CN113065531A - Vehicle identification method for three-dimensional spliced video of expressway service area - Google Patents
Vehicle identification method for three-dimensional spliced video of expressway service area Download PDFInfo
- Publication number
- CN113065531A CN113065531A CN202110523304.6A CN202110523304A CN113065531A CN 113065531 A CN113065531 A CN 113065531A CN 202110523304 A CN202110523304 A CN 202110523304A CN 113065531 A CN113065531 A CN 113065531A
- Authority
- CN
- China
- Prior art keywords
- feature points
- plane
- points
- sampling
- image
- 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
- 238000000034 method Methods 0.000 title claims abstract description 29
- 238000005070 sampling Methods 0.000 claims abstract description 51
- 238000012216 screening Methods 0.000 claims abstract description 13
- 238000013507 mapping Methods 0.000 claims abstract description 7
- 239000011159 matrix material Substances 0.000 claims description 15
- 238000001514 detection method Methods 0.000 claims description 7
- 238000000354 decomposition reaction Methods 0.000 claims description 6
- 239000013598 vector Substances 0.000 claims description 6
- 239000003086 colorant Substances 0.000 claims description 3
- 238000010276 construction Methods 0.000 description 3
- 230000004075 alteration Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 239000002360 explosive Substances 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/49—Segmenting video sequences, i.e. computational techniques such as parsing or cutting the sequence, low-level clustering or determining units such as shots or scenes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/08—Detecting or categorising vehicles
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Multimedia (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computing Systems (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Image Analysis (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention discloses a vehicle identification method for a three-dimensional spliced video of a highway service area, which comprises the following steps: s1, selecting a three-dimensional frame image from the three-dimensional spliced video, carrying out UV expansion on the three-dimensional frame image, selecting a minimum rectangular frame for the expanded plane, and selecting the frame of the minimum rectangular frame as a plane M to be detected; s2, down-sampling a plane M to be detected to obtain a multi-level sampling plane, dividing the multi-level sampling plane into a plurality of small image blocks to determine the feature points and the feature point inner directions of the small image blocks, clustering the feature points to obtain key feature points of the multi-level sampling plane, screening key feature points with high correlation, and matching the screened key feature points; and S3, calibrating the vehicle detected in the sampling plane, and mapping the vehicle to the original three-dimensional frame image through UV to realize vehicle identification.
Description
Technical Field
The invention relates to the field of computer vision and the field of intelligent transportation, in particular to a vehicle identification method for a three-dimensional spliced video of a highway service area.
Background
With the rapid development of economy in China, the continuous increase of the total mileage of the expressway and the explosive increase of the number of passenger cars and freight cars, the construction of the expressway service area gradually moves to informatization and modernization. However, due to the large passenger flow in the service area of the expressway, a large number of vehicles are driven, and occasionally traffic events such as vehicle collision and anchorage, road construction, cargo scattering and the like occur, so that transient congestion of roads is caused at light, and the traffic flow is interrupted and the traffic is blocked at heavy. In order to reduce traffic delay and ensure road safety, traffic events need to be detected accurately and quickly.
The expressway service area is a place for providing services such as dining, shopping, refueling, maintaining and vehicle repairing and the like along the expressway, and is very important for safety in trip, fatigue driving prevention and driving efficiency improvement. As the construction mileage of the highway increases year by year, more service areas are put into use, the pedestrian flow and the vehicle flow of the highway service areas are large, and the following safety problem cannot be ignored. The driver keeps high tension when driving on the highway all the time, the driver can feel tired when driving on the highway for a long time, and the driver is easy to be careless and not concentrated when entering the service area, so the high-speed service area is also a high-incidence area of safety accidents. In order to reduce traffic delay and ensure road safety, real-time dynamic statistics on traffic flow information in an area needs to be performed.
Disclosure of Invention
The invention aims to provide a vehicle identification method for a three-dimensional spliced video of a highway service area, which has feasibility and high efficiency for detecting small targets and solves the problem of vehicle detection of the highway service area.
In order to achieve the purpose, the invention is realized by the following technical scheme:
a vehicle identification method for a three-dimensional spliced video of a highway service area is characterized by comprising the following steps:
s1, selecting a three-dimensional frame image from the three-dimensional spliced video, carrying out UV expansion on the three-dimensional frame image, selecting a minimum rectangular frame for the expanded plane, and selecting the frame of the minimum rectangular frame as a plane M to be detected;
s2, down-sampling a plane M to be detected to obtain a multi-level sampling plane, dividing the multi-level sampling plane into a plurality of small image blocks to determine the feature points and the feature point inner directions of the small image blocks, clustering the feature points to obtain key feature points of the multi-level sampling plane, screening key feature points with high correlation, and matching the screened key feature points;
and S3, calibrating the vehicle detected in the sampling plane, and mapping the vehicle to the original three-dimensional frame image through UV to realize vehicle identification.
The down-sampling of the plane M to be detected in step S2 to obtain a multi-level sampling plane includes:
the plane M to be detected is downsampled according to the proportion of an image 1/2 to obtain a first-level image pyramid M1, M1 is downsampled to obtain M2, and 5 levels of image pyramids M, M1, M2, M3 and M4 are obtained by repeated downsampling.
The step S2 of dividing the multi-level sampling plane into a plurality of small image blocks to determine the feature points and the directions within the feature points of each small image block includes:
setting a decomposition threshold, equally dividing the image into m parts to obtain image blocks Ni (i is 1, 2 and 3.. m), equally dividing the image blocks Ni for j times until the image blocks Ni are smaller than the set decomposition threshold, and stopping decomposing the image blocks to obtain h small image blocks, wherein h is m × j;
traversing the brightness of pixel points in the small image block, comparing the brightness of the selected pixel point with the brightness of surrounding pixel points, and if the brightness of the selected pixel point is greater than or less than the brightness of the surrounding pixel points, taking the pixel point as a feature point;
after the image block feature points are found, the feature principal direction of the image feature points is determined, and in each small image block B, the mass center is obtained: where g (X, Y) is a gray scale expression and X, Y are gray scale values, then:
the direction of the internal characteristic point is obtained through the position of the center of mass point as follows:
before clustering the feature points in step S2, the method further includes:
the extracted characteristic points are standardized by a mean absolute deviation method, and the formula is
Wherein n is the number of selected characteristic points in the image, and the influence of outliers is reduced by a mean absolute deviation method so as to reduce the data set D to { a }iEach characteristic point a of 1, 2, …, n | i ═iCalculating the distance sum S between the calculated distance sum and the rest of the feature pointsiAnd the distances of all feature points are equal to H, if Si>H, then aiIs regarded as an isolated point, is deleted, and eliminates the influence of noise and the isolated point, wherein
And S2, clustering the feature points, obtaining key feature points of a multilevel sampling plane after clustering, screening key feature points with high correlation, and matching the screened key feature points, wherein the matching comprises the following steps:
s231, selecting a proper K value to cluster the obtained feature points in the fixed internal feature direction;
step S232, calculating cosine distances of each point position of the plane, comparing cosine values of plane vectors to measure different individual differences, comparing included angle values and screening direction differences of the characteristic vectors;
step 233, repeating step 231 and step 232 to obtain cluster blocks and cluster centers which reach a stable state, and selecting different colors to label the clusters;
step S234, establishing a characteristic matrix T for the obtained characteristic points, wherein each row in the matrix represents one characteristic point, extracting a plurality of sampling points for each characteristic point, and P sampling pairs are established by the plurality of sampling points, wherein the characteristic matrix T comprises P columns;
step S235, calculating the average value and variance of each row of the matrix, sorting each row of the feature matrix T according to the descending of the variance of each row, selecting the front Q rows, and finishing the screening of key feature points;
step S236, a binary descriptor of the Q columns is obtained, the columns of the binary descriptor are arranged from high variance to low variance, and the first Q/2 columns are selected for matching.
After step S236, the method further includes:
and if the distance between the front Q/2 columns of the two feature points to be matched is smaller than a set threshold value, matching by using the residual column information.
The step S3 further includes:
if a blank area is detected in the sampling plane, removing the blank grids after blank area grids are blanked, and not carrying out subsequent UV (ultraviolet) mapping work;
and if the object filling area is detected in the sampling plane, carrying out vehicle detection on the object filling area, and calibrating the detected grid outer frame of the vehicle area.
Compared with the prior art, the invention has the following advantages:
the method has feasibility and high efficiency for detecting the small target, and solves the problem of vehicle detection in the expressway service area.
Drawings
FIG. 1 is a flow chart of a vehicle identification method for three-dimensional spliced videos of a highway service area, which is disclosed by the invention;
fig. 2 is a flowchart of step S2;
fig. 3 is a flow chart for determining the direction of an inner feature point.
Detailed Description
The present invention will now be further described by way of the following detailed description of a preferred embodiment thereof, taken in conjunction with the accompanying drawings.
As shown in fig. 1, a vehicle identification method facing a three-dimensional spliced video of a highway service area includes the following steps:
s1, selecting a three-dimensional frame image from the three-dimensional spliced video, carrying out UV expansion on the three-dimensional frame image, selecting a minimum rectangular frame for the expanded plane, and selecting the frame of the minimum rectangular frame as a plane M to be detected;
s2, down-sampling a plane M to be detected to obtain a multi-level sampling plane, dividing the multi-level sampling plane into a plurality of small image blocks to determine the feature points and the feature point inner directions of the small image blocks, clustering the feature points to obtain key feature points of the multi-level sampling plane, screening key feature points with high correlation, and matching the screened key feature points;
and S3, calibrating the vehicle detected in the sampling plane, and mapping the vehicle to the original three-dimensional frame image through UV to realize vehicle identification.
As shown in fig. 3, the down-sampling the plane M to be detected in step S2 to obtain a multi-level sampling plane includes:
downsampling a plane M to be detected according to the proportion of an image 1/2 to obtain a first-level image pyramid M1, downsampling M1 to obtain M2, and repeatedly downsampling to obtain 5-level image pyramids M, M1, M2, M3 and M4, wherein M1 is obtained by downsampling M2 is obtained by downsampling M1, and the like, and M4 is obtained by downsampling M3; has the following relationship: the purpose of constructing M, M1, M2, M3, M4 is to get more useful information about the image.
As shown in fig. 2, the step S2 of dividing the multi-level sampling plane into a plurality of small image blocks to determine the feature points and the feature point inner directions of each of the small image blocks includes:
setting a decomposition threshold, equally dividing the image into m parts to obtain image blocks Ni (i is 1, 2 and 3.. m), equally dividing the image blocks Ni for j times until the image blocks Ni are smaller than the set decomposition threshold, and stopping decomposing the image blocks to obtain h small image blocks, wherein h is m × j;
traversing the brightness of pixel points in the image block, comparing the brightness of the selected pixel point with the brightness of surrounding pixel points, and if the brightness of the selected pixel point is greater than or less than the brightness of the surrounding pixel points, taking the pixel point as a feature point;
after the image block feature points are found, the feature principal direction of the image feature points is determined, and in each small image block B, the mass center is obtained: where g (X, Y) is a gray scale expression and X, Y are gray scale values, then:
the direction of the internal characteristic point is obtained through the position of the center of mass point as follows:
before clustering the feature points in step S2, the method further includes:
the extracted characteristic points are standardized by a mean absolute deviation method, and the formula is
Wherein n is the number of selected characteristic points in the image, and the influence of outliers is reduced by a mean absolute deviation method so as to reduce the data set D to { a }i1, 2, …, niCalculating the distance sum S between the calculated distance sum and the rest of the feature pointsiAnd the distances of all feature points are equal to H, if Si>H, then aiIs regarded as an isolated point, is deleted, and eliminates the influence of noise and the isolated point, wherein
And S2, clustering the feature points, obtaining key feature points of a multilevel sampling plane after clustering, screening key feature points with high correlation, and matching the screened key feature points, wherein the matching comprises the following steps:
s231, selecting a proper K value to cluster the obtained feature points in the fixed internal feature direction, so that the features in the image are more obvious and are convenient to distinguish, wherein the K value is 0.98 in the embodiment, and the clustering center is a point which takes the K value as a central point and is in a certain range of cosine included angle values of the K value, namely, the feature points have similar features, so that the cycle process of mismatching and identification is reduced, the matching accuracy is improved, and the matching rate is accelerated;
step S232, calculating cosine distances of each point position of the plane, comparing cosine values of plane vectors to measure different individual differences, comparing included angle values and screening direction differences of the characteristic vectors;
step 233, repeating step 231 and step 232 to obtain cluster blocks and cluster centers which reach a stable state, and selecting different colors to label the clusters;
step S234, establishing a feature matrix T for the obtained feature points, where each row in the matrix represents a feature point, 43 sampling points are extracted for each feature point, P sampling pairs are established for the 43 sampling points, where P is 903, the feature matrix T includes 903 columns, and there are 6 pairing modes among sampling points, such as three sampling points a1, a2, and a3, and each pairing mode is called a sampling pair;
step S235, calculating the average value and variance of each row of the matrix, sorting each row of the characteristic matrix T according to the variance of each row from large to small, selecting the front Q rows, and finishing the screening of key characteristic points, wherein Q is 256;
step S236, 256 binary descriptors are obtained, the binary descriptors are arranged from high variance to low variance, the high variance represents the fuzzy information, the low variance represents the detail information, and the first 128 columns are selected for matching.
Further, the binary descriptor is a series of binary strings, F denotes the binary descriptor, Pa is a sampling point pair, N denotes a binary code length,
which represents the pixel value of the previous sample point in the pair Pa, and, similarly,
representing the pixel value of the next sample point.
After step S236, the method further includes:
and if the distance between the front 128 columns of the two feature points to be matched is smaller than a set threshold value, matching by using the residual column information.
The step S3 further includes:
if a blank area is detected in the sampling plane, removing the blank grids after blank area grids are blanked, and not carrying out subsequent UV (ultraviolet) mapping work;
if an object filling area is detected in the sampling plane, vehicle detection is carried out on the object filling area, the detected grid outer frame of the vehicle area is calibrated, and vehicle calibration frames are divided into three types: car, bus, and truck.
The calibrated vehicle area is mapped into the original three-dimensional video frame image according to the UV, and vehicle detection work on the frame image is completed; and (5) calibrating and identifying the vehicles in the rest frames of the three-dimensional video according to the steps.
In conclusion, the vehicle identification method for the three-dimensional spliced video of the expressway service area has feasibility and high efficiency for detecting small targets, and solves the problem of vehicle detection of the expressway service area.
While the present invention has been described in detail with reference to the preferred embodiments, it should be understood that the above description should not be taken as limiting the invention. Various modifications and alterations to this invention will become apparent to those skilled in the art upon reading the foregoing description. Accordingly, the scope of the invention should be determined from the following claims.
Claims (7)
1. A vehicle identification method for a three-dimensional spliced video of a highway service area is characterized by comprising the following steps:
s1, selecting a three-dimensional frame image from the three-dimensional spliced video, carrying out UV expansion on the three-dimensional frame image, selecting a minimum rectangular frame for the expanded plane, and selecting the frame of the minimum rectangular frame as a plane M to be detected;
s2, down-sampling a plane M to be detected to obtain a multi-level sampling plane, dividing the multi-level sampling plane into a plurality of small image blocks to determine the feature points and the feature point inner directions of the small image blocks, clustering the feature points to obtain key feature points of the multi-level sampling plane, screening key feature points with high correlation, and matching the screened key feature points;
and S3, calibrating the vehicle detected in the sampling plane, and mapping the vehicle to the original three-dimensional frame image through UV to realize vehicle identification.
2. The method for identifying vehicles based on the three-dimensional spliced video of the service area of the expressway as claimed in claim 1, wherein the down-sampling the plane M to be detected to obtain a multi-level sampling plane in step S2 comprises:
the plane M to be detected is downsampled according to the proportion of an image 1/2 to obtain a first-level image pyramid M1, M1 is downsampled to obtain M2, and 5 levels of image pyramids M, M1, M2, M3 and M4 are obtained by repeated downsampling.
3. The method for identifying vehicles oriented to the three-dimensional spliced video of the service area of the expressway as claimed in claim 2, wherein the step S2 of dividing the multi-level sampling plane into a plurality of small image blocks to determine the feature points and the directions within the feature points of each small image block comprises:
setting a decomposition threshold, equally dividing the image into m parts to obtain image blocks Mi (i is 1, 2 and 3.. m), equally dividing the image blocks Mi j times until the image blocks are smaller than the set decomposition threshold, and stopping decomposing the image blocks to obtain h small image blocks, wherein h is m multiplied by j;
traversing the brightness of pixel points in the small image block, comparing the brightness of the selected pixel point with the brightness of surrounding pixel points, and if the brightness of the selected pixel point is greater than or less than the brightness of the surrounding pixel points, taking the pixel point as a feature point;
after the image block feature points are found, the feature principal direction of the image feature points is determined, and in each small image block B, the mass center is obtained: where g (X, Y) is a gray scale expression and X, Y are gray scale values, then:
the direction of the internal characteristic point is obtained through the position of the center of mass point as follows:
4. the method for identifying vehicles based on the three-dimensional spliced video of the service area of the expressway as claimed in claim 3, wherein before clustering the feature points in step S2, the method further comprises:
the extracted characteristic points are standardized by a mean absolute deviation method, and the formula is
Wherein n is the number of selected characteristic points in the image, and the influence of outliers is reduced by a mean absolute deviation method so as to reduce the data set D to { a }iEach characteristic point a of 1, 2, …, n | i ═iCalculating the distance sum S between the calculated distance sum and the rest of the feature pointsiAnd the distances of all feature points are equal to H, if Si>H, then aiIs regarded as an isolated point, is deleted, and eliminates the influence of noise and the isolated point, wherein
5. The method for identifying the vehicles oriented to the three-dimensional spliced video of the service area of the expressway as recited in claim 4, wherein the step S2 of clustering the feature points to obtain key feature points of a multilevel sampling plane after clustering, and the step S of screening the key feature points with high correlation comprises the steps of:
s231, selecting a proper K value to cluster the obtained feature points in the fixed internal feature direction;
step S232, calculating cosine distances of each point position of the plane, comparing cosine values of plane vectors to measure different individual differences, comparing included angle values and screening direction differences of the characteristic vectors;
step 233, repeating step 231 and step 232 to obtain cluster blocks and cluster centers which reach a stable state, and selecting different colors to label the clusters;
step S234, establishing a characteristic matrix T for the obtained characteristic points, wherein each row in the matrix represents one characteristic point, extracting a plurality of sampling points for each characteristic point, and P sampling pairs are established by the plurality of sampling points, wherein the characteristic matrix T comprises P columns;
step S235, calculating the average value and variance of each row of the matrix, sorting each row of the feature matrix T according to the descending of the variance of each row, selecting the front Q rows, and finishing the screening of key feature points;
step S236, a binary descriptor of the Q columns is obtained, the columns of the binary descriptor are arranged from high variance to low variance, and the first Q/2 columns are selected for matching.
6. The method for identifying vehicles based on the three-dimensional spliced video of the service area of the expressway as claimed in claim 5, wherein the step S236 is followed by further comprising:
and if the distance between the front Q/2 columns of the two feature points to be matched is smaller than a set threshold value, matching by using the residual column information.
7. The method for identifying vehicles based on the three-dimensional spliced video of the service area of the expressway as claimed in claim 1, wherein the step S3 further comprises:
if a blank area is detected in the sampling plane, removing the blank grids after blank area grids are blanked, and not carrying out subsequent UV (ultraviolet) mapping work;
and if the object filling area is detected in the sampling plane, carrying out vehicle detection on the object filling area, and calibrating the detected grid outer frame of the vehicle area.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110523304.6A CN113065531B (en) | 2021-05-13 | 2021-05-13 | Vehicle identification method for three-dimensional spliced video of expressway service area |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110523304.6A CN113065531B (en) | 2021-05-13 | 2021-05-13 | Vehicle identification method for three-dimensional spliced video of expressway service area |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113065531A true CN113065531A (en) | 2021-07-02 |
| CN113065531B CN113065531B (en) | 2024-05-14 |
Family
ID=76568652
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110523304.6A Active CN113065531B (en) | 2021-05-13 | 2021-05-13 | Vehicle identification method for three-dimensional spliced video of expressway service area |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113065531B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115205666A (en) * | 2022-09-16 | 2022-10-18 | 太平金融科技服务(上海)有限公司深圳分公司 | Image analysis method, apparatus, server, medium, and computer program product |
| CN116523852A (en) * | 2023-04-13 | 2023-08-01 | 成都飞机工业（集团）有限责任公司 | Foreign matter detection method of carbon fiber composite material based on feature matching |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105654507A (en) * | 2015-12-24 | 2016-06-08 | 北京航天测控技术有限公司 | Vehicle outer contour dimension measuring method based on image dynamic feature tracking |
| WO2018130016A1 (en) * | 2017-01-10 | 2018-07-19 | 哈尔滨工业大学深圳研究生院 | Parking detection method and device based on monitoring video |
| WO2019054561A1 (en) * | 2017-09-15 | 2019-03-21 | 서울과학기술대학교 산학협력단 | 360-degree image encoding device and method, and recording medium for performing same |
| CN111798560A (en) * | 2020-06-09 | 2020-10-20 | 同济大学 | Three-dimensional real-scene model visualization method for infrared thermal image temperature measurement data of power equipment |
-
2021
- 2021-05-13 CN CN202110523304.6A patent/CN113065531B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105654507A (en) * | 2015-12-24 | 2016-06-08 | 北京航天测控技术有限公司 | Vehicle outer contour dimension measuring method based on image dynamic feature tracking |
| WO2018130016A1 (en) * | 2017-01-10 | 2018-07-19 | 哈尔滨工业大学深圳研究生院 | Parking detection method and device based on monitoring video |
| WO2019054561A1 (en) * | 2017-09-15 | 2019-03-21 | 서울과학기술대학교 산학협력단 | 360-degree image encoding device and method, and recording medium for performing same |
| CN111798560A (en) * | 2020-06-09 | 2020-10-20 | 同济大学 | Three-dimensional real-scene model visualization method for infrared thermal image temperature measurement data of power equipment |
Non-Patent Citations (2)
| Title |
|---|
| 李子健;阮秋琦;: "基于LPP和改进SIFT的copy-move篡改检测", 信号处理, no. 04 * |
| 杨昊瑜;戴华林;王丽;张蕊;: "基于K-means的ORB算法在前方车辆识别中的应用", 传感器与微系统, no. 10 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115205666A (en) * | 2022-09-16 | 2022-10-18 | 太平金融科技服务(上海)有限公司深圳分公司 | Image analysis method, apparatus, server, medium, and computer program product |
| CN115205666B (en) * | 2022-09-16 | 2023-03-07 | 太平金融科技服务(上海)有限公司深圳分公司 | Image analysis method, image analysis device, image analysis server, and image analysis medium |
| CN116523852A (en) * | 2023-04-13 | 2023-08-01 | 成都飞机工业（集团）有限责任公司 | Foreign matter detection method of carbon fiber composite material based on feature matching |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113065531B (en) | 2024-05-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20210197851A1 (en) | Method for building virtual scenario library for autonomous vehicle | |
| CN102208013B (en) | Landscape coupling reference data generation system and position measuring system | |
| US9058744B2 (en) | Image based detecting system and method for traffic parameters and computer program product thereof | |
| US11208101B2 (en) | Device and method for measuring transverse distribution of wheel path | |
| CN113065531A (en) | Vehicle identification method for three-dimensional spliced video of expressway service area | |
| Yang et al. | Image-based visibility estimation algorithm for intelligent transportation systems | |
| CN111027430B (en) | Traffic scene complexity calculation method for intelligent evaluation of unmanned vehicles | |
| CN109448397B (en) | Group fog monitoring method based on big data | |
| CN113147733B (en) | Intelligent speed limiting system and method for automobile in rain, fog and sand dust weather | |
| CN101567041A (en) | Method for recognizing characters of number plate images of motor vehicles based on trimetric projection | |
| CN112381101B (en) | Infrared road scene segmentation method based on category prototype regression | |
| CN113378751A (en) | Traffic target identification method based on DBSCAN algorithm | |
| CN112084890A (en) | Multi-scale traffic signal sign identification method based on GMM and CQFL | |
| CN114612883A (en) | Forward vehicle distance detection method based on cascade SSD and monocular depth estimation | |
| CN114926984B (en) | Real-time traffic conflict collection and road safety evaluation method | |
| Oh et al. | In-depth understanding of lane changing interactions for in-vehicle driving assistance systems | |
| Ries et al. | Trajectory-based clustering of real-world urban driving sequences with multiple traffic objects | |
| BARODI et al. | Improved deep learning performance for real-time traffic sign detection and recognition applicable to intelligent transportation systems | |
| CN116630702A (en) | Pavement adhesion coefficient prediction method based on semantic segmentation network | |
| CN115544888A (en) | Dynamic scene boundary assessment method based on physical mechanism and machine learning hybrid theory | |
| CN114912689A (en) | Map grid index and XGBOST-based over-limit vehicle destination prediction method and system | |
| Ning et al. | Automatic driving scene target detection algorithm based on improved yolov5 network | |
| CN113095387B (en) | Road risk identification method based on networking vehicle-mounted ADAS | |
| CN112559968A (en) | Driving style representation learning method based on multi-situation data | |
| CN114882393B (en) | Road reverse running and traffic accident event detection method based on target detection |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB03 | Change of inventor or designer information |
Inventor after: Fu Chao Inventor after: Huang Shifeng Inventor after: Ma Yiming Inventor after: Shen Tianxing Inventor after: Yang Ju Inventor after: Liu Yi Inventor after: Chen Yunfei Inventor after: Ni Zhaoxin Inventor before: Huang Shifeng Inventor before: Ma Yiming Inventor before: Shen Tianxing Inventor before: Yang Ju Inventor before: Liu Yi Inventor before: Chen Yunfei Inventor before: Ni Zhaoxin Inventor before: Fu Chao |
|
| CB03 | Change of inventor or designer information | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |