CN103150904A - Bayonet vehicle image identification method based on image features - Google Patents
Bayonet vehicle image identification method based on image features Download PDFInfo
- Publication number
- CN103150904A CN103150904A CN2013100451514A CN201310045151A CN103150904A CN 103150904 A CN103150904 A CN 103150904A CN 2013100451514 A CN2013100451514 A CN 2013100451514A CN 201310045151 A CN201310045151 A CN 201310045151A CN 103150904 A CN103150904 A CN 103150904A
- Authority
- CN
- China
- Prior art keywords
- image
- vehicle
- template
- identified
- color
- 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
- 238000000034 method Methods 0.000 title claims abstract description 40
- 239000003086 colorant Substances 0.000 claims abstract description 13
- 230000000295 complement effect Effects 0.000 claims description 53
- 230000008447 perception Effects 0.000 claims description 15
- 238000000746 purification Methods 0.000 claims description 12
- 239000011159 matrix material Substances 0.000 claims description 5
- BQCADISMDOOEFD-UHFFFAOYSA-N Silver Chemical compound [Ag] BQCADISMDOOEFD-UHFFFAOYSA-N 0.000 claims description 3
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims description 3
- 229910052709 silver Inorganic materials 0.000 claims description 3
- 239000004332 silver Substances 0.000 claims description 3
- 230000007547 defect Effects 0.000 abstract 1
- 238000010586 diagram Methods 0.000 description 8
- 230000008878 coupling Effects 0.000 description 7
- 238000010168 coupling process Methods 0.000 description 7
- 238000005859 coupling reaction Methods 0.000 description 7
- 238000004040 coloring Methods 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 2
- 238000011403 purification operation Methods 0.000 description 2
- 241001270131 Agaricus moelleri Species 0.000 description 1
- 241001573881 Corolla Species 0.000 description 1
- 230000002411 adverse Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 230000013011 mating Effects 0.000 description 1
Images
Landscapes
- Image Analysis (AREA)
Abstract
The invention relates to the field of traffic image processing, in particular to a bayonet vehicle image identification method based on image features. The steps of bayonet vehicle image recognition method based on the image features includes that a template dada bank is established, the data bank stores template images which are shot with different style and different color vehicles, vehicle attribute data, vehicle body color data and vehicle scale invariant feature transform (SIFT) feature data are stored corresponding to each template image, wherein the SIFT feature data are obtained through data processing; vehicle recognition is conducted to to-be-identified images, color identification is conducted to the to-be-identified images, the template images which are fitted with the vehicle body colors are selected form the template data bank according to color identification results; the to-be-identified images are extracted the SIFT features and compared with the selected template images; the template images which are matched with the to-be-identified images are obtained; the vehicle attribute data are output, wherein the vehicle attribute data correspond to the matched template images. The bayonet vehicle image identification method based on the image features combines the vehicle color identification and the recognition based on a SIFT operator principle. Color information is added in the identification process. The defect that the SIFT operator discard the color information is overcome. Identification accuracy can be improved.
Description
Technical field
The present invention relates to the traffic image process field, more specifically, relate to a kind of bayonet socket vehicle image recognition methods based on characteristics of image.
Background technology
Vehicle identification system (Vehicle Recognition System, VRS) is an important component part of intelligent transportation system.The vehicle identification system of broad sense refers to identify the information such as the car plate, color, producer, model, special marking of vehicle.But the vehicle identification system of using at present substantially all only identify by car plate and is realized that testing vehicle register is definite, be mainly used in testing the speed in the interval and highway, parking lot access management in the middle of.Although car plate unique legal sign that is vehicle, there are the problems such as accuracy rate is not high, service condition is limited in the license plate recognition technology of processing based on image in application, especially for false-trademark, deck, block the licence plate situation and can't tackle.Therefore, it is very important other vehicle characteristics beyond car plate being identified.Current also have distinguishing the technology of the vehicle such as large, medium and small, and identify the technology of not carrying out the vehicle brand recognition by car, but classification is very rough, and the discrimination of color is not high yet.Still depend on workman's judgement and manual retrieval for the identification of concrete vehicle model, style and the search of specific vehicle, this work often needs to spend a large amount of human resources and time, and when searching for for mass image data, workload is unable to estimate.Therefore substitute with computing machine artificial, carry out the identification of vehicle model, style and more accurately color distinguish right form wrong normal necessary.
It is different that the object of identifying is pressed in main vehicle image recognition methods at present, can be divided three classes: Direct Recognition, vehicle local feature recognition and characteristics of image identification.
Direct Recognition.The perception information of these class methods Direct Recognition mankind to vehicle mainly comprises vehicle license and vehicle appearance information (as color, shape, size etc.).Although the former recognition result can be used as the law enforcement foundation, accuracy rate is limited; The latter has special requirement usually for photographed scene, and often needs prior demarcation, can't accurate Calculation.
The vehicle local feature.The vehicle local feature mainly comprises vehicle mark information and car face information.But the former only comprises man of carshop information, and the latter often causes being difficult to extracting and accurately definition because car face information is too complicated.
Characteristics of image identification.Characteristics of image does not generally represent the mankind to the perception information of vehicle, therefore to Characteristic of Image identification, often needs to have template image, identifies by the matching degree of calculating image to be identified and template image feature.The vehicle image feature mainly comprises Color Distribution Features (as histogram, color moment etc.) and image space local feature.For the former, different vehicle images may corresponding identical distribution characteristics; Latter is take SIFT, SURF, PCA-SIFT as representative, and computation complexity is higher.
Summary of the invention
Technical matters to be solved by this invention is to provide the high bayonet socket vehicle image recognition methods based on characteristics of image of a kind of recognition accuracy.
For solving the problems of the technologies described above, technical scheme of the present invention is as follows:
A kind of bayonet socket vehicle image recognition methods based on characteristics of image comprises the steps:
Set up template database:
Store the template image of the different style different colours vehicles of taking gained in template database, corresponding every template image store car attribute data, body color data in template database;
Every in template database template image is carried out data process, obtain vehicle SIFT characteristic and be stored in template database;
Treating recognition image identifies:
Treat recognition image and carry out color identification;
Choose the template image that meets body color from template database according to the color recognition result;
Utilize SIFT operator principle to treat recognition image and extract feature, and compare with selected template image, obtain the template image that is complementary with image to be identified;
The corresponding vehicle attribute data reading of the template image that will be complementary.
The present invention combines with vehicle color identification with based on the identification of SIFT operator principle, the template image data of selecting colouring information to be consistent under every kind of vehicle in template database according to the color recognition result, and carry out comparison based on SIFT operator principle, make the template image of image to be identified and corresponding color carry out matching operation, for identifying has added colouring information, overcome the shortcoming that the SIFT operator abandons colouring information, can improve the accuracy of identification; And corresponding each vehicle all stores corresponding vehicle attribute data in template database, and is meticulousr to the classification of vehicle, makes recognition result more concrete, and recognition accuracy is higher.
One of improve: every in template database template image is carried out data process and obtain the concrete steps that vehicle SIFT characteristic is stored in template database and be:
Image coordinate to vehicle license plate place in every template image is carried out mark;
Every template image is carried out determining of vehicle SIFT unique point, be about to be positioned in the masterplate image SIFT characteristic point information rejecting of car plate marked region, in reservation masterplate image, remaining SIFT characteristic information is as the vehicle SIFT characteristic of every template image.
The present invention gets rid of the unique point of license plate area in template image, make the follow-up interference of having got rid of the license plate area unique point when image to be identified and template image are mated, can greatly reduce the possibility that is identified as different automobile types, thereby improve the accuracy rate of identification.
Two of improvement: treating the concrete steps that recognition image carries out color identification is:
In advance body color is divided into green, yellow, red, blue, white and black six classes, wherein yellow comprises the yellow of Human Perception, orange and brown, redness comprises redness, pink colour and the purple of Human Perception, white comprises white, silver color, the light gray of Human Perception, and black comprises black, the Dark grey of Human Perception;
For green, yellow, red, blue, white five kinds of colors, in conjunction with r, g, b difference in twos, empirical value is set r, g, b, h, s, v value are marked certain scope, add up the ratio that green, yellow, red, blue, white five kinds of colored pixels points in vehicle body scope on image to be identified account for the interior pixel of this image vehicle body scope;
The shades of colour ratio for the treatment of recognition image in green, yellow, red, blue, white order judges, when the ratio of current color surpasses the empirical value of corresponding color, the vehicle body that judges image to be identified is current color, when image body color ratio to be identified does not all exceed the empirical value of green, yellow, red, blue, white five kinds of colors, the body color that judges this image to be identified is black, thereby obtains the color recognition result.
The present invention combines RGB and the HSV value of image and carries out the judgement of body color, and according to certain color sequences, body color is judged, makes the result of color identification more stable.
Three of improvement: the vehicle body scope of described image to be identified is got rid of vehicle window scope, the front face exhaust grille scope of car and car light scope.The present invention has rejected face exhaust grille scope and car light scope before vehicle window scope, car when body color is identified, can reduce the colors such as vehicle window, the front face exhaust grille of car and car light to the adverse effect of vehicle color judgement, and then improve the accuracy of body color judgement.
Four of improvement: the described SIFT of utilization operator principle is treated recognition image and is extracted feature, and compares with selected template image, and the concrete steps of obtaining the template image that is complementary with image to be identified are:
Treating recognition image utilizes SIFT operator principle to determine its SIFT characteristic;
The SIFT characteristic of the template image that SIFT characteristic and every of image to be identified is chosen is compared, and obtains the unique point that is complementary pair;
To calculating the matching degree of two images, determine the template image that is complementary with this image to be identified according to the unique point that is complementary according to matching degree.
Five of improvement: obtain the unique point that is complementary to the rear data purification step of also carrying out, specific as follows:
The input value of the right position mapping relations of the corresponding unique point that is complementary in every template image of choosing and image to be identified as the RANSAC algorithm, use the homography matrix of RANSAC method estimation image conversion, reject not the conforming unique point of meeting geometric pair, the unique point of obtaining reservation is to as the final unique point that is complementary pair.
The present invention uses the RANSAC algorithm to carry out data purification to matching result in order to improve the robustness of algorithm.
Six of improvement: obtain the unique point that is complementary and judge also that to rear whether the right logarithm of every template image of choosing be complementary unique point corresponding with image to be identified is greater than threshold value μ, if greater than executing data purification step, otherwise judge that directly two images match degree are 0.Due in the RANSAC algorithm, the unique point that the estimation of homography matrix is needed at least 4 pairs of couplings is in order to strengthen the stability of algorithm, after coupling finishes, just carry out the data purification operation at the right logarithm of the unique point that is complementary during greater than threshold value μ, threshold value μ suggestion value is 6.
Seven of improvement: the unique point that described basis is complementary is to calculating the matching degree of two images, and the concrete steps that determine the template image that is complementary with this image to be identified according to matching degree are:
The unique point that is complementary according to image to be identified and each template image of choosing is to computed image matching degree IMD=N/ N
0, wherein N represents the unique point logarithm that is complementary, N
0Sum for SIFT unique point in image to be identified;
Select maximal value IMD from all IMD values
max, with IMD
maxCompare judgement IMD with setting threshold λ
maxWhether greater than λ, if judge this IMD
maxCorresponding template image and images match to be identified success, otherwise in the judge templet database less than the template image that is complementary with this image to be identified.
The present invention adopts the relative value of unique point logarithm of coupling as images match degree IMD, can weigh the matching degree between image more objectively, can improve the accuracy of mating between image.
Nine of improvement: described vehicle attribute comprise vehicle brand, vehicle model and vehicle year money.
The present invention with vehicle brand, vehicle model and vehicle year money deposit in template database as the vehicle attribute data, realization is to vehicle sophisticated category more, when image to be identified finds the template image that is complementary in template database, can find corresponding vehicle attribute data reading from template database, make recognition result more concrete, greatly improved the precision of identification.
Compared with prior art, the beneficial effect of technical solution of the present invention is:
(1) the present invention combines with vehicle color identification with based on the identification of SIFT operator principle, for identifying has added colouring information, has overcome the shortcoming that the SIFT operator abandons colouring information, can improve the accuracy of identification.
(2) the present invention is identifying RGB and the HSV value that combines image to vehicle color, and judges according to certain color sequences, makes recognition result more reliable and more stable.
(3) the present invention's corresponding each vehicle in template database all stores corresponding vehicle attribute data, when image to be identified matches corresponding template image in template database, can export corresponding vehicle attribute data, make recognition result more concrete, recognition accuracy is higher.
(4) the present invention can reduce the color of vehicle window, the front face exhaust grille of car and car light to the impact of vehicle color judgement, can improve the accuracy rate of vehicle color identification.
(5) the present invention has got rid of the interference of license plate area unique point when identifying based on the vehicle of SIFT operator, can greatly reduce the possibility that is identified as different automobile types.
(6) the present invention by the similarity that definition images match degree IMD represents image to be identified and template image, by with an objective appraisal method, recognition result being described, can improve the accuracy rate of vehicle identification.
(7) the present invention by using the RANSAC algorithm to carry out data purification to matching result, can improve the robustness of algorithm, reduces the identification error rate.
Description of drawings
Fig. 1 is the process flow diagram of a kind of bayonet socket vehicle image recognition methods specific embodiment based on characteristics of image in the present invention.
Fig. 2 is the exemplary plot of the template image in template database.
Fig. 3 is the SIFT unique point result of calculation exemplary plot of a template image in template database.
Fig. 4 is image to be identified and SIFT examples of features thereof.
The schematic diagram of Fig. 5 for the image to be identified of Fig. 4 and white template image shown in Figure 2 are mated.
The result schematic diagram of Fig. 6 for the matching result of Fig. 5 is carried out data purification.
Fig. 7 is the process flow diagram of a kind of bayonet socket vehicle image recognition methods one preferred embodiment based on characteristics of image in the present invention.
Embodiment
Below in conjunction with drawings and Examples, technical scheme of the present invention is described further.
Embodiment 1
As shown in Figure 1, be the process flow diagram of a kind of bayonet socket vehicle image recognition methods specific embodiment based on characteristics of image in the present invention.Referring to as 1, the concrete steps of a kind of bayonet socket vehicle image recognition methods based on characteristics of image of this specific embodiment comprise:
Step S100: set up template database:
Step S101: the template image that stores the different style different colours vehicles of taking gained in template database, corresponding every template image store car attribute data, body color data in template database, wherein the vehicle attribute data can comprise vehicle brand, vehicle model and vehicle year money; As shown in Figure 2, be two template images in " Toyota, corolla, the 9th generation ";
Step S102: every in template database template image is carried out data process, obtain vehicle SIFT characteristic and be stored in template database, wherein the SIFT characteristic generally comprises the SIFT descriptor of SIFT unique point and each unique point;
Step S200: treat recognition image and carry out vehicle identification:
Step S201: treat recognition image and carry out color identification;
Step S202: choose the template image that meets body color according to the color recognition result from template database;
Step S203: utilize SIFT operator principle to treat recognition image and extract feature, and compare with selected template image, obtain the template image that is complementary with image to be identified;
Step S204: the corresponding vehicle attribute data reading of the template image that will be complementary.
This specific embodiment combines with vehicle color identification with based on the identification of SIFT operator principle, the template image data of selecting to meet color result according to the color recognition result in template database under every kind of vehicle carries out the comparison based on SIFT operator principle, find the matching template image in template database, and export corresponding vehicle attribute data according to the template image of coupling.
In step S102, for getting rid of vehicle license plate provincial characteristics point to the interference of images match, improve the accuracy rate of identification, this specific embodiment is in specific implementation process, in advance every template image in template database is carried out the delimitation of license plate area, and the SIFT feature within every template image license plate area is removed, concrete steps are as follows:
Step S1021: the image coordinate to vehicle license plate place in every template image is carried out mark;
Step S1022: every template image is carried out determining of vehicle SIFT unique point, be about to be positioned in the masterplate image SIFT characteristic point information rejecting of car plate marked region, in reservation masterplate image, remaining SIFT characteristic information is as the vehicle SIFT characteristic of every template image.If there is car plate in template image, the image coordinate at car plate place is carried out mark, the SIFT characteristic point information that is positioned at car plate mark coordinate range is rejected, and calculate the SIFT descriptor of all the other unique points, finally obtain the vehicle SIFT characteristic of every template image.As shown in Figure 3, be the SIFT feature calculation result of template image, wherein, illustrated arrow starting point representation feature point position, direction of arrow representation feature point principal direction, the direction of arrow represents to describe the submodule value.
In step S201, this specific embodiment is based on the difference in perception of human eye to color, factor in conjunction with illumination effect, carry out self-defined to body color, thereby according to the color of making by oneself, body color is judged, and the RGB and the HSV value that combine image carry out the judgement of body color, and according to certain color sequences, body color is judged, makes the result of color identification more stable.Concrete steps are as follows:
Step S2011: in advance body color is divided into green, yellow, red, blue, white and black six classes, wherein yellow comprises the yellow of Human Perception, orange and brown, redness comprises redness, pink colour and the purple of Human Perception, white comprises white, silver color, the light gray of Human Perception, and black comprises black, the Dark grey of Human Perception;
Step S2012: for green, yellow, red, blue, white five kinds of colors, in conjunction with r, g, b difference in twos, empirical value is set r, g, b, h, s, v value are marked certain scope; As shown in table 1 below:
The judgement of table 1 color
| Color | hThe value scope | sThe value scope | vThe value scope | r, g, bScope | Other conditions |
| Green | 70< h≤170 | s> ThreS | v> ThreV | diff(r,g,b)> ThreRGB | - |
| Yellow | 20< h≤70 | s> ThreS | v> ThreV | diff(r,g,b)> ThreRGB | - |
| Red | h≤ 20 or h>300 | s> ThreS | v> ThreV | diff(r,g,b)> ThreRGB | - |
| Blue | 170< h≤260 | s> ThreS | v> ThreV | diff(r,g,b)> ThreRGB | - |
| White | - | ? | ? | r, g, bAll greater than ThreWhite | ? |
| Black | - | - | - | - | Do not meet other above situations |
Referring to table 1, wherein
ThreS,
ThreV,
ThreRGB,
ThreWhiteBe respectively empirical value, the suggestion value is respectively 30,20,15,160, diff (
r,
g,
b) expression
r,
g,
bIn any two the value differences;
As shown in table 1, when its h value of a certain pixel in image to be identified is between 70＜h≤170, and the value of its s, v is respectively greater than ThreS, ThreV, and in r, g, b, the differences of any two values greater than ThreRGB, can judge that this treats that pixel is green.The judgement of other colors in like manner can get.
Step S2013: add up the ratio that green, yellow, red, blue, white five kinds of colored pixels points in vehicle body scope on image to be identified account for the interior pixel of this vehicle body scope; Wherein the vehicle body scope is got rid of vehicle window scope, the front face exhaust grille scope of car and car light scope;
Step S2014: the shades of colour ratio for the treatment of recognition image in green, yellow, red, blue, white order judges, when the ratio of current color surpasses the empirical value of corresponding color, the vehicle body that judges image to be identified is current color, when image body color ratio to be identified does not all exceed the empirical value of green, yellow, red, blue, white five kinds of colors, the body color that judges this image to be identified is black, thereby obtains the color recognition result.
In this specific embodiment, when treating recognition image and carrying out vehicle identification, determine the SIFT characteristic of vehicle image according to SIFT operator principle, color combining recognition result input template database is compared, and obtains at last the result of identifying.Therefore, the concrete steps of the step S203 of this specific embodiment are;
S2031: treat recognition image and utilize SIFT operator principle to determine its SIFT characteristic; As shown in Figure 4, be image to be identified and SIFT examples of features thereof.
S2032: the SIFT characteristic of the template image that SIFT characteristic and every of image to be identified is chosen is compared, and obtains the unique point that is complementary pair;
S2033: to calculating the matching degree of two images, determine the template image that is complementary with this image to be identified according to matching degree according to the unique point that is complementary.
Wherein, the unique point that is complementary in step S2032 is to representing the degree that it is complementary by Euclidean distance, and concrete steps are as follows:
S20321: preset threshold epsilon;
S20322: for the SIFT unique point P in image to be identified, calculate the Euclidean distance between the proper vector descriptor of all unique points and P in every template image of choosing, therefrom find minimum value and sub-minimum d1 and the d2 of this Euclidean distance, and record d1 and d2 corresponding SIFT unique point Q1 and Q2 in template image respectively;
S20323: if d2 is 0, make parameter
RatioBe 0; Otherwise calculate
S20324: with parameter
RatioCompare with threshold epsilon, when
RatioDuring less than threshold epsilon, the match is successful to judge SIFT unique point P in image to be identified and the unique point Q1 of template image, otherwise judgement P mates unsuccessful with Q1;
S20325: add up and record unique point that every template image of choosing and image to be identified be complementary pair.As shown in Figure 5, the image to be identified of Fig. 4 (color recognition result for white) and white template image shown in Figure 2 are mated, result as shown in Figure 5, the unique point that the Based on Feature Points that connects by straight line between two cars in diagram is complementary pair.
After the coupling of step S20324 finished, in order to improve the robustness of algorithm, this specific embodiment can also use the RANSAC algorithm to carry out data purification to matching result, and concrete steps are as follows:
The input value of the right position mapping relations of the unique point that is complementary corresponding in every template image of choosing and image to be identified as the RANSAC algorithm, use the homography matrix of RANSAC method estimation image conversion, reject not the conforming unique point of meeting geometric pair, the unique point of obtaining reservation is to as the final unique point that is complementary pair.As shown in Figure 6, for the matching result of Fig. 5 being carried out the result schematic diagram of data purification.
Due in the RANSAC algorithm, the estimation of homography matrix is needed at least the unique point of 4 pairs of couplings, so that strengthen the stability of algorithm, after coupling finishes, just carry out the data purification operation when advancing the right logarithm of the unique point that is complementary greater than the threshold value μ that sets.Specific as follows:
Obtain the unique point that is complementary and whether also judge the right logarithm of image to be identified the be complementary unique point corresponding with every template image of choosing greater than threshold value μ to rear, if greater than executing data purification step, otherwise judge that directly two images match degree are 0.Wherein threshold value μ suggestion is made as 6.
in this specific embodiment, consider that in actual match, an image to be identified may need to mate with many template images, the image to be identified of different colours also may mate from the template image of different gray scales, what have certain impact to the unique point logarithm that the match is successful at last and the template image feature is counted, therefore, in order further to improve the accuracy rate of identification, this specific embodiment represents two matching degrees between image by the relative value that adopts the match point logarithm as the images match degree, in the hope of more objectively the matching degree between image being weighed.Therefore, the unique point that in step S2033, basis is complementary is to calculating the matching degree of two images, and the concrete steps that determine the template image that is complementary with this image to be identified according to matching degree are as follows:
S20331: the unique point that is complementary according to image to be identified and each template image of choosing is to computed image matching degree IMD=N/ N
0, wherein N represents the unique point logarithm that is complementary, N
0Sum for SIFT unique point in image to be identified;
S20332: select maximal value IMD from all IMD values
max, with IMD
maxCompare judgement IMD with setting threshold λ
maxWhether greater than λ, if judge this IMD
maxCorresponding template image and images match to be identified success, otherwise in the judge templet database less than the template image that is complementary with this image to be identified.
Embodiment 2
As shown in Figure 7, be the process flow diagram of a kind of bayonet socket vehicle image recognition methods one preferred embodiment based on characteristics of image in the present invention.Referring to Fig. 7, the concrete steps of this preferred embodiment are as follows:
Set up template database: the template image of the different style vehicles of pre-stored different automobile types in template database, and with the vehicle attribute data of template image, body color, SIFT characteristic corresponding stored in template database; Wherein to every template image, at first it delimit the zone at car plate place, and the SIFT characteristic information that every template image is positioned at license plate area is deleted;
When needs are treated recognition image and carried out vehicle identification, first treat recognition image and carry out color identification, and can calculate simultaneously the SIFT feature of this image to be identified;
Image to be identified is input in template database compares, first select the template image of suitable color to carry out the SIFT characteristic matching according to its color recognition result in template database under different automobile types;
Matching result is carried out data purification;
Matching result after data purification is carried out IMD calculate, get template image corresponding to maximum IMD value, get the vehicle attribute data of this template image and export.
Claims (9)
1. the bayonet socket vehicle image recognition methods based on characteristics of image, is characterized in that, comprises the steps:
Set up template database:
Store the template image of the different style different colours vehicles of taking gained in template database, corresponding every template image store car attribute data, body color data in template database;
Every in template database template image is carried out data process, obtain vehicle SIFT characteristic and be stored in template database;
Treat recognition image and carry out vehicle identification:
Treat recognition image and carry out color identification;
Choose the template image that meets body color from template database according to the color recognition result;
Utilize SIFT operator principle to treat recognition image and extract feature, and compare with selected template image, obtain the template image that is complementary with image to be identified;
The corresponding vehicle attribute data reading of the template image that will be complementary.
2. the bayonet socket vehicle image recognition methods based on characteristics of image according to claim 1, it is characterized in that, every in template database template image is carried out data processes, obtain vehicle SIFT characteristic and the concrete steps that are stored in template database are:
Image coordinate to vehicle license plate place in every template image is carried out mark;
Every template image is carried out determining of vehicle SIFT unique point, be about to be positioned in the masterplate image SIFT characteristic point information rejecting of car plate marked region, in reservation masterplate image, remaining SIFT characteristic information is as the vehicle SIFT characteristic of every template image.
3. the bayonet socket vehicle image recognition methods based on characteristics of image according to claim 1, is characterized in that, treats the concrete steps that recognition image carries out color identification to be:
In advance body color is divided into green, yellow, red, blue, white and black six classes, wherein yellow comprises the yellow of Human Perception, orange and brown, redness comprises redness, pink colour and the purple of Human Perception, white comprises white, silver color, the light gray of Human Perception, and black comprises black, the Dark grey of Human Perception;
For green, yellow, red, blue, white five kinds of colors, in conjunction with r, g, b difference in twos, empirical value is set r, g, b, h, s, v value are marked certain scope;
Add up green, yellow, red, blue, white five kinds of colored pixels points in vehicle body scope on image to be identified and account for the ratio of the interior pixel of this image vehicle body scope;
The shades of colour ratio for the treatment of recognition image in green, yellow, red, blue, white order judges, when the ratio of current color surpasses the empirical value of corresponding color, the vehicle body that judges image to be identified is current color, when image body color ratio to be identified does not all exceed the empirical value of green, yellow, red, blue, white five kinds of colors, the body color that judges this image to be identified is black, thereby obtains the color recognition result.
4. the bayonet socket vehicle image recognition methods based on characteristics of image according to claim 3, is characterized in that, the vehicle body scope of described image to be identified is got rid of vehicle window scope, the front face exhaust grille scope of car and car light scope.
5. the bayonet socket vehicle image recognition methods based on characteristics of image according to claim 1, it is characterized in that, utilize SIFT operator principle to treat recognition image and extract feature, and compare with selected template image, the concrete steps of obtaining the template image that is complementary with image to be identified are:
Treating recognition image utilizes SIFT operator principle to determine its SIFT characteristic;
The SIFT characteristic of the template image that SIFT characteristic and every of image to be identified is chosen is compared, and obtains the unique point that is complementary pair;
To calculating the matching degree of two images, determine the template image that is complementary with this image to be identified according to the unique point that is complementary according to matching degree.
6. the bayonet socket vehicle image recognition methods based on characteristics of image according to claim 5, is characterized in that, obtains the unique point that is complementary to the rear data purification step of also carrying out, and is specific as follows:
The input value of the right position mapping relations of the corresponding unique point that is complementary in every template image of choosing and image to be identified as the RANSAC algorithm, use the homography matrix of RANSAC method estimation image conversion, reject not the conforming unique point of meeting geometric pair, the unique point of obtaining reservation is to as the final unique point that is complementary pair.
7. the bayonet socket vehicle image recognition methods based on characteristics of image according to claim 6, it is characterized in that, obtain the unique point that is complementary and judge also that to rear whether the right logarithm of image to be identified the be complementary unique point corresponding with every template image of choosing is greater than threshold value μ, if greater than executing data purification step, otherwise judge that directly the matching degree of two images is 0.
8. the described bayonet socket vehicle image recognition methods based on characteristics of image of according to claim 5-7 any one, it is characterized in that, the unique point that described basis is complementary is to calculating the matching degree of two images, and the concrete steps that determine the template image that is complementary with this image to be identified according to matching degree are:
The unique point that is complementary according to image to be identified and each template image of choosing is to computed image matching degree IMD=N/ N
0, wherein N represents the unique point logarithm that is complementary, N
0Sum for SIFT unique point in image to be identified;
Select maximal value IMD from all IMD values
max, with IMD
maxCompare judgement IMD with setting threshold λ
maxWhether greater than λ, if judge this IMD
maxCorresponding template image and images match to be identified success, otherwise in the judge templet database less than the template image that is complementary with this image to be identified.
9. the described bayonet socket vehicle image recognition methods based on characteristics of image of according to claim 1-7 any one is characterized in that, described vehicle attribute comprise vehicle brand, vehicle model and vehicle year money.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2013100451514A CN103150904A (en) | 2013-02-05 | 2013-02-05 | Bayonet vehicle image identification method based on image features |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2013100451514A CN103150904A (en) | 2013-02-05 | 2013-02-05 | Bayonet vehicle image identification method based on image features |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN103150904A true CN103150904A (en) | 2013-06-12 |
Family
ID=48548937
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2013100451514A Pending CN103150904A (en) | 2013-02-05 | 2013-02-05 | Bayonet vehicle image identification method based on image features |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103150904A (en) |
Cited By (44)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103325259A (en) * | 2013-07-09 | 2013-09-25 | 西安电子科技大学 | Illegal parking detection method based on multi-core synchronization |
| CN103390166A (en) * | 2013-07-17 | 2013-11-13 | 中科联合自动化科技无锡有限公司 | Vehicle model consistency distinguishing method based on vehicle front face characteristics |
| CN103678558A (en) * | 2013-12-06 | 2014-03-26 | 中科联合自动化科技无锡有限公司 | Suspicion vehicle search method based on sift characteristic |
| WO2015003446A1 (en) * | 2013-07-10 | 2015-01-15 | Tencent Technology (Shenzhen) Company Limited | Methods and systems for image recognition |
| CN104658265A (en) * | 2015-03-08 | 2015-05-27 | 无锡桑尼安科技有限公司 | System for recognizing vehicle crossing full line and changing lanes at traffic intersection |
| US9195896B2 (en) | 2013-07-10 | 2015-11-24 | Tencent Technology (Shenzhen) Company Limited | Methods and systems for image recognition |
| CN105160691A (en) * | 2015-08-29 | 2015-12-16 | 电子科技大学 | Color histogram based vehicle body color identification method |
| CN105205444A (en) * | 2015-08-14 | 2015-12-30 | 合肥工业大学 | Vehicle logo identification method based on dot pair characteristics |
| CN105206060A (en) * | 2015-10-14 | 2015-12-30 | 武汉烽火众智数字技术有限责任公司 | Vehicle type recognition device and method based on SIFT characteristics |
| CN105243668A (en) * | 2015-10-13 | 2016-01-13 | 中山大学 | License plate image privacy removal method |
| CN105321350A (en) * | 2014-08-05 | 2016-02-10 | 北京大学 | Method and device for detection of fake plate vehicles |
| CN105320708A (en) * | 2014-08-05 | 2016-02-10 | 北京大学 | Establishment method of vehicle model database and server |
| CN105354530A (en) * | 2015-09-22 | 2016-02-24 | 浙江宇视科技有限公司 | Vehicle body color identification method and apparatus |
| CN105373813A (en) * | 2015-12-24 | 2016-03-02 | 四川华雁信息产业股份有限公司 | Equipment state image monitoring method and device |
| CN105652332A (en) * | 2016-02-24 | 2016-06-08 | 北京君和信达科技有限公司 | Radiation source control method and rapid pass type security check system |
| CN106127228A (en) * | 2016-06-16 | 2016-11-16 | 北方工业大学 | Remote sensing image ship detection candidate area identification method based on decision template classifier fusion |
| CN106169076A (en) * | 2016-07-22 | 2016-11-30 | 中山大学 | A kind of angle license plate image storehouse based on perspective transform building method |
| CN106295526A (en) * | 2016-07-28 | 2017-01-04 | 浙江宇视科技有限公司 | The method and device of Car image matching |
| CN107256394A (en) * | 2017-06-09 | 2017-10-17 | 北京深瞐科技有限公司 | Driver information and information of vehicles checking method, device and system |
| CN107609457A (en) * | 2016-07-12 | 2018-01-19 | 比亚迪股份有限公司 | Car body Color Recognition System and vehicle and car body color identification method |
| CN107967482A (en) * | 2017-10-24 | 2018-04-27 | 广东中科南海岸车联网技术有限公司 | Icon-based programming method and device |
| CN105320923B (en) * | 2014-08-05 | 2018-06-12 | 北京大学 | Model recognizing method and device |
| CN108389396A (en) * | 2018-02-28 | 2018-08-10 | 北京精英智通科技股份有限公司 | A kind of vehicle matching process, device and charge system based on video |
| CN108629366A (en) * | 2018-03-16 | 2018-10-09 | 佛山科学技术学院 | A kind of image-recognizing method of high-tension line steel tower |
| CN109033175A (en) * | 2018-06-25 | 2018-12-18 | 高新兴科技集团股份有限公司 | A kind of method and system to scheme to search vehicle |
| CN109189959A (en) * | 2018-09-06 | 2019-01-11 | 腾讯科技（深圳）有限公司 | A kind of method and device constructing image data base |
| CN109741406A (en) * | 2019-01-03 | 2019-05-10 | 广州广电银通金融电子科技有限公司 | A kind of body color recognition methods under monitoring scene |
| CN109800693A (en) * | 2019-01-08 | 2019-05-24 | 西安交通大学 | A kind of vehicle detection at night method based on Color Channel composite character |
| CN110427982A (en) * | 2019-07-12 | 2019-11-08 | 北京航天光华电子技术有限公司 | A kind of automatic wiring machine route correction method and system based on image procossing |
| CN110472643A (en) * | 2019-08-20 | 2019-11-19 | 山东浪潮人工智能研究院有限公司 | A kind of optical imagery employee's card identification method based on Feature Points Matching |
| CN110619401A (en) * | 2019-08-01 | 2019-12-27 | 重庆捷旭科技有限公司 | Visitor information vehicle management method and management system |
| CN111325722A (en) * | 2020-02-17 | 2020-06-23 | 江苏诚印科技有限公司 | Stamp image accurate identification method, stamp image identification processing method and stamp image identification system |
| CN111523473A (en) * | 2020-04-23 | 2020-08-11 | 北京百度网讯科技有限公司 | Mask wearing identification method, device, equipment and readable storage medium |
| CN111696217A (en) * | 2020-05-25 | 2020-09-22 | 上海金亥通信设备有限公司 | Park parking management system |
| CN112215234A (en) * | 2020-10-16 | 2021-01-12 | 神思电子技术股份有限公司 | Method for checking information of vehicle personnel at road interface |
| WO2021012521A1 (en) * | 2019-07-19 | 2021-01-28 | 平安科技（深圳）有限公司 | Search-based webpage forensics method and device, readable storage medium and server |
| CN112489436A (en) * | 2020-10-29 | 2021-03-12 | 浙江预策科技有限公司 | Vehicle identity recognition method, device and system and electronic device |
| US11068549B2 (en) | 2019-11-15 | 2021-07-20 | Capital One Services, Llc | Vehicle inventory search recommendation using image analysis driven by machine learning |
| CN113362612A (en) * | 2021-06-02 | 2021-09-07 | 国电内蒙古东胜热电有限公司 | Vehicle identification method and system |
| CN113469216A (en) * | 2021-05-31 | 2021-10-01 | 浙江中烟工业有限责任公司 | Retail terminal poster identification and integrity judgment method, system and storage medium |
| WO2021238855A1 (en) * | 2020-05-28 | 2021-12-02 | 同方威视技术股份有限公司 | Vehicle inspection method and system |
| CN114550113A (en) * | 2022-02-25 | 2022-05-27 | 阿波罗智联(北京)科技有限公司 | Image recognition method, device, equipment and medium for intelligent traffic |
| CN116563770A (en) * | 2023-07-10 | 2023-08-08 | 四川弘和数智集团有限公司 | Method, device, equipment and medium for detecting vehicle color |
| CN116740383A (en) * | 2023-06-19 | 2023-09-12 | 清远市保鸿涂料有限公司 | Vehicle paint formula accurate identification method and system based on image identification |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101436252A (en) * | 2008-12-22 | 2009-05-20 | 北京中星微电子有限公司 | Method and system for recognizing vehicle body color in vehicle video image |
| US20090202107A1 (en) * | 2008-02-08 | 2009-08-13 | Tk Holdings Inc. | Object detection and recognition system |
| CN102194130A (en) * | 2011-05-19 | 2011-09-21 | 苏州两江科技有限公司 | Vehicle classification method based on image recognition |
| CN102610102A (en) * | 2012-02-27 | 2012-07-25 | 安科智慧城市技术（中国）有限公司 | Suspect vehicle inspection and control method and system |
| CN102855758A (en) * | 2012-08-27 | 2013-01-02 | 无锡北邮感知技术产业研究院有限公司 | Detection method for vehicle in breach of traffic rules |
-
2013
- 2013-02-05 CN CN2013100451514A patent/CN103150904A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090202107A1 (en) * | 2008-02-08 | 2009-08-13 | Tk Holdings Inc. | Object detection and recognition system |
| CN101436252A (en) * | 2008-12-22 | 2009-05-20 | 北京中星微电子有限公司 | Method and system for recognizing vehicle body color in vehicle video image |
| CN102194130A (en) * | 2011-05-19 | 2011-09-21 | 苏州两江科技有限公司 | Vehicle classification method based on image recognition |
| CN102610102A (en) * | 2012-02-27 | 2012-07-25 | 安科智慧城市技术（中国）有限公司 | Suspect vehicle inspection and control method and system |
| CN102855758A (en) * | 2012-08-27 | 2013-01-02 | 无锡北邮感知技术产业研究院有限公司 | Detection method for vehicle in breach of traffic rules |
Non-Patent Citations (1)
| Title |
|---|
| 肖敬文等: "基于几何与颜色特征的公交车辆视频检测算", 《中山大学学报( 自然科学版)》 * |
Cited By (62)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103325259A (en) * | 2013-07-09 | 2013-09-25 | 西安电子科技大学 | Illegal parking detection method based on multi-core synchronization |
| WO2015003446A1 (en) * | 2013-07-10 | 2015-01-15 | Tencent Technology (Shenzhen) Company Limited | Methods and systems for image recognition |
| US9195896B2 (en) | 2013-07-10 | 2015-11-24 | Tencent Technology (Shenzhen) Company Limited | Methods and systems for image recognition |
| CN103390166B (en) * | 2013-07-17 | 2016-08-17 | 中科联合自动化科技无锡有限公司 | Vehicle model consistency discrimination method based on front face feature |
| CN103390166A (en) * | 2013-07-17 | 2013-11-13 | 中科联合自动化科技无锡有限公司 | Vehicle model consistency distinguishing method based on vehicle front face characteristics |
| CN103678558A (en) * | 2013-12-06 | 2014-03-26 | 中科联合自动化科技无锡有限公司 | Suspicion vehicle search method based on sift characteristic |
| CN105321350B (en) * | 2014-08-05 | 2018-01-23 | 北京大学 | Fake-licensed car detection method and device |
| CN105320923B (en) * | 2014-08-05 | 2018-06-12 | 北京大学 | Model recognizing method and device |
| CN105321350A (en) * | 2014-08-05 | 2016-02-10 | 北京大学 | Method and device for detection of fake plate vehicles |
| CN105320708A (en) * | 2014-08-05 | 2016-02-10 | 北京大学 | Establishment method of vehicle model database and server |
| CN105320708B (en) * | 2014-08-05 | 2018-12-25 | 北京大学 | The method for building up and server in model data library |
| CN104658265A (en) * | 2015-03-08 | 2015-05-27 | 无锡桑尼安科技有限公司 | System for recognizing vehicle crossing full line and changing lanes at traffic intersection |
| CN105205444A (en) * | 2015-08-14 | 2015-12-30 | 合肥工业大学 | Vehicle logo identification method based on dot pair characteristics |
| CN105160691A (en) * | 2015-08-29 | 2015-12-16 | 电子科技大学 | Color histogram based vehicle body color identification method |
| CN105354530B (en) * | 2015-09-22 | 2019-07-16 | 浙江宇视科技有限公司 | A kind of body color recognition methods and device |
| CN105354530A (en) * | 2015-09-22 | 2016-02-24 | 浙江宇视科技有限公司 | Vehicle body color identification method and apparatus |
| CN105243668B (en) * | 2015-10-13 | 2018-04-27 | 中山大学 | A kind of method that license plate image goes privacy |
| CN105243668A (en) * | 2015-10-13 | 2016-01-13 | 中山大学 | License plate image privacy removal method |
| CN105206060B (en) * | 2015-10-14 | 2017-12-29 | 武汉烽火众智数字技术有限责任公司 | A kind of vehicle type recognition device and its method based on SIFT feature |
| CN105206060A (en) * | 2015-10-14 | 2015-12-30 | 武汉烽火众智数字技术有限责任公司 | Vehicle type recognition device and method based on SIFT characteristics |
| CN105373813A (en) * | 2015-12-24 | 2016-03-02 | 四川华雁信息产业股份有限公司 | Equipment state image monitoring method and device |
| CN105652332A (en) * | 2016-02-24 | 2016-06-08 | 北京君和信达科技有限公司 | Radiation source control method and rapid pass type security check system |
| CN108897055A (en) * | 2016-02-24 | 2018-11-27 | 北京君和信达科技有限公司 | A kind of method for controlling radiation source and fast general formula safe examination system |
| CN106127228A (en) * | 2016-06-16 | 2016-11-16 | 北方工业大学 | Remote sensing image ship detection candidate area identification method based on decision template classifier fusion |
| CN107609457A (en) * | 2016-07-12 | 2018-01-19 | 比亚迪股份有限公司 | Car body Color Recognition System and vehicle and car body color identification method |
| CN107609457B (en) * | 2016-07-12 | 2021-11-12 | 比亚迪股份有限公司 | Vehicle body color recognition system, vehicle and vehicle body color recognition method |
| CN106169076A (en) * | 2016-07-22 | 2016-11-30 | 中山大学 | A kind of angle license plate image storehouse based on perspective transform building method |
| CN106169076B (en) * | 2016-07-22 | 2019-05-14 | 中山大学 | A kind of angle license plate image library building method based on perspective transform |
| CN106295526A (en) * | 2016-07-28 | 2017-01-04 | 浙江宇视科技有限公司 | The method and device of Car image matching |
| CN106295526B (en) * | 2016-07-28 | 2019-10-18 | 浙江宇视科技有限公司 | The method and device of Car image matching |
| CN107256394A (en) * | 2017-06-09 | 2017-10-17 | 北京深瞐科技有限公司 | Driver information and information of vehicles checking method, device and system |
| CN107967482A (en) * | 2017-10-24 | 2018-04-27 | 广东中科南海岸车联网技术有限公司 | Icon-based programming method and device |
| CN108389396A (en) * | 2018-02-28 | 2018-08-10 | 北京精英智通科技股份有限公司 | A kind of vehicle matching process, device and charge system based on video |
| CN108629366A (en) * | 2018-03-16 | 2018-10-09 | 佛山科学技术学院 | A kind of image-recognizing method of high-tension line steel tower |
| CN109033175A (en) * | 2018-06-25 | 2018-12-18 | 高新兴科技集团股份有限公司 | A kind of method and system to scheme to search vehicle |
| CN109189959B (en) * | 2018-09-06 | 2020-11-10 | 腾讯科技（深圳）有限公司 | Method and device for constructing image database |
| CN109189959A (en) * | 2018-09-06 | 2019-01-11 | 腾讯科技（深圳）有限公司 | A kind of method and device constructing image data base |
| CN109741406A (en) * | 2019-01-03 | 2019-05-10 | 广州广电银通金融电子科技有限公司 | A kind of body color recognition methods under monitoring scene |
| CN109800693B (en) * | 2019-01-08 | 2021-05-28 | 西安交通大学 | Night vehicle detection method based on color channel mixing characteristics |
| CN109800693A (en) * | 2019-01-08 | 2019-05-24 | 西安交通大学 | A kind of vehicle detection at night method based on Color Channel composite character |
| CN110427982A (en) * | 2019-07-12 | 2019-11-08 | 北京航天光华电子技术有限公司 | A kind of automatic wiring machine route correction method and system based on image procossing |
| WO2021012521A1 (en) * | 2019-07-19 | 2021-01-28 | 平安科技（深圳）有限公司 | Search-based webpage forensics method and device, readable storage medium and server |
| CN110619401A (en) * | 2019-08-01 | 2019-12-27 | 重庆捷旭科技有限公司 | Visitor information vehicle management method and management system |
| CN110472643A (en) * | 2019-08-20 | 2019-11-19 | 山东浪潮人工智能研究院有限公司 | A kind of optical imagery employee's card identification method based on Feature Points Matching |
| US11775597B2 (en) | 2019-11-15 | 2023-10-03 | Capital One Services, Llc | Vehicle inventory search recommendation using image analysis driven by machine learning |
| US11068549B2 (en) | 2019-11-15 | 2021-07-20 | Capital One Services, Llc | Vehicle inventory search recommendation using image analysis driven by machine learning |
| CN111325722B (en) * | 2020-02-17 | 2024-02-20 | 江苏诚印科技有限公司 | Seal image accurate identification method and system and seal image identification processing method |
| CN111325722A (en) * | 2020-02-17 | 2020-06-23 | 江苏诚印科技有限公司 | Stamp image accurate identification method, stamp image identification processing method and stamp image identification system |
| CN111523473A (en) * | 2020-04-23 | 2020-08-11 | 北京百度网讯科技有限公司 | Mask wearing identification method, device, equipment and readable storage medium |
| CN111523473B (en) * | 2020-04-23 | 2023-09-26 | 北京百度网讯科技有限公司 | Mask wearing recognition method, device, equipment and readable storage medium |
| CN111696217A (en) * | 2020-05-25 | 2020-09-22 | 上海金亥通信设备有限公司 | Park parking management system |
| WO2021238855A1 (en) * | 2020-05-28 | 2021-12-02 | 同方威视技术股份有限公司 | Vehicle inspection method and system |
| CN112215234A (en) * | 2020-10-16 | 2021-01-12 | 神思电子技术股份有限公司 | Method for checking information of vehicle personnel at road interface |
| CN112489436A (en) * | 2020-10-29 | 2021-03-12 | 浙江预策科技有限公司 | Vehicle identity recognition method, device and system and electronic device |
| CN112489436B (en) * | 2020-10-29 | 2022-05-03 | 浙江预策科技有限公司 | Vehicle identity recognition method, device and system and electronic device |
| CN113469216A (en) * | 2021-05-31 | 2021-10-01 | 浙江中烟工业有限责任公司 | Retail terminal poster identification and integrity judgment method, system and storage medium |
| CN113469216B (en) * | 2021-05-31 | 2024-02-23 | 浙江中烟工业有限责任公司 | Retail terminal poster identification and integrity judgment method, system and storage medium |
| CN113362612A (en) * | 2021-06-02 | 2021-09-07 | 国电内蒙古东胜热电有限公司 | Vehicle identification method and system |
| CN114550113A (en) * | 2022-02-25 | 2022-05-27 | 阿波罗智联(北京)科技有限公司 | Image recognition method, device, equipment and medium for intelligent traffic |
| CN116740383A (en) * | 2023-06-19 | 2023-09-12 | 清远市保鸿涂料有限公司 | Vehicle paint formula accurate identification method and system based on image identification |
| CN116563770A (en) * | 2023-07-10 | 2023-08-08 | 四川弘和数智集团有限公司 | Method, device, equipment and medium for detecting vehicle color |
| CN116563770B (en) * | 2023-07-10 | 2023-09-29 | 四川弘和数智集团有限公司 | Method, device, equipment and medium for detecting vehicle color |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103150904A (en) | Bayonet vehicle image identification method based on image features | |
| CN103810505B (en) | Vehicles identifications method and system based on multiple layer description | |
| CN108596277B (en) | Vehicle identity recognition method and device and storage medium | |
| CN106778583B (en) | Vehicle attribute identification method and device based on convolutional neural network | |
| CN110837807A (en) | Identification method and system for fake-licensed vehicle | |
| CN103413147B (en) | A kind of licence plate recognition method and system | |
| CN108154160A (en) | Color recognizing for vehicle id method and system | |
| CN109657664B (en) | License plate type identification method and device and electronic equipment | |
| CN108960055B (en) | Lane line detection method based on local line segment mode characteristics | |
| CN104200228B (en) | Recognizing method and system for safety belt | |
| CN109190513A (en) | In conjunction with the vehicle of saliency detection and neural network again recognition methods and system | |
| CN110245691A (en) | A kind of intelligent identification Method of vehicle appearance color discoloration repacking | |
| CN103425989A (en) | Vehicle color identification method based on significance analysis | |
| CN105825233A (en) | Pedestrian detection method based on random fern classifier of online learning | |
| CN109215364A (en) | Traffic signals recognition methods, system, equipment and storage medium | |
| CN109993806A (en) | A kind of color identification method, device and electronic equipment | |
| CN109635637A (en) | A kind of licence plate recognition method, device and calculate equipment | |
| CN106650611B (en) | Method and device for recognizing color of vehicle body | |
| CN110378254A (en) | Recognition methods, system, electronic equipment and the storage medium of vehicle damage amending image trace | |
| CN109325487A (en) | A kind of full type licence plate recognition method based on target detection | |
| Pondenkandath et al. | Identifying cross-depicted historical motifs | |
| CN108510483B (en) | Method for generating color image tampering detection by adopting VLAD coding and SVM calculation | |
| CN104123553B (en) | A kind of license plate locating method and system based on cascade morphological transformation | |
| CN110458234B (en) | Vehicle searching method with map based on deep learning | |
| CN111652200A (en) | Processing method, device and equipment for distinguishing multiple vehicles from pictures in vehicle insurance case |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20130612 |
|
| RJ01 | Rejection of invention patent application after publication |