or computer program
Product.Therefore, the present invention can use the reality in terms of complete hardware embodiment, complete software embodiment or combination software and hardware
Apply the form of example.Moreover, the present invention can use the computer for wherein including computer usable program code in one or more
The shape for the computer program product that usable storage medium is implemented on (including but is not limited to magnetic disk storage and optical memory etc.)
Formula.
The present invention is the flow with reference to method according to embodiments of the present invention, equipment (system) and computer program product
Figure and/or block diagram describe.It should be understood that can be by every first-class in computer program instructions implementation process figure and/or block diagram
Journey and/or the flow in square frame and flow chart and/or block diagram and/or the combination of square frame.These computer programs can be provided
The processors of all-purpose computer, special-purpose computer, Embedded Processor or other programmable data processing devices is instructed to produce
A raw machine so that produced by the instruction of computer or the computing device of other programmable data processing devices for real
The device for the function of being specified in present one flow of flow chart or one square frame of multiple flows and/or block diagram or multiple square frames.
These computer program instructions, which may be alternatively stored in, can guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works so that the instruction being stored in the computer-readable memory, which produces, to be included referring to
Make the manufacture of device, the command device realize in one flow of flow chart or multiple flows and/or one square frame of block diagram or
The function of being specified in multiple square frames.
These computer program instructions can be also loaded into computer or other programmable data processing devices so that counted
Series of operation steps is performed on calculation machine or other programmable devices to produce computer implemented processing, so as in computer or
The instruction performed on other programmable devices is provided for realizing in one flow of flow chart or multiple flows and/or block diagram one
The step of function of being specified in individual square frame or multiple square frames.
Obviously, those skilled in the art can carry out the essence of various changes and modification without departing from the present invention to the present invention
God and scope.So, if these modifications and variations of the present invention belong to the scope of the claims in the present invention and its equivalent technologies
Within, then the present invention is also intended to comprising including these changes and modification.
Claims (22)
- A kind of 1. target surface image processing method, it is characterised in that including:Template two field picture is obtained, and extracts target surface edge contour;Determine some first angle points on the target surface edge contour of the template two field picture;Current frame image is obtained, and extracts target surface edge contour;Determine some second angle points on the target surface edge contour of the current frame image;Determine that the second angle point corresponding with each first angle point forms characteristic point according to characteristic matching degree on each first angle point periphery It is right;According to each characteristic point to carrying out the current frame image after being corrected to current frame image.
- 2. the method as described in claim 1, it is characterised in that the template two field picture is when detecting shell hole, to detect bullet The image in hole.
- 3. the method as described in claim 1, it is characterised in that the current video two field picture is the newest frame collected Image.
- 4. the method as described in claim 1, it is characterised in that according to each characteristic point to carrying out correction acquisition to current frame image Current frame image after correction, it is to the present frame figure after being corrected after being fitted to current frame image according to each characteristic point Picture.
- 5. method as claimed in claim 4, it is characterised in that according to each characteristic point to being fitted to current frame image When, it is fitted using RANSAC algorithms.
- 6. the method as described in claim 1, it is characterised in that further comprise:The shell hole on image is determined on current frame image;After carrying out Connected area disposal$ to current frame image, determined whether according to connected region area for repeated hole.
- 7. method as claimed in claim 6, it is characterised in that the shell hole on image is determined on current frame image, including:After being corrected to current frame image, the ash between the current frame image after current frame image and previous frame correction is determined Angle value;The shell hole on current frame image is determined according to the difference of the grey scale change degree of gray value and average gray.
- 8. method as claimed in claim 6, it is characterised in that determined whether according to connected region area for repeated hole, including:History shell hole position is determined, and determines that connected region area is less than the shell hole of predetermined threshold value;Determine whether the shell hole is repeated hole according to the shell hole position and history shell hole position.
- 9. the method as described in claim 1 to 8 is any, it is characterised in that further comprise:After being corrected to current frame image, the gray value differences between the current frame image after being corrected according to previous frame are carried out Interference is handled.
- 10. method as claimed in claim 9, it is characterised in that between the current frame image after being corrected according to previous frame Gray value differences carry out interference processing, including:It is determined that on current frame image after correction pixel position, and the first gray value on the position;Determine the region that this is positioned adjacent on the current frame image after previous frame correction, and second gray value in the region;First gray value and the second gray value are carried out making poor processing.
- 11. method as claimed in claim 10, it is characterised in that the second gray value is most connect with the first gray value in the region Near gray value.
- A kind of 12. target surface image processing apparatus, it is characterised in that including:Angle point determining module, for obtaining template two field picture, and extract target surface edge contour;Determine the target of the template two field picture Some first angle points on the edge contour of face;Current frame image is obtained, and extracts target surface edge contour;Determine the present frame figure Some second angle points on the target surface edge contour of picture;Characteristic point is to module, for determining corresponding with each first angle point the according to characteristic matching degree on each first angle point periphery Two angle points form characteristic point pair;Rectification module, for according to each characteristic point to carrying out the current frame image after being corrected to current frame image.
- 13. device as claimed in claim 12, it is characterised in that angle point determining module is further used for that shell hole will be being detected When detect the image of shell hole as the template two field picture.
- 14. device as claimed in claim 12, it is characterised in that angle point determining module is further used for newest by what is collected A two field picture as the current video two field picture.
- 15. device as claimed in claim 12, it is characterised in that rectification module is further used for according to each characteristic point to working as Prior image frame corrected after being fitted after current frame image.
- 16. device as claimed in claim 15, it is characterised in that rectification module be further used for according to each characteristic point to right When current frame image is fitted, it is fitted using RANSAC algorithms.
- 17. device as claimed in claim 12, it is characterised in that further comprise:Shell hole determining module, for determining the shell hole on image on current frame image;Repeated hole determining module, for current frame image carry out Connected area disposal$ after, according to connected region area determine whether for Repeated hole.
- 18. device as claimed in claim 17, it is characterised in that shell hole determining module is further used on current frame image When determining the shell hole on image, including:After being corrected to current frame image, the ash between the current frame image after current frame image and previous frame correction is determined Angle value;The shell hole on current frame image is determined according to the difference of the grey scale change degree of gray value and average gray.
- 19. device as claimed in claim 17, it is characterised in that repeated hole determining module is further used for according to connected region Area determine whether for repeated hole when, including:History shell hole position is determined, and determines that connected region area is less than the shell hole of predetermined threshold value;Determine whether the shell hole is repeated hole according to the shell hole position and history shell hole position.
- 20. the device as described in claim 12 to 19 is any, it is characterised in that further comprise:Processing module is disturbed, for after being corrected to current frame image, according to the current frame image after being corrected with previous frame Between gray value differences carry out interference processing.
- 21. device as claimed in claim 20, it is characterised in that interference processing module is further used in basis and previous frame When the gray value differences between current frame image after correction carry out interference processing, including:It is determined that on current frame image after correction pixel position, and the first gray value on the position;Determine the region that this is positioned adjacent on the current frame image after previous frame correction, and second gray value in the region;First gray value and the second gray value are carried out making poor processing.
- 22. device as claimed in claim 21, it is characterised in that interference processing module be further used for by the region with the The immediate gray value of one gray value is as the second gray value.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710488510.1A CN107341793A (en) | 2017-06-23 | 2017-06-23 | A kind of target surface image processing method and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710488510.1A CN107341793A (en) | 2017-06-23 | 2017-06-23 | A kind of target surface image processing method and device |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107341793A true CN107341793A (en) | 2017-11-10 |
Family
ID=60221177
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710488510.1A Pending CN107341793A (en) | 2017-06-23 | 2017-06-23 | A kind of target surface image processing method and device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107341793A (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108168375A (en) * | 2017-12-18 | 2018-06-15 | 浙江华睿科技有限公司 | A kind of target scoring method and device |
CN108805210A (en) * | 2018-06-14 | 2018-11-13 | 深圳深知未来智能有限公司 | A kind of shell hole recognition methods based on deep learning |
CN108805144A (en) * | 2018-06-01 | 2018-11-13 | 杭州晨鹰军泰科技有限公司 | Shell hole recognition methods based on morphology correction and system, indication of shots equipment |
CN109990657A (en) * | 2019-05-07 | 2019-07-09 | 武汉高德红外股份有限公司 | It is a kind of based on image registration without target single-shot school rifle method |
CN111914857A (en) * | 2020-08-11 | 2020-11-10 | 上海柏楚电子科技股份有限公司 | Layout method, device and system for excess sheet material, electronic equipment and storage medium |
CN115661798A (en) * | 2022-12-23 | 2023-01-31 | 小米汽车科技有限公司 | Method and device for determining target area, vehicle and storage medium |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102609918A (en) * | 2012-02-15 | 2012-07-25 | 国家***第二海洋研究所 | Image characteristic registration based geometrical fine correction method for aviation multispectral remote sensing image |
CN106408527A (en) * | 2016-08-25 | 2017-02-15 | 安徽水滴科技有限责任公司 | Automatic target scoring method based on video analysis |
CN106802113A (en) * | 2016-12-23 | 2017-06-06 | 西安交通大学 | Intelligent hit telling system and method based on many shell hole algorithm for pattern recognitions |
-
2017
- 2017-06-23 CN CN201710488510.1A patent/CN107341793A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102609918A (en) * | 2012-02-15 | 2012-07-25 | 国家***第二海洋研究所 | Image characteristic registration based geometrical fine correction method for aviation multispectral remote sensing image |
CN106408527A (en) * | 2016-08-25 | 2017-02-15 | 安徽水滴科技有限责任公司 | Automatic target scoring method based on video analysis |
CN106802113A (en) * | 2016-12-23 | 2017-06-06 | 西安交通大学 | Intelligent hit telling system and method based on many shell hole algorithm for pattern recognitions |
Non-Patent Citations (3)
Title |
---|
傅军栋等: "《实景图像拼接及其漫游控制技术》", 30 June 2011, 成都:西南交通大学出版社 * |
李瑞峰: "《工业机器人设计与应用》", 31 January 2017, 哈尔滨工业大学出版社 * |
杨常青: "《舰基图像处理技术原理与应用》", 31 May 2015 * |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108168375A (en) * | 2017-12-18 | 2018-06-15 | 浙江华睿科技有限公司 | A kind of target scoring method and device |
CN108168375B (en) * | 2017-12-18 | 2019-11-19 | 浙江华睿科技有限公司 | A kind of target scoring method and device |
CN108805144A (en) * | 2018-06-01 | 2018-11-13 | 杭州晨鹰军泰科技有限公司 | Shell hole recognition methods based on morphology correction and system, indication of shots equipment |
CN108805210A (en) * | 2018-06-14 | 2018-11-13 | 深圳深知未来智能有限公司 | A kind of shell hole recognition methods based on deep learning |
CN108805210B (en) * | 2018-06-14 | 2022-03-04 | 深圳深知未来智能有限公司 | Bullet hole identification method based on deep learning |
CN109990657A (en) * | 2019-05-07 | 2019-07-09 | 武汉高德红外股份有限公司 | It is a kind of based on image registration without target single-shot school rifle method |
CN109990657B (en) * | 2019-05-07 | 2021-10-01 | 武汉高德红外股份有限公司 | Image registration-based non-target gun calibration method |
CN111914857A (en) * | 2020-08-11 | 2020-11-10 | 上海柏楚电子科技股份有限公司 | Layout method, device and system for excess sheet material, electronic equipment and storage medium |
CN111914857B (en) * | 2020-08-11 | 2023-05-09 | 上海柏楚电子科技股份有限公司 | Layout method, device and system for plate excess material, electronic equipment and storage medium |
CN115661798A (en) * | 2022-12-23 | 2023-01-31 | 小米汽车科技有限公司 | Method and device for determining target area, vehicle and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107341793A (en) | A kind of target surface image processing method and device | |
CN110349126B (en) | Convolutional neural network-based marked steel plate surface defect detection method | |
CN111325721A (en) | Gas leakage detection method and system based on infrared thermal imaging | |
US8331650B2 (en) | Methods, systems and apparatus for defect detection | |
CN111047568B (en) | Method and system for detecting and identifying steam leakage defect | |
CN111126174A (en) | Visual detection method for robot to grab parts | |
CN110610150B (en) | Tracking method, device, computing equipment and medium of target moving object | |
JP6899189B2 (en) | Systems and methods for efficiently scoring probes in images with a vision system | |
CN112037203A (en) | Side surface defect detection method and system based on complex workpiece outer contour registration | |
CN112734761B (en) | Industrial product image boundary contour extraction method | |
CN108537787B (en) | Quality judgment method for face image | |
CN117372432B (en) | Electronic cigarette surface defect detection method and system based on image segmentation | |
CN110189375A (en) | A kind of images steganalysis method based on monocular vision measurement | |
CN114549441B (en) | Straw defect detection method based on image processing | |
CN111539927B (en) | Detection method of automobile plastic assembly fastening buckle missing detection device | |
CN103473776A (en) | Method and system for comparing image definition and automatic focusing control method | |
CN108109154A (en) | A kind of new positioning of workpiece and data capture method | |
CN110047063A (en) | A kind of detection method that material is fallen, device, equipment and storage medium | |
CN114612410A (en) | Novel clothing detects device | |
CN112419261A (en) | Visual acquisition method and device with abnormal point removing function | |
CN109671084B (en) | Method for measuring shape of workpiece | |
CN112613456A (en) | Small target detection method based on multi-frame differential image accumulation | |
CN117218054A (en) | Space target detection method under fixed star background | |
CN116524269A (en) | Visual recognition detection system | |
CN111473944B (en) | PIV data correction method and device for observing complex wall surface in flow field |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20171110 |