CN102436576A - Multi-scale self-adaptive high-efficiency target image identification method based on multi-level structure - Google Patents
Multi-scale self-adaptive high-efficiency target image identification method based on multi-level structure Download PDFInfo
- Publication number
- CN102436576A CN102436576A CN2011103215618A CN201110321561A CN102436576A CN 102436576 A CN102436576 A CN 102436576A CN 2011103215618 A CN2011103215618 A CN 2011103215618A CN 201110321561 A CN201110321561 A CN 201110321561A CN 102436576 A CN102436576 A CN 102436576A
- Authority
- CN
- China
- Prior art keywords
- thumbnail
- target image
- image
- far infrared
- layer
- 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
Abstract
The invention relates to a multi-scale self-adaptive high-efficiency target image identification method based on a multi-level structure. The method comprises the steps of acquiring a target image; performing multi-layer zooming, decomposition and display on the acquired original image and identifying the target image. By the method, the problems of instability of an identification system, low corresponding rate of the relationship between the identification accuracy rate and the identification speed, low identification rate and the like caused by factors such as size change of the image and the like in image identification are completely solved. In addition, the multi-scale self-adaptive high-efficiency target image identification method based on the multi-level structure has high identification speed.
Description
Technical field
The present invention relates to a kind of object features image-recognizing method, especially relate to the efficient target image recognition methods of a kind of multi-scale self-adaptive based on the multilayer level structure.
Background technology
How from image, can identify specific objective as human eye, for example: desk, automobile etc. are human dream and pursuits always.Image object identification is the major issue of a research of artificial intelligence field, is solving automated production and detection, productive life the very corn of a subject methods such as intelligent image analysis and retrieval.
Present stage, target accurately location is the gordian technique that identifying information is handled, and is widely used in the systems such as recognition of face, man-machine interaction, Intelligent Human-Machine Interface.The accurate location of eyes also is one and has challenging problem under the complex background.Because factors such as illumination, size, attitude, plane rotation, picture quality bring the complicated variation for the eyes outward appearance, blocking etc. of reflective, the hair of the switching of eyes, glasses and picture frame also brings a lot of difficulties to the accurate location of eyes except being for this; Particularly under the situation of eyes closed, eyebrow and thick picture frame all can bring larger interference to eye location.
The eyes that propose are at present accurately located the method that main stream approach is based on heuristic rule.These class methods mainly are to formulate locating rule according to the priori of eyes.These prioris comprise organ distributed knowledge, shape knowledge, color knowledge, physical characteristics etc.It is relatively poor that these class methods generally adapt to the extraneous ability that changes, one or more variations that often can only treatment of organs, and stability and precision and requirement of actual application also have gap.Causing the reason of this phenomenon mainly is the local appearance that they have only considered organ, and does not consider the restriction relation between organ and adjacent domain or organ.When there be the object similar with the target organ outward appearance in face, will influence like this to positioning belt.Outward appearance during such as eyes closed and eyebrow, thick picture frame are very similar.So the global characteristics of taking all factors into consideration the organ local feature and can expressing this restriction relation could obtain the more locating effect of robust.
Along with development of technology, in Target Recognition, how rapidly and efficiently the extraction target signature and to set up corresponding identification system be an important and crucial problem in the Target Recognition.Template matches (template matching) is a kind of important method in the Target Recognition Algorithms commonly used at present.The masterplate coupling comes target is discerned to the method that image carries out global registration through adopting masterplate.What have is a little that algorithm is simple, realizes easily.Shortcoming is that the robustness (Robustness) of algorithm is poor especially; Size when image; When directions etc. had conversion, the non-constant of matching effect was though there is the people to propose to adopt the method for the masterplate of a plurality of different size and Orientations to solve this type of problem; But this solution has increased the time complexity of algorithm greatly, and making becomes in the middle of practical application hardly maybe.David Marr; It is the method for other a kind of Target Recognition that Cannon etc. have proposed different edge extracting (edge detection); It had is a little the geometric properties of response diagram picture to a certain extent, has certain adaptability to light and shade conversion etc. is arranged.Recently David lowe is at " Objectrecognitionfrom local scale-invariant features ", and " SIFT " algorithm that proposes in the documents such as " Distinctiveimage features from scale-invariant keypoints " obtains increasing attention and application in Target Recognition.The SIFT algorithm carries out filtering through adopting Gaussian filter bank to image, extracts the method for scale-invaraint characteristic then and discerns.This method is to specific objective, for example: the books that certain is specific, the identification aspect of building etc. has obtained effect preferably.Yet the SIFT algorithm is to one type object identification, and for example: it is unsatisfactory from image, to pick out effects such as all desks and automobile, is the combination of serial of methods in addition on the SIFT algorithm essence, and the time complexity of algorithm is very high.And in document " Object recognition from local scale-invariant features ", author David lowe mentions, and the SIFT algorithm just meets the optic nerve principle of human body to a certain extent.Because variations such as the image object that will discern of institute possibly exist between size, direction, illumination, deformation, class, image recognition is a very complicated problems.Generally speaking, because Target Recognition the complex nature of the problem, existing method can well not felt above problem aspect the identification accuracy of method and the speed.Exist especially:
(1) size of image, direction, illumination, when individualities etc. are vicissitudinous, the recognition objective of robustness how.
(2) how can the quick identification target in rational time range.
Deficiency or defective to present recognition methods; The present invention proposes a kind of brand-new multiple dimensioned efficient target identification method with hierarchical structure: mainly comprise following which floor: at first; With the analysis tool (wavelet analysis or pyramid multiscale analysis method) of Multi-scale, image is carried out Filtering Processing, obtain the decomposition texture of a level; Self-adaptation is chosen matching layer and is mated; Obtain the secondary characteristics of image,, can carry out high efficiency Target Recognition based on the validity feature of this extraction.
Summary of the invention
The objective of the invention is to image identification system deficiency of the prior art and defective, a kind of effectively image-recognizing method rapidly and efficiently is provided, specifically is the efficient target image recognition methods of a kind of multi-scale self-adaptive based on the multilayer level structure.Multiple dimensioned (multi-scale) image object recognition methods that has concrete hierarchical structure (Hierarchical) through employing; Solved more comprehensively in image recognition that recognition system is stable inadequately because the factors such as size conversion of image cause, concerned problems such as corresponding rate is lower, discrimination is not high between recognition accuracy and recognition speed, the efficient target image recognition methods of the multi-scale self-adaptive based on the multilayer level structure of the present invention simultaneously also has very high recognition speed.
For realizing above-mentioned purpose, the present invention realizes through following technical scheme:
The efficient target image recognition methods of a kind of multi-scale self-adaptive based on the multilayer level structure of the present invention is characterized in that, should may further comprise the steps successively based on the multiple dimensioned efficient target image recognition methods of multilayer level structure:
Step 1, collection target image; The acquisition method of said target image is following:
At first, on the far infrared digital simulation field camera of the subsidiary image pick-up card of a plurality of outputs, be separately installed with initiatively far infrared light source; Wherein, Select the far red light spectral coverage of whole spectrum medium wavelength between 9um and 11um as active far infrared light source, initiatively the far infrared light source is made up of the N of wavelength between a 9.5um and 10.5um LED, and wherein N is a natural number; N ∈ [26,36]; When the far infrared digital simulation field camera of said active far infrared light source of combination and the subsidiary image pick-up card of said output, with the coaxial arrangement of far infrared digital simulation field camera of LED and the subsidiary image pick-up card of said output; The be del shape of said light emitting diode in the far infrared digital simulation field camera plane of the subsidiary image pick-up card of said output evenly put, with the light source stepless action in target image;
Collection for target image; Use the target site in the visible light digital simulation field camera alignment image of the subsidiary image pick-up card of a said output to gather; And the visible light digital simulation field camera that is connected the subsidiary image pick-up card of above-mentioned output on the same image pick-up card and the far infrared digital simulation field camera of the subsidiary image pick-up card of above-mentioned output being carried out synchronous acquisition control through image capture software, IMAQ has simultaneity;
The original image multilayer convergent-divergent of step 2, collection decomposes and shows; The original image convergent-divergent decomposition display packing of said collection is following: generate the target image thumbnail data according to the IMAQ data in the step 1, the thumbnail data data type is the two-dimensional array that and display device show lattice match; Generate corresponding thumbnail according to said target image thumbnail data then, this is generated corresponding thumbnail, carry out turriform anisotropic filter group conversion more than 5 layers; Again the thumbnail of said bank of filters conversion is arranged layering again and sort out, the thumbnail layer that the del top in the above-mentioned steps one is corresponding is classified as S
1Layer, the corresponding thumbnail layer in del bottom in the step 1 is classified as S
nLayer, said S
1Layer and S
nContain several layers, its S between the layer
nIn n be natural number, n ∈ [34,39]; Wherein, adjacent two thumbnail interlayer thumbnail image scale size ratios are 1～2: 1, and the thumbnail image of all thumbnail layers forms pyramid;
Step 3, target image identification; Said target image recognition methods is following: all the thumbnail layers in the pyramid in the above-mentioned steps two adopt linear SVMs to classify; Carry out the branch time-like; Choose 2/5ths to 3/5ths thumbnail layer as training set; Remaining thumbnail layer carries out the multiple dimensioned efficient target image identification of multilayer level structure then as test set as the standard of test accuracy through the equilibrium point on the demonstration curve that generates in the said linear SVMs.
As optimized technical scheme:
Adjacent two thumbnail interlayer thumbnail image scale size ratios in the above-mentioned steps two are 1.414: 1.
S in the above-mentioned steps two
1Layer and S
nContain 36 layers between the layer.
The thumbnail corresponding to this generation in the above-mentioned steps two carries out the conversion of 16 layers of turriform anisotropic filter group.
The thumbnail layer of choosing half in the above-mentioned steps three is as training set.
The far red light spectral coverage that the whole spectrum medium wavelength of selection in the above-mentioned steps one is 10um is as active far infrared light source.
Active far infrared light source in the above-mentioned steps one is that N the LED of 10.5um formed by wavelength.
The present invention is through researching and analysing the relative merits of image identification system in the past; Various challenges and problem that analysis image identification is faced; Take all factors into consideration recognition accuracy and the relation between recognition speed in image recognition; Multiple dimensioned (multi-scale) image object recognition methods that has concrete hierarchical structure (Hierarchical) through employing; More comprehensively solved in image recognition because problem such as the factors such as size conversion of image cause recognition system stable inadequately, and discrimination is not high, the method has very high recognition speed simultaneously.
Simultaneously, the system that whole recognition methods of the present invention constitutes adopts the Hierarchical hierarchy that meets human visual nervous system (Visual cortex), can also be by S1, and C1, S2, C2 four big layers constitute.The method of every big layer is described as follows:
S1: in this layer, the method that can adopt Multi-scale to analyze is decomposed image, then the image that decomposes is carried out obtaining the Multi-resolution processing scheme of pyramid from group.
After adopting the method for Multi-scale to decompose, can obtain the structure of a pyramid to image.Usually, in the structure that this kind method obtains, the image tool has plenty of the decay of 2: 1 sizes, and promptly the image size is 2: 1 between adjacent two layers.For make image layer and layer directly a size is arbitrary proportion, for example 1.414: 1, and traditional method adopts the method for interpolation to obtain intermediate image, have distortion to a certain degree but the method through interpolation obtains image, computing velocity is slow simultaneously.
At this, we propose a kind ofly in advance image to be carried out convergent-divergent (rescale), and then carry out multi-scale and decompose (wavelet decomposition etc.), then the image after decomposing are arranged (Reorder) again, obtain the S1 layer of image then.So just can obtain any band, the size between band is the method for the structure of arbitrary proportion.
The C1:C1 layer is main relatively and the complex cell (Complex cell) in the optic nerve (visual cortex).Through to merger many between adjacent layer, obtain the c1 layer of image.
The S2:S2 layer is equivalent to V2 and the simple cell in the v4 zone in the optic nerve (Visual Cortex).And, can carry out adaptively choosing different band and mating at this layer.In the middle of this layer, a selected k characteristic and former masterplate characteristic are carried out the coupling of following formula,
C2: among this layer, adopt the operation of global maximum,, obtain a corresponding with it eigenwert to each masterplate characteristic.The characteristic that obtains is the biological characteristic that meets human body optic nerve system.
Classification: adopt sorter that the eigenwert of being obtained is classified then, realize Target Recognition.
Recognition methods of the present invention has above through adopting the multiple dimensioned structure of level; Having avoided traditional template matches target is the method that needs method the template of a plurality of different scales; And through the adaptive method of choosing coupling level (Band); Realized the high-level efficiency of coupling; And whole Target Recognition structure matches with people's visual system to a certain extent, has more comprehensively solved the robustness of the recognizer that in the Target Recognition field, exists and the problem of the contradiction between the algorithm time complexity, and experimental result shows that this method has good recognition accuracy and has higher recognition speed simultaneously.
Embodiment
Below in conjunction with embodiment, further set forth the present invention.
Embodiment 1:
The efficient target image recognition methods of a kind of multi-scale self-adaptive based on the multilayer level structure is characterized in that, should may further comprise the steps successively based on the multiple dimensioned efficient target image recognition methods of multilayer level structure:
Step 1, collection target image; The acquisition method of said target image is following:
At first, on the far infrared digital simulation field camera of the subsidiary image pick-up card of a plurality of outputs, be separately installed with initiatively far infrared light source; Wherein, Select the far red light spectral coverage of whole spectrum medium wavelength between 9um and 11um as active far infrared light source, initiatively the far infrared light source is made up of the N of wavelength between a 9.5um and 10.5um LED, and wherein N is a natural number; N ∈ [26,36]; When the far infrared digital simulation field camera of said active far infrared light source of combination and the subsidiary image pick-up card of said output, with the coaxial arrangement of far infrared digital simulation field camera of LED and the subsidiary image pick-up card of said output; The be del shape of said light emitting diode in the far infrared digital simulation field camera plane of the subsidiary image pick-up card of said output evenly put, with the light source stepless action in target image;
Collection for target image; Use the target site in the visible light digital simulation field camera alignment image of the subsidiary image pick-up card of a said output to gather; And the visible light digital simulation field camera that is connected the subsidiary image pick-up card of above-mentioned output on the same image pick-up card and the far infrared digital simulation field camera of the subsidiary image pick-up card of above-mentioned output being carried out synchronous acquisition control through image capture software, IMAQ has simultaneity;
The original image multilayer convergent-divergent of step 2, collection decomposes and shows; The original image convergent-divergent decomposition display packing of said collection is following: generate the target image thumbnail data according to the IMAQ data in the step 1, the thumbnail data data type is the two-dimensional array that and display device show lattice match; Generate corresponding thumbnail according to said target image thumbnail data then, this is generated corresponding thumbnail, carry out turriform anisotropic filter group conversion more than 16 layers; Again the thumbnail of said bank of filters conversion is arranged layering again and sort out, the thumbnail layer that the del top in the above-mentioned steps one is corresponding is classified as S
1Layer, the corresponding thumbnail layer in del bottom in the step 1 is classified as S
nLayer, said S
1Layer and S
nContain several layers, its S between the layer
nIn n be natural number, n ∈ [34,36]; Wherein, adjacent two thumbnail interlayer thumbnail image scale size ratios are 1: 1, and the thumbnail image of all thumbnail layers forms pyramid;
Step 3, target image identification; Said target image recognition methods is following: all the thumbnail layers in the pyramid in the above-mentioned steps two adopt linear SVMs to classify; Carry out the branch time-like; Choose 2/5ths thumbnail layer as training set; Remaining thumbnail layer carries out the multiple dimensioned efficient target image identification of multilayer level structure then as test set as the standard of test accuracy through the equilibrium point on the demonstration curve that generates in the said linear SVMs.
Embodiment 2:
The efficient target image recognition methods of a kind of multi-scale self-adaptive based on the multilayer level structure is characterized in that, should may further comprise the steps successively based on the multiple dimensioned efficient target image recognition methods of multilayer level structure:
Step 1, collection target image; The acquisition method of said target image is following:
At first, on the far infrared digital simulation field camera of the subsidiary image pick-up card of a plurality of outputs, be separately installed with initiatively far infrared light source; Wherein, Select the far red light spectral coverage of whole spectrum medium wavelength between 9um and 11um as active far infrared light source, initiatively the far infrared light source is made up of the N of wavelength between a 9.5um and 10.5um LED, and wherein N is a natural number; N ∈ [26,36]; When the far infrared digital simulation field camera of said active far infrared light source of combination and the subsidiary image pick-up card of said output, with the coaxial arrangement of far infrared digital simulation field camera of LED and the subsidiary image pick-up card of said output; The be del shape of said light emitting diode in the far infrared digital simulation field camera plane of the subsidiary image pick-up card of said output evenly put, with the light source stepless action in target image;
Collection for target image; Use the target site in the visible light digital simulation field camera alignment image of the subsidiary image pick-up card of a said output to gather; And the visible light digital simulation field camera that is connected the subsidiary image pick-up card of above-mentioned output on the same image pick-up card and the far infrared digital simulation field camera of the subsidiary image pick-up card of above-mentioned output being carried out synchronous acquisition control through image capture software, IMAQ has simultaneity;
The original image multilayer convergent-divergent of step 2, collection decomposes and shows; The original image convergent-divergent decomposition display packing of said collection is following: generate the target image thumbnail data according to the IMAQ data in the step 1, the thumbnail data data type is the two-dimensional array that and display device show lattice match; Generate corresponding thumbnail according to said target image thumbnail data then, this is generated corresponding thumbnail, carry out turriform anisotropic filter group conversion more than 16 layers; Again the thumbnail of said bank of filters conversion is arranged layering again and sort out, the thumbnail layer that the del top in the above-mentioned steps one is corresponding is classified as S
1Layer, the corresponding thumbnail layer in del bottom in the step 1 is classified as S
nLayer, said S
1Layer and S
nContain several layers, its S between the layer
nIn n be natural number, n ∈ [36,39]; Wherein, adjacent two thumbnail interlayer thumbnail image scale size ratios are 1.414: 1, and the thumbnail image of all thumbnail layers forms pyramid;
Step 3, target image identification; Said target image recognition methods is following: all the thumbnail layers in the pyramid in the above-mentioned steps two adopt linear SVMs to classify; Carry out the branch time-like; Choose half the thumbnail layer as training set; Remaining thumbnail layer carries out the multiple dimensioned efficient target image identification of multilayer level structure then as test set as the standard of test accuracy through the equilibrium point on the demonstration curve that generates in the said linear SVMs.
Embodiment 3:
The efficient target image recognition methods of a kind of multi-scale self-adaptive based on the multilayer level structure is characterized in that, should may further comprise the steps successively based on the multiple dimensioned efficient target image recognition methods of multilayer level structure:
Step 1, collection target image; The acquisition method of said target image is following:
At first, on the far infrared digital simulation field camera of the subsidiary image pick-up card of a plurality of outputs, be separately installed with initiatively far infrared light source; Wherein, select whole spectrum medium wavelength be far red light spectral coverage between the 10um as active far infrared light source, initiatively the far infrared light source is that N LED between the 10.5um formed by wavelength, wherein N is a natural number, N ∈ [26,36]; When the far infrared digital simulation field camera of said active far infrared light source of combination and the subsidiary image pick-up card of said output, with the coaxial arrangement of far infrared digital simulation field camera of LED and the subsidiary image pick-up card of said output; The be del shape of said light emitting diode in the far infrared digital simulation field camera plane of the subsidiary image pick-up card of said output evenly put, with the light source stepless action in target image;
Collection for target image; Use the target site in the visible light digital simulation field camera alignment image of the subsidiary image pick-up card of a said output to gather; And the visible light digital simulation field camera that is connected the subsidiary image pick-up card of above-mentioned output on the same image pick-up card and the far infrared digital simulation field camera of the subsidiary image pick-up card of above-mentioned output being carried out synchronous acquisition control through image capture software, IMAQ has simultaneity;
The original image multilayer convergent-divergent of step 2, collection decomposes and shows; The original image convergent-divergent decomposition display packing of said collection is following: generate the target image thumbnail data according to the IMAQ data in the step 1, the thumbnail data data type is the two-dimensional array that and display device show lattice match; Generate corresponding thumbnail according to said target image thumbnail data then, this is generated corresponding thumbnail, carry out turriform anisotropic filter group conversion more than 8 layers; Again the thumbnail of said bank of filters conversion is arranged layering again and sort out, the thumbnail layer that the del top in the above-mentioned steps one is corresponding is classified as S
1Layer, the corresponding thumbnail layer in del bottom in the step 1 is classified as S
nLayer, said S
1Layer and S
nContain several layers, its S between the layer
nIn n be natural number, n ∈ [36,39]; Wherein, adjacent two thumbnail interlayer thumbnail image scale size ratios are 1.414: 1, and the thumbnail image of all thumbnail layers forms pyramid;
Step 3, target image identification; Said target image recognition methods is following: all the thumbnail layers in the pyramid in the above-mentioned steps two adopt linear SVMs to classify; Carry out the branch time-like; Choose 3/5ths thumbnail layer as training set; Remaining thumbnail layer carries out the multiple dimensioned efficient target image identification of multilayer level structure then as test set as the standard of test accuracy through the equilibrium point on the demonstration curve that generates in the said linear SVMs.
Embodiment 4:
Can adopt equipment such as camera and video camera to obtain required image, then the image that is obtained discerned.
For method is carried out objective appraisal, adopted standard database that method is tested, database comprises 186 leaves, 1155 automobiles, 450 people's faces, 1074 Zhang Fei's machines, 826 motorcycles and 900 background pictures constitute.At the S1 layer, we adopt 4 band, and each band has 2 scale, and the scale between two adjacent scale differs 1.414, then according to recombinating like the mode of figure Fig.2.Carrying out the branch time-like; The characteristic of choosing 1000 C2 layers is as test feature; Adopt linear SVMs (Support Vector Machine SVM) classifies. carry out the branch time-like, we choose half the image as training set; Half the in addition as test set, seek equilibrium point on the RoC curve then as the standard of test accuracy.
When testing, the hardware environment that is adopted is: Intel Core2, and 6600 2.4GHz CPU, 3.24GB memory, 32bit Matlab 2007a,
We can find out from above-mentioned test, and after the employing new method, the preparation rate of identification all has lifting in various degree, is not targets such as good leaf and automobile to former recognition effect particularly, and the accuracy rate of identification has the raising of significance degree.And all identification of targets degree of readiness have all been reached the high discrimination more than 96%, satisfy most of identification personage's requirement.
Another advantage of algorithm is that recognition speed is very fast; The above-mentioned about 5000 pictures storehouses of data that comprise 5 class targets are tested; Adopting the matlab platform and do not carrying out under the situation of hardware optimization; The processing time of average every pictures was less than for 2 seconds, and total system is expected to reach the requirement of real time processing system, at a high speed shape library and video was implemented to handle.Following table is the average handling time of every kind of target:
| The target classification | Leaf | Automobile | People's face | Aircraft | Motorcycle |
| Time | 1.8 second | 1.5 second | 1.9 second | 1.6 second | 1.8 second |
In average every pictures processing time of last table, be total to about 5000 pictures, the very potential application requirements that reaches.
The present invention is not limited to top explanation and embodiment.On the contrary, be intended to extensively be suitable in the determined boundary of the described below claim of the present invention.
Claims (9)
1. the efficient target image recognition methods of the multi-scale self-adaptive based on the multilayer level structure is characterized in that, should may further comprise the steps successively based on the multiple dimensioned efficient target image recognition methods of multilayer level structure:
Step 1, collection target image; The acquisition method of said target image is following:
At first, on the far infrared digital simulation field camera of the subsidiary image pick-up card of a plurality of outputs, be separately installed with initiatively far infrared light source; Wherein, Select the far red light spectral coverage of whole spectrum medium wavelength between 9um and 11um as active far infrared light source, initiatively the far infrared light source is made up of the N of wavelength between a 9.5um and 10.5um LED, and wherein N is a natural number; N ∈ [26,36]; When the far infrared digital simulation field camera of said active far infrared light source of combination and the subsidiary image pick-up card of said output, with the coaxial arrangement of far infrared digital simulation field camera of LED and the subsidiary image pick-up card of said output; The be del shape of said light emitting diode in the far infrared digital simulation field camera plane of the subsidiary image pick-up card of said output evenly put, with the light source stepless action in target image;
Collection for target image; Use the target site in the visible light digital simulation field camera alignment image of the subsidiary image pick-up card of a said output to gather; And the visible light digital simulation field camera that is connected the subsidiary image pick-up card of above-mentioned output on the same image pick-up card and the far infrared digital simulation field camera of the subsidiary image pick-up card of above-mentioned output being carried out synchronous acquisition control through image capture software, IMAQ has simultaneity;
The original image multilayer convergent-divergent of step 2, collection decomposes and shows; The original image convergent-divergent decomposition display packing of said collection is following: generate the target image thumbnail data according to the IMAQ data in the step 1, the thumbnail data data type is the two-dimensional array that and display device show lattice match; Generate corresponding thumbnail according to said target image thumbnail data then, this is generated corresponding thumbnail, carry out turriform anisotropic filter group conversion more than 5 layers; Again the thumbnail of said bank of filters conversion is arranged layering again and sort out, the thumbnail layer that the del top in the above-mentioned steps one is corresponding is classified as S
1Layer, the corresponding thumbnail layer in del bottom in the step 1 is classified as S
nLayer, said S
1Layer and S
nContain several layers, its S between the layer
nIn n be natural number, n ∈ [34,39]; Wherein, adjacent two thumbnail interlayer thumbnail image scale size ratios are 1～2: 1, and the thumbnail image of all thumbnail layers forms pyramid;
Step 3, target image identification; Said target image recognition methods is following: all the thumbnail layers in the pyramid in the above-mentioned steps two adopt linear SVMs to classify; Carry out the branch time-like; Choose 2/5ths to 3/5ths thumbnail layer as training set; Remaining thumbnail layer carries out the multiple dimensioned efficient target image identification of multilayer level structure then as test set as the standard of test accuracy through the equilibrium point on the demonstration curve that generates in the said linear SVMs.
2. the multiple dimensioned efficient target image recognition methods based on the multilayer level structure according to claim 1 is characterized in that: the adjacent two thumbnail interlayer thumbnail image scale size ratios in the above-mentioned steps two are 1.414: 1.
3. the multiple dimensioned efficient target image recognition methods based on the multilayer level structure according to claim 1 is characterized in that: the S in the above-mentioned steps two
1Layer and S
nContain 36 layers between the layer.
4. the multiple dimensioned efficient target image recognition methods based on the multilayer level structure according to claim 1 is characterized in that: the thumbnail corresponding to this generation in the above-mentioned steps two, carry out the conversion of 16 layers of turriform anisotropic filter group.
5. the multiple dimensioned efficient target image recognition methods based on the multilayer level structure according to claim 1 is characterized in that: the thumbnail layer of choosing half in the above-mentioned steps three is as training set.
6. the multiple dimensioned efficient target image recognition methods based on the multilayer level structure according to claim 1 is characterized in that: the far red light spectral coverage that the whole spectrum medium wavelength of the selection in the above-mentioned steps one is 10um is as active far infrared light source.
7. the multiple dimensioned efficient target image recognition methods based on the multilayer level structure according to claim 1 is characterized in that: the active far infrared light source in the above-mentioned steps one is that N the LED of 10.5um formed by wavelength.
8. the multiple dimensioned efficient target image recognition methods based on the multilayer level structure according to claim 2 is characterized in that: the active far infrared light source in the above-mentioned steps one is that N the LED of 10.5um formed by wavelength.
9. according to claim 3 or 9 described multiple dimensioned efficient target image recognition methodss based on the multilayer level structure, it is characterized in that: the active far infrared light source in the above-mentioned steps one is that N the LED of 10.5um formed by wavelength.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110321561.8A CN102436576B (en) | 2011-10-21 | 2011-10-21 | Multi-scale self-adaptive high-efficiency target image identification method based on multi-level structure |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110321561.8A CN102436576B (en) | 2011-10-21 | 2011-10-21 | Multi-scale self-adaptive high-efficiency target image identification method based on multi-level structure |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102436576A true CN102436576A (en) | 2012-05-02 |
| CN102436576B CN102436576B (en) | 2013-11-06 |
Family
ID=45984631
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201110321561.8A Expired - Fee Related CN102436576B (en) | 2011-10-21 | 2011-10-21 | Multi-scale self-adaptive high-efficiency target image identification method based on multi-level structure |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102436576B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107967454A (en) * | 2017-11-24 | 2018-04-27 | 武汉理工大学 | Take the two-way convolutional neural networks Classification in Remote Sensing Image method of spatial neighborhood relation into account |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6389169B1 (en) * | 1998-06-08 | 2002-05-14 | Lawrence W. Stark | Intelligent systems and methods for processing image data based upon anticipated regions of visual interest |
| CN101236599A (en) * | 2007-12-29 | 2008-08-06 | 浙江工业大学 | Human face recognition detection device based on multi- video camera information integration |
| CN201927079U (en) * | 2011-03-07 | 2011-08-10 | 山东电力研究院 | Rapid real-time integration processing system for visible image and infrared image |
-
2011
- 2011-10-21 CN CN201110321561.8A patent/CN102436576B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6389169B1 (en) * | 1998-06-08 | 2002-05-14 | Lawrence W. Stark | Intelligent systems and methods for processing image data based upon anticipated regions of visual interest |
| CN101236599A (en) * | 2007-12-29 | 2008-08-06 | 浙江工业大学 | Human face recognition detection device based on multi- video camera information integration |
| CN201927079U (en) * | 2011-03-07 | 2011-08-10 | 山东电力研究院 | Rapid real-time integration processing system for visible image and infrared image |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107967454A (en) * | 2017-11-24 | 2018-04-27 | 武汉理工大学 | Take the two-way convolutional neural networks Classification in Remote Sensing Image method of spatial neighborhood relation into account |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102436576B (en) | 2013-11-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Pflug et al. | Ear biometrics: a survey of detection, feature extraction and recognition methods | |
| Mishra et al. | Region duplication forgery detection technique based on SURF and HAC | |
| Akhtar et al. | Automated plant disease analysis (APDA): performance comparison of machine learning techniques | |
| CN106446933B (en) | Multi-target detection method based on contextual information | |
| Bandeira et al. | Detection of sub-kilometer craters in high resolution planetary images using shape and texture features | |
| CN105718889B (en) | Based on GB (2D)2The face personal identification method of PCANet depth convolution model | |
| CN103646255A (en) | Face detection method based on Gabor characteristics and extreme learning machine | |
| Goel et al. | Are buildings only instances? Exploration in architectural style categories | |
| CN102508917B (en) | Multi-dimensional object robust high-speed retrieval and positioning method for some feature images | |
| Bouchaffra et al. | Structural hidden Markov models for biometrics: Fusion of face and fingerprint | |
| Villaruz | Deep convolutional neural network feature extraction for berry trees classification | |
| CN106408037A (en) | Image recognition method and apparatus | |
| Mughees et al. | Efficient deep auto-encoder learning for the classification of hyperspectral images | |
| Siraj et al. | Flower image classification modeling using neural network | |
| Uba | Land use and land cover classification using deep learning techniques | |
| CN101964055B (en) | Visual perception mechansim simulation natural scene type identification method | |
| CN105069403A (en) | Three-dimensional ear recognition based on block statistic features and dictionary learning sparse representation classification | |
| CN108537266A (en) | A kind of cloth textured fault sorting technique of depth convolutional network | |
| CN102436576B (en) | Multi-scale self-adaptive high-efficiency target image identification method based on multi-level structure | |
| Abed et al. | Real-time color image classification based on deep learning network | |
| Bilang et al. | Cactaceae detection using MobileNet architecture | |
| Proença et al. | SHREC’15 Track: Retrieval of Oobjects captured with kinect one camera | |
| Zhao et al. | Gabor-modulated grouped separable convolutional network for hyperspectral image classification | |
| Yang et al. | Color image representation using invariant exponent moments | |
| Shrivastava et al. | Bridging the semantic gap with human perception based features for scene categorization |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| DD01 | Delivery of document by public notice |
Addressee: Hong Tao Document name: Notification that Application Deemed not to be Proposed |
|
| GR01 | Patent grant | ||
| DD01 | Delivery of document by public notice |
Addressee: Hong Tao Document name: Notification to Pay the Fees |
|
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20131106 Termination date: 20141021 |
|
| EXPY | Termination of patent right or utility model | ||
| DD01 | Delivery of document by public notice |
Addressee: Hong Tao Document name: Notification of Decision on Request for Restoration of Right |