CN106682641A - Pedestrian identification method based on image with FHOG- LBPH feature - Google Patents
Pedestrian identification method based on image with FHOG- LBPH feature Download PDFInfo
- Publication number
- CN106682641A CN106682641A CN201710006490.XA CN201710006490A CN106682641A CN 106682641 A CN106682641 A CN 106682641A CN 201710006490 A CN201710006490 A CN 201710006490A CN 106682641 A CN106682641 A CN 106682641A
- Authority
- CN
- China
- Prior art keywords
- fhog
- image
- lbph
- characteristic
- pedestrian
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/20—Movements or behaviour, e.g. gesture recognition
- G06V40/23—Recognition of whole body movements, e.g. for sport training
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Psychiatry (AREA)
- Social Psychology (AREA)
- Human Computer Interaction (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention provides a pedestrian identification method based on image with FHOG-LBPH feature. By statistical average of fusion HOG features (FHOG) , the optimal features are selected according to the combination with single optimal feature and the separable criterion of bhattacharyya distance, and improved FHOG-LBPH feature are obtained through fusion LBPH characteristics, fundamentally reducing the feature dimension; A classifier is obtained by using a support vector machine (SVM) to train the sample characteristics in order to obtain the classification of the test sample. The experimental results show that the method makes the pedestrian detection accuracy and real-time performance a certain improvement. Effectiveness of the method is validated by the image automatically shot, and the method has certain application value in the real pedestrian recognition.
Description
The technical field is as follows:
the invention provides an image pedestrian identification method based on FHOG-LBPH characteristics, and belongs to the technical field of computer vision monitoring.
Secondly, background art:
pedestrian detection is an important research direction in the field of computer vision and pattern recognition, and has good application value in the fields of intelligent transportation, video monitoring, crowd safety prediction and management, robots and advanced human-computer interaction and the like. Since the human body is a non-rigid target and is easily affected by the posture, clothing, vision, illumination and the like, and the precision of pedestrian detection is affected by the complex background environment, how to quickly and accurately detect pedestrians from video images is still a research focus at present.
Among many technical methods for detecting pedestrians, a detection method based on robotics is the mainstream method at present. The method comprises two important aspects: feature extraction and classifier learning. In the feature extraction process, the common pedestrian features include HOG features, LBP features and the like. The most applied in classifier learning is to train the sample characteristics by using SVM to obtain a classifier with higher performance. However, the detection of histogram of gradient directions (HOG) and Local Binary Pattern (LBP) has the problems of high feature dimension, more redundant information, influence on the pedestrian detection speed in video images, and the like.
Therefore, the invention provides an image pedestrian identification method based on FHOG-LBPH characteristics. The fusion HOG Features (FHOOG) are subjected to statistical averaging, the optimal features are selected by combining the single optimal feature combination and the Bhattacharyya distance separability criterion, and are fused with the LBPH features to obtain improved FHOOG-LBPH features, so that the feature dimension is reduced fundamentally; and training the sample characteristics by using a Support Vector Machine (SVM) to obtain a classifier, thereby achieving the purpose of classifying the test samples. The experimental result shows that the accuracy and the real-time performance of the pedestrian detection are improved to a certain extent by the method, the effectiveness of the method is verified by the self-shot image, and the method has a certain application value in the actual pedestrian identification.
Thirdly, the invention content:
(1) the purpose is as follows: the invention aims to provide an image pedestrian recognition method based on FHOG-LBPH characteristics, namely, the method comprises the steps of carrying out statistical averaging on fused HOG characteristics (FHOG), selecting optimal characteristics, fusing the optimal characteristics with the LBPH characteristics to obtain improved FHOG-LBPH characteristics, training sample characteristics by using a Support Vector Machine (SVM) to obtain a classifier, and achieving the purpose of classifying test samples while improving accuracy and instantaneity.
(2) The invention provides an image pedestrian identification method based on FHOG-LBPH characteristics, which specifically comprises the following steps:
the method comprises the following steps: and extracting FHOG characteristics. The FHOG descriptor is calculated by a process of mutually overlapping the area characteristics, the target area is divided into small areas through cell units, the small areas are combined into a block unit, and the FHOG characteristics are the combination of the area characteristics.
The FHOG descriptor is used for calculating the statistic value of the direction information of the local image gradient, describing the appearance and the shape of a local target and representing the contour information of a pedestrian. FHOG feature extraction comprises the following steps: the method comprises the steps of gray scale space normalization, gradient calculation, gradient direction histogram and overlapped block histogram normalization, and FHOG characteristic generation.
Step two: and (4) extracting LBPH characteristics. Since the feature description in step one is pedestrian contour edge information, a flat surface is ignored more, and the processing capability and robustness of the classifier for noisy edge information are poor. An efficient texture descriptor (LBPH) is thus introduced, which extracts and measures texture features of local neighbors in the gray-scale map. The LBPH characteristic extraction comprises the following steps: LBP characteristic image calculation, LBP characteristic image blocking, histogram normalization and LBPH characteristic vector calculation.
Because the texture characteristics are stable, the method is not easily influenced by background colors and illumination, and is beneficial to distinguishing images, thereby improving the accuracy of pedestrian detection.
Step three: FHOG-LBPH feature fusion. And (4) connecting the FHOG characteristic obtained in the first step and the second step with the LBPH characteristic in series to form a characteristic of an image to detect the pedestrian. The FHOG-LBPH fusion feature improves the robustness of the classifier as a whole.
Step four: and (5) identifying the pedestrian. And (4) training the training sample characteristics extracted in the step three by using an SVM to obtain a classifier, and detecting the test data set by using the classifier so as to judge whether the image is a pedestrian.
The invention has the advantages and positive effects that: in the aspect of image pedestrian identification, the image features of FHOG + LBPH are combined based on the description of FHOG features on pedestrian contour edge information and the description of LBPH features on local adjacent region texture information in a gray scale image, so that the image identification method is more beneficial to image distinguishing, the processing capability and robustness of a classifier on noisy edge information are enhanced, the pedestrian detection efficiency is improved, and the method has a certain application value in actual pedestrian identification.
Fourthly, explanation of the attached drawings:
FIG. 1 is a flowchart illustrating the overall steps of an image pedestrian recognition method based on FHOG-LBPH characteristics according to the present invention;
FIG. 2 shows the experimental effect of the image pedestrian recognition method based on FHOG-LBPH characteristics.
The fifth embodiment is as follows:
the image pedestrian recognition method based on FHOG-LBPH characteristics of the invention is further described with reference to the figures 1 and 2:
the method comprises the following steps: extracting FHOG characteristics; and calculating a statistical value of the direction information of the local image gradient, and representing the contour information of the pedestrian.
Step 1.1, normalizing the input image by using a Gamma method.
Where I (x, y) represents the gray value of the current pixel. The method aims to adjust the contrast of an image, reduce the influence caused by local shadow and illumination change of the image and simultaneously inhibit the interference of noise.
Step 1.2, pixel gradient calculation.
The FHOG characteristic is very sensitive to a template operator in calculation, and the simplest one-dimensional discrete differential template (-1, 0, +1) and the transpose thereof are found to have the best detection effect by performing gradient calculation on each pixel of the image in the horizontal direction and the vertical direction through comparison. The calculation process is as follows:
Gx(x,y)＝H(x+1,y)-H(x-1,y)
Gy(x,y)＝H(x,y+1)-H(x,y-1)
ω(x,y)＝tan-1(Gy(x,y)/Gx(x,y))
Gx(x,y)，Gy(x, y) are respectively the horizontal gradient and the vertical gradient of the current pixel point; g (x, y) is the gradient amplitude of the current pixel point; and omega (x, y) is the gradient direction of the current pixel point.
Limiting the gradient direction in the range of [0, pi ], equally dividing the gradient direction into 9 intervals, and then using the amplitude in the direction to make a weighted statistical histogram for each interval.
Step 1.3, intra-block normalization. Dividing the image into cells and blocks with the same size; unlike HOG which extracts 36-dimensional features per cell, FHOG extracts 31-dimensional features per cell and normalizes the 31-dimensional feature vectors.
The purpose of normalization is to further eliminate the influence of illumination and shadow:
wherein,and viThe gradient values of each pixel after the initial and normalization are respectively expressed, and are a constant, and the value is 0.001.
And step 1.4, acquiring FHOG characteristics. And (3) normalizing each Block by adopting a sliding window search method according to the normalization method in the step 1.3, and connecting the feature vectors in each normalized Block in series to obtain the final feature vector.
Step two: and (4) extracting LBPH characteristics. LBPH is an effective texture descriptor, which extracts and measures texture features of local neighboring areas in a gray-scale map, and features thereof are relatively stable and are not easily affected by background color and illumination.
Step 2.1, calculating an LBP characteristic image (LBP value of a rotation invariant equivalence mode) of the image;
step 2.2, partitioning the LBP characteristic image;
step 2.3, calculating a histogram of each regional characteristic image, and normalizing the histogram;
step 2.4, arranging the histograms of the regional characteristic images in a row according to the spatial sequence of the blocks to form an LBPH characteristic vector;
step three: FHOG-LBPH feature fusion. Since FHOG feature dimension is high, a flat surface is ignored, and the robustness of the classifier is poor, the pedestrian is detected by a method of fusing the selected FHOG feature and the LBPH feature, and the fusing method is as follows:
and (4) connecting the FHOG characteristic obtained in the first step and the second step with the LBPH characteristic in series to form a characteristic of an image to detect the pedestrian. Wherein the cell size is 8 multiplied by 8 when the FHOG characteristic is extracted, the padding size is 1 multiplied by 1, and the FHOG characteristic of the whole image is 10 multiplied by 18 multiplied by 31 dimensions; when extracting LBPH characteristics, the cell size is 8 multiplied by 8, each block area counts 9 histograms, and 12 multiplied by 20 multiplied by 9 dimensional characteristics are extracted from the whole image.
Step four: and (5) identifying the pedestrian. And (4) training the training sample characteristics extracted in the step three by using an SVM to obtain a classifier, and detecting the test data set by using the classifier so as to judge whether the image is a pedestrian.
Finally, the pedestrian identification is verified through tests; the experiment uses the pedestrian image library of INRIA and adds a part of the pedestrian pictures as positive samples for training the classifier. Dividing the data set into a training set and a testing set according to the proportion of 9:1, training a linear SVM classifier by using the training set, and testing the classification effect of the classifier by using the testing set. The result shows that the FHOG-LBPH characteristic firms the characteristics of the image in two aspects of shape and texture, and enhances the processing capability and robustness of the classifier on noisy edge information; through classification training of the SVM, pedestrians and non-pedestrians can be well recognized, the efficiency of pedestrian detection is improved, and the method has a certain application value in actual pedestrian recognition.
It should be noted that this example is merely illustrative of the method of application of the present invention and is not intended to limit the present invention. Any person skilled in the art can modify the above embodiments without departing from the spirit and scope of the present invention. Therefore, the scope of the invention should be determined from the following claims.
Claims (4)
1. An image pedestrian recognition method based on FHOG-LBPH characteristics is characterized in that:
the method comprises the following concrete steps:
the method comprises the following steps: and extracting FHOG characteristics. The FHOG descriptor is calculated by a process of mutually overlapping the area characteristics, the target area is divided into small areas through cell units, the small areas are combined into a block unit, and the FHOG characteristics are the combination of the area characteristics.
The FHOG descriptor is used for calculating the statistic value of the direction information of the local image gradient, describing the appearance and the shape of a local target and representing the contour information of a pedestrian. FHOG feature extraction comprises the following steps: the method comprises the steps of gray scale space normalization, gradient calculation, gradient direction histogram and overlapped block histogram normalization, and FHOG characteristic generation.
Step two: and (4) extracting LBPH characteristics. Since the feature description in step one is pedestrian contour edge information, a flat surface is ignored more, and the processing capability and robustness of the classifier for noisy edge information are poor. An efficient texture descriptor (LBPH) is thus introduced, which extracts and measures texture features of local neighbors in the gray-scale map. The LBPH characteristic extraction comprises the following steps: LBP characteristic image calculation, LBP characteristic image blocking, histogram normalization and LBPH characteristic vector calculation.
Because the texture characteristics are stable, the method is not easily influenced by background colors and illumination, and is beneficial to distinguishing images, thereby improving the accuracy of pedestrian detection.
Step three: FHOG-LBPH feature fusion. And (4) connecting the FHOG characteristic obtained in the first step and the second step with the LBPH characteristic in series to form a characteristic of an image to detect the pedestrian. The FHOG-LBPH fusion feature improves the robustness of the classifier as a whole.
Step four: and (5) identifying the pedestrian. And (4) training the training sample characteristics extracted in the step three by using an SVM to obtain a classifier, and detecting the test data set by using the classifier so as to judge whether the image is a pedestrian.
2. An image pedestrian recognition method based on FHOG-LBPH characteristics is characterized in that:
the FHOG characteristic extraction in the step one comprises the following specific steps:
step 1.1, normalizing the input image by using a Gamma method.
Where I (x, y) represents the gray value of the current pixel. The method aims to adjust the contrast of an image, reduce the influence caused by local shadow and illumination change of the image and simultaneously inhibit the interference of noise.
Step 1.2, pixel gradient calculation.
The FHOG characteristic is very sensitive to a template operator in calculation, and the simplest one-dimensional discrete differential template (-1, 0, +1) and the transpose thereof are found to have the best detection effect by performing gradient calculation on each pixel of the image in the horizontal direction and the vertical direction through comparison. The calculation process is as follows:
Gx(x,y)＝H(x+1,y)-H(x-1,y)
Gy(x,y)＝H(x,y+1)-H(x,y-1)
ω(x,y)＝tan-1(Gy(x,y)/Gx(x,y))
Gx(x,y)，Gy(x, y) are respectively the horizontal gradient and the vertical gradient of the current pixel point; g (x, y) is the gradient amplitude of the current pixel point; and omega (x, y) is the gradient direction of the current pixel point.
Limiting the gradient direction in the range of [0, pi ], equally dividing the gradient direction into 9 intervals, and then using the amplitude in the direction to make a weighted statistical histogram for each interval.
Step 1.3, intra-block normalization. Dividing the image into cells and blocks with the same size; unlike HOG which extracts 36-dimensional features per cell, FHOG extracts 31-dimensional features per cell and normalizes the 31-dimensional feature vectors.
The purpose of normalization is to further eliminate the influence of illumination and shadow:
wherein,and viThe gradient values of each pixel after the initial and normalization are respectively expressed, and are a constant, and the value is 0.001.
And step 1.4, acquiring FHOG characteristics. And (3) normalizing each Block by adopting a sliding window search method according to the normalization method in the step 1.3, and connecting the feature vectors in each normalized Block in series to obtain the final feature vector.
3. An image pedestrian recognition method based on FHOG-LBPH characteristics is characterized in that:
the LBPH characteristic extraction in the second step comprises the following specific steps:
step 2.1, calculating an LBP characteristic image (LBP value of a rotation invariant equivalence mode) of the image;
step 2.2, partitioning the LBP characteristic image;
step 2.3, calculating a histogram of each regional characteristic image, and normalizing the histogram;
and 2.4, sequentially arranging the histograms of the regional characteristic images in a row according to the spatial sequence of the blocks to form an LBPH characteristic vector.
4. An image pedestrian recognition method based on FHOG-LBPH characteristics is characterized in that:
FHOG-LBPH characteristic fusion in the third step comprises the following specific steps:
the dimensionality of the FHOG features is high, a flat surface is omitted, the robustness of the classifier is poor, and therefore the pedestrian is detected by a method of fusing the selected FHOG features and the LBPH features, and the fusing method is as follows: and (4) connecting the FHOG characteristic obtained in the first step and the second step with the LBPH characteristic in series to form a characteristic of an image to detect the pedestrian. Wherein the cell size is 8 multiplied by 8 when the FHOG characteristic is extracted, the padding size is 1 multiplied by 1, and the FHOG characteristic of the whole image is 10 multiplied by 18 multiplied by 31 dimensions; when extracting LBPH characteristics, the cell size is 8 multiplied by 8, each block area counts 9 histograms, and 12 multiplied by 20 multiplied by 9 dimensional characteristics are extracted from the whole image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710006490.XA CN106682641A (en) | 2017-01-05 | 2017-01-05 | Pedestrian identification method based on image with FHOG- LBPH feature |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710006490.XA CN106682641A (en) | 2017-01-05 | 2017-01-05 | Pedestrian identification method based on image with FHOG- LBPH feature |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106682641A true CN106682641A (en) | 2017-05-17 |
Family
ID=58849813
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710006490.XA Pending CN106682641A (en) | 2017-01-05 | 2017-01-05 | Pedestrian identification method based on image with FHOG- LBPH feature |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106682641A (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108388920A (en) * | 2018-03-01 | 2018-08-10 | 福州大学 | A kind of Copy of ID Card detection method of fusion HOG and LBPH features |
| CN108960029A (en) * | 2018-03-23 | 2018-12-07 | 北京交通大学 | A kind of pedestrian diverts one's attention behavioral value method |
| CN109063619A (en) * | 2018-07-25 | 2018-12-21 | 东北大学 | A kind of traffic lights detection method and system based on adaptive background suppression filter and combinations of directions histogram of gradients |
| CN109086801A (en) * | 2018-07-06 | 2018-12-25 | 湖北工业大学 | A kind of image classification method based on improvement LBP feature extraction |
| CN109377624A (en) * | 2018-11-23 | 2019-02-22 | 卢伟涛 | A kind of door intelligent opening system based on facial image identification |
| CN109740693A (en) * | 2019-01-18 | 2019-05-10 | 北京细推科技有限公司 | Data identification method and device |
| CN109740618A (en) * | 2019-01-14 | 2019-05-10 | 河南理工大学 | Automatic statistic method and device for online test paper scores based on FHOG features |
| CN110287990A (en) * | 2019-05-21 | 2019-09-27 | 山东大学 | Microalgae image classification method, system, device and storage medium |
| CN110866420A (en) * | 2018-08-28 | 2020-03-06 | 天津理工大学 | A 2D Deceptive Pedestrian Recognition Method Based on Light Field Camera, HOG, SVM |
| CN110866422A (en) * | 2018-08-28 | 2020-03-06 | 天津理工大学 | Two-dimensional false pedestrian recognition method based on combination of light field imaging and LBP (local binary pattern) and SVM (support vector machine) |
| CN111476271A (en) * | 2020-03-10 | 2020-07-31 | 杭州易现先进科技有限公司 | Icon identification method, device, system, computer equipment and storage medium |
| CN113505695A (en) * | 2021-07-09 | 2021-10-15 | 上海工程技术大学 | AEHAL characteristic-based track fastener state detection method |
| CN113815531A (en) * | 2021-10-28 | 2021-12-21 | 贵州省交通规划勘察设计研究院股份有限公司 | Vehicle-mounted device and detection method for intelligent detection of pedestrians on road at night |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104573707A (en) * | 2014-12-17 | 2015-04-29 | 安徽清新互联信息科技有限公司 | Vehicle license plate Chinese character recognition method based on multi-feature fusion |
| CN104866862A (en) * | 2015-04-27 | 2015-08-26 | 中南大学 | Strip steel surface area type defect identification and classification method |
| CN105447503A (en) * | 2015-11-05 | 2016-03-30 | 长春工业大学 | Sparse-representation-LBP-and-HOG-integration-based pedestrian detection method |
-
2017
- 2017-01-05 CN CN201710006490.XA patent/CN106682641A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104573707A (en) * | 2014-12-17 | 2015-04-29 | 安徽清新互联信息科技有限公司 | Vehicle license plate Chinese character recognition method based on multi-feature fusion |
| CN104866862A (en) * | 2015-04-27 | 2015-08-26 | 中南大学 | Strip steel surface area type defect identification and classification method |
| CN105447503A (en) * | 2015-11-05 | 2016-03-30 | 长春工业大学 | Sparse-representation-LBP-and-HOG-integration-based pedestrian detection method |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108388920A (en) * | 2018-03-01 | 2018-08-10 | 福州大学 | A kind of Copy of ID Card detection method of fusion HOG and LBPH features |
| CN108960029A (en) * | 2018-03-23 | 2018-12-07 | 北京交通大学 | A kind of pedestrian diverts one's attention behavioral value method |
| CN109086801A (en) * | 2018-07-06 | 2018-12-25 | 湖北工业大学 | A kind of image classification method based on improvement LBP feature extraction |
| CN109063619A (en) * | 2018-07-25 | 2018-12-21 | 东北大学 | A kind of traffic lights detection method and system based on adaptive background suppression filter and combinations of directions histogram of gradients |
| CN110866420A (en) * | 2018-08-28 | 2020-03-06 | 天津理工大学 | A 2D Deceptive Pedestrian Recognition Method Based on Light Field Camera, HOG, SVM |
| CN110866422A (en) * | 2018-08-28 | 2020-03-06 | 天津理工大学 | Two-dimensional false pedestrian recognition method based on combination of light field imaging and LBP (local binary pattern) and SVM (support vector machine) |
| CN109377624A (en) * | 2018-11-23 | 2019-02-22 | 卢伟涛 | A kind of door intelligent opening system based on facial image identification |
| CN109740618A (en) * | 2019-01-14 | 2019-05-10 | 河南理工大学 | Automatic statistic method and device for online test paper scores based on FHOG features |
| CN109740618B (en) * | 2019-01-14 | 2022-11-04 | 河南理工大学 | Test paper score automatic statistical method and device based on FHOG characteristics |
| CN109740693A (en) * | 2019-01-18 | 2019-05-10 | 北京细推科技有限公司 | Data identification method and device |
| CN109740693B (en) * | 2019-01-18 | 2021-05-18 | 北京细推科技有限公司 | Data identification method and device |
| CN110287990A (en) * | 2019-05-21 | 2019-09-27 | 山东大学 | Microalgae image classification method, system, device and storage medium |
| CN111476271A (en) * | 2020-03-10 | 2020-07-31 | 杭州易现先进科技有限公司 | Icon identification method, device, system, computer equipment and storage medium |
| CN111476271B (en) * | 2020-03-10 | 2023-07-21 | 杭州易现先进科技有限公司 | Icon identification method, device, system, computer equipment and storage medium |
| CN113505695A (en) * | 2021-07-09 | 2021-10-15 | 上海工程技术大学 | AEHAL characteristic-based track fastener state detection method |
| CN113815531A (en) * | 2021-10-28 | 2021-12-21 | 贵州省交通规划勘察设计研究院股份有限公司 | Vehicle-mounted device and detection method for intelligent detection of pedestrians on road at night |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106682641A (en) | Pedestrian identification method based on image with FHOG- LBPH feature | |
| CN108520226B (en) | A pedestrian re-identification method based on body decomposition and saliency detection | |
| Hou et al. | People counting and human detection in a challenging situation | |
| CN105512683B (en) | Target localization method and device based on convolutional neural network | |
| US9639748B2 (en) | Method for detecting persons using 1D depths and 2D texture | |
| CN103034852B (en) | The detection method of particular color pedestrian under Still Camera scene | |
| CN103310194B (en) | Pedestrian based on crown pixel gradient direction in a video shoulder detection method | |
| Grundmann et al. | 3D shape context and distance transform for action recognition | |
| Ye et al. | A two-stage real-time YOLOv2-based road marking detector with lightweight spatial transformation-invariant classification | |
| CN107316031A (en) | The image characteristic extracting method recognized again for pedestrian | |
| CN104166841A (en) | Rapid detection identification method for specified pedestrian or vehicle in video monitoring network | |
| CN105022982A (en) | Hand motion identifying method and apparatus | |
| Monteiro et al. | Vision-based pedestrian detection using haar-like features | |
| CN101996308A (en) | Human face identification method and system and human face model training method and system | |
| Kim et al. | Autonomous vehicle detection system using visible and infrared camera | |
| CN106557740B (en) | A Recognition Method of Oil Depot Targets in Remote Sensing Images | |
| Masmoudi et al. | Vision based system for vacant parking lot detection: Vpld | |
| Ye et al. | Overhead ground wire detection by fusion global and local features and supervised learning method for a cable inspection robot | |
| Hu et al. | Depth sensor based human detection for indoor surveillance | |
| CN104063682A (en) | Pedestrian detection method based on edge grading and CENTRIST characteristic | |
| CN105447873A (en) | RGB-D significant object detection method based on random forest learning | |
| Hariyono et al. | Motion segmentation using optical flow for pedestrian detection from moving vehicle | |
| CN119169393B (en) | Photovoltaic module big data AI analysis and diagnosis system | |
| Jindal et al. | Landmark points detection in the case of human facial tracking and detection | |
| CN108985216B (en) | Pedestrian head detection method based on multivariate logistic regression feature fusion |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170517 |