CN107247926A - A kind of human body detecting method and device - Google Patents
A kind of human body detecting method and device Download PDFInfo
- Publication number
- CN107247926A CN107247926A CN201710359529.6A CN201710359529A CN107247926A CN 107247926 A CN107247926 A CN 107247926A CN 201710359529 A CN201710359529 A CN 201710359529A CN 107247926 A CN107247926 A CN 107247926A
- Authority
- CN
- China
- Prior art keywords
- cloud
- data
- cloud data
- cluster
- peak
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Human Computer Interaction (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The embodiment of the invention discloses a kind of human body detecting method and device, to solve parallel overlapping standing, the problem of accuracy is not high when carrying out human testing before and after having two people even many people in the prior art.This method includes：For the first cloud data cluster to be detected, the multiple first sub- cloud data clusters are divided into according to default rule；Determine each local peak in the first cloud data cluster, if the projector distance of two local peaks in the horizontal plane is less than given threshold, the first sub- cloud data cluster in the corresponding region of the first sub- cloud data cluster where this two local peaks is merged into second point cloud aggregate of data, otherwise, it determines the second point cloud aggregate of data comprising the local peak；According to the HOG values of second point cloud aggregate of data correspondence image, whether the image for determining second point cloud aggregate of data is human body image.This mode for determining human body, can more distinguish the situation of parallel overlapping standing before and after two people even many people, improve the accuracy of human testing.
Description
Technical field
The present invention relates to technical field of computer vision, more particularly to a kind of human body detecting method and device.
Background technology
At present, human testing is the study hotspot of computer vision field, and the identification and feature extraction for human body have very
High scientific research and commercial value, have broad application prospects.Traditional human testing is based primarily upon monocular camera acquisition
What the video data that two dimensional image or monocular camera are obtained was carried out.Due to the single biography in monocular camera or monocular-camera
Sensor obtains the limited in one's ability of information, and characteristics of human body, which extracts, relies primarily on the color and Gradient Features of image, the precision of detection and
Speed is all relatively limited.
With the appearance of depth camera, such as binocular camera, structure light camera and TOF camera, what depth camera was obtained
Cloud data can directly obtain the three-dimensional information of specified point in image, and the information of acquisition is more enriched, and compensate for traditional list
The deficiency of mesh camera or monocular-camera, improves the precision of human testing.
In the prior art, depth camera is when carrying out human testing, first using the consistent (RANdom of random sampling
SAmple Consensus, RANSAC) algorithm determines cloud data and the deletion of ground level.Secondly using Euclidean distance cluster pair
Remaining cloud data is clustered, and primarily determines that the cloud data cluster that there may be human body.Then it is true by projection histogram
The problem of vertical Wave crest and wave trough solves one people of the class of many people one and multiclass that cluster is produced, determines to there may be the point cloud number of human body again
According to cluster.Finally supported using histograms of oriented gradients (Histograms Of Oriented Gradients, HOG) features training
Vector machine (Support Vector Machine, SVM) grader carries out human testing.But when environment is complicated, two even
The parallel overlapping class of many people one and multiclass one stood, Wave crest and wave trough solution cluster generation is established by projection histogram before and after many people
During the problem of people, the projection of two people or many people on XOY plane is located in same projection histogram, so as to cause people's physical examination
The accuracy of survey is not high.
The content of the invention
The present invention provides a kind of human body detecting method and device, to solve in the prior art when having before two people even many people
Afterwards during parallel overlapping standing, when carrying out human testing, the problem of accuracy is not high.
To reach above-mentioned purpose, the embodiment of the invention discloses a kind of human body detecting method, methods described includes：
For the first cloud data cluster primarily determined that as human body to be detected, by the first cloud data cluster according to pre-
If regular partition be the multiple first sub- cloud data clusters；
According to the peak in each first sub- cloud data cluster, each local highest in the first cloud data cluster is determined
Point, if the projector distance of two local peaks in the horizontal plane is less than the first threshold of setting, by this two local highests
The first sub- cloud data cluster in the corresponding region of the first sub- cloud data cluster where point merges into second point cloud aggregate of data, such as
The distance of the projection in the horizontal plane of the local peak of fruit and other any projections of local peak in the horizontal plane is not small
In the first threshold of setting, according to the local peak, it is determined that the second point cloud aggregate of data comprising the local peak；
For each second point cloud aggregate of data, the target direction gradient Nogata of the second point cloud aggregate of data correspondence image is obtained
Scheme HOG values, according to the forecast model of training in advance, determine whether the corresponding image of the second point cloud aggregate of data is human figure
Picture.
Further, it is determined that the process of the first cloud data cluster includes：
Obtain the cloud data of image to be detected；
The cloud data of ground level is determined in the cloud data of described image using the consistent RANSAC algorithms of random sampling,
The cloud data of the ground level is deleted in the cloud data of described image；
Clustered in the cloud data after deleting the cloud data of ground level according to default cluster radius, after cluster
Obtained each classification is defined as the first cloud data cluster.
Further, it is described that the first cloud data cluster is divided into multiple first son point cloud numbers according to default rule
Before cluster, methods described also includes：
Recognize the highs and lows in each first cloud data cluster；
If peak threshold value of the peak higher than setting in the first cloud data cluster, or minimum point would be less than setting most
Low spot threshold value, then delete the first cloud data cluster.
Further, it is described that the first cloud data cluster is divided into multiple first son point cloud numbers according to default rule
Include according to cluster：
By the first cloud data cluster according to default length and width by the first cloud data cluster according to default
Regular partition be the multiple first sub- cloud data clusters, wherein the default length and width be less than human body cube it is corresponding
Length and width.
Further, the peak in the sub- cloud data cluster of the basis each first, is determined in the first cloud data cluster
Each local peak include：
For the peak in each first sub- cloud data cluster, judge that the peak in the first sub- cloud data cluster is
Peak in no the first sub- cloud data cluster for predetermined number adjacent thereto, if it is, by the first sub- cloud data
Peak in cluster is used as the local peak in the first cloud data cluster.
Further, if the projection of the local peak in the horizontal plane and other any local peaks are in level
The distance of projection on face is all not less than the first threshold of setting, according to the local peak, it is determined that including the local peak
Second point cloud aggregate of data includes：
According to the local peak and default cylinder radius, cylinder is determined, by the institute in the cylinder
There is cloud data to be defined as second point cloud aggregate of data.
Further, before the target HOG values for obtaining the second point cloud aggregate of data correspondence image, methods described is also wrapped
Include：
For each second point cloud aggregate of data, the quantity of the point cloud in the second point cloud aggregate of data is recognized；
The second point cloud aggregate of data that the quantity of a cloud is less than default Second Threshold is deleted.
Further, it is described to be directed to each second point cloud aggregate of data, obtain the mesh of the second point cloud aggregate of data correspondence image
Mark HOG values include：
For each second point cloud aggregate of data, the peak in the second point cloud aggregate of data is recognized, according to the peak
And the proportionate relationship of default height and the width, determine the human region in the corresponding image of the second point cloud aggregate of data；
Obtain the target HOG values of the human region.
The embodiment of the invention discloses a kind of human body detection device, described device includes：
Division module, for for the first cloud data cluster primarily determined that as human body to be detected, by described first point
Cloud aggregate of data is divided into the multiple first sub- cloud data clusters according to default rule；
Determining module, for the peak in each first sub- cloud data cluster, is determined in the first cloud data cluster
Each local peak, will if the projector distance of two local peaks in the horizontal plane is less than the first threshold of setting
The first sub- cloud data cluster in the corresponding region of the first sub- cloud data cluster where this two local peaks merges into the
Two cloud data clusters, if the projection of local peak in the horizontal plane and other any throwings of local peak in the horizontal plane
The distance of shadow is all not less than the first threshold of setting, according to the local peak, it is determined that including the second point of the local peak
Cloud aggregate of data；
Detection module, for for each second point cloud aggregate of data, obtaining the mesh of the second point cloud aggregate of data correspondence image
Histograms of oriented gradients HOG values are marked, according to the forecast model of training in advance, the corresponding image of the second point cloud aggregate of data are determined
Whether it is human body image.
Further, the determining module, is additionally operable to obtain the cloud data of image to be detected；Using random sampling one
RANSAC algorithms are caused to determine the cloud data of ground level in the cloud data of described image, in the cloud data of described image
Delete the cloud data of the ground level；According to default cluster radius in the cloud data after deleting the cloud data of ground level
Clustered, each classification obtained after cluster is defined as the first cloud data cluster.
Further, described device also includes：
First identification removing module, for recognizing the highs and lows in each first cloud data cluster；If the
Peak threshold value of the peak higher than setting in some cloud aggregates of data, or minimum point then will less than the minimum point threshold value of setting
The first cloud data cluster is deleted.
Further, the division module, specifically for by the first cloud data cluster according to default length and width
The first cloud data cluster is divided into the multiple first sub- cloud data clusters by degree according to default rule, wherein described default
Length and width is less than the corresponding length and width of human body cube.
Further, the determining module, specifically for for the peak in each first sub- cloud data cluster, judging
Peak in the first sub- cloud data cluster whether be predetermined number adjacent thereto the first sub- cloud data cluster in most
High point, if it is, regarding the peak in the first sub- cloud data cluster as the local peak in the first cloud data cluster.
Further, the determining module, specifically according to the local peak and default cylinder radius, it is determined that circle
All cloud datas in the cylinder are defined as second point cloud aggregate of data by cylinder.
Further, described device also includes：
Second identification removing module, for for each second point cloud aggregate of data, recognizing in the second point cloud aggregate of data
The quantity of point cloud；The second point cloud aggregate of data that the quantity of a cloud is less than default Second Threshold is deleted.
Further, the detection module, specifically for for each second point cloud aggregate of data, recognizing the second point cloud number
According to the peak in cluster, according to the peak and the proportionate relationship of default height and the width, the second point cloud number is determined
According to the human region in the corresponding image of cluster；Obtain the target HOG values of the human region.
The embodiment of the invention discloses a kind of human body detecting method and device, methods described includes：At the beginning of to be detected
Step is defined as the first cloud data cluster of human body, and the first cloud data cluster is divided into multiple first according to default rule
Sub- cloud data cluster；According to the peak in each first sub- cloud data cluster, each office in the first cloud data cluster is determined
Portion's peak, if the projector distance of two local peaks in the horizontal plane is less than the first threshold of setting, by two offices
The first sub- cloud data cluster in the corresponding region of the first sub- cloud data cluster where portion's peak merges into second point cloud number
According to cluster, if the distance of the projection of local peak in the horizontal plane and other any projections of local peak in the horizontal plane
All not less than the first threshold of setting, according to the local peak, it is determined that the second point cloud aggregate of data comprising the local peak；
For each second point cloud aggregate of data, the target direction histogram of gradients HOG values of the second point cloud aggregate of data correspondence image are obtained,
According to the forecast model of training in advance, determine whether the corresponding image of the second point cloud aggregate of data is human body image.Due to
In the embodiment of the present invention, for the first cloud data cluster primarily determined that as human body to be detected, with the first cloud data cluster
Local peak projector distance in the horizontal plane clustered again, determine second point cloud aggregate of data.By obtaining second point
The target HOG values of cloud aggregate of data correspondence image, determine whether the corresponding image of the second point cloud aggregate of data is human body image.This
The mode for determining human body is planted, the situation of parallel overlapping standing before and after two people even many people can be more distinguished, improve human testing
Accuracy.
Brief description of the drawings
In order to illustrate more clearly about the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing
There is the accompanying drawing used required in technology description to be briefly described, it should be apparent that, drawings in the following description are only this
Some embodiments of invention, for those of ordinary skill in the art, on the premise of not paying creative work, can be with
Other accompanying drawings are obtained according to these accompanying drawings.
Fig. 1 is a kind of human testing process schematic that the embodiment of the present invention 1 is provided；
Fig. 2 is a kind of human testing process schematic that the embodiment of the present invention 3 is provided；
Fig. 2A is a kind of first cloud data cluster schematic diagram provided in an embodiment of the present invention；
Fig. 3 is a kind of human testing process schematic that the embodiment of the present invention 7 is provided；
Fig. 4 is a kind of human detection result schematic diagram provided in an embodiment of the present invention；
Fig. 5 is a kind of human body detection device structure chart that the embodiment of the present invention 1 is provided；
Fig. 6 is a kind of human body detection device structure chart provided in an embodiment of the present invention.
Embodiment
In order to improve the accuracy of the human testing in parallel overlapping standing before and after two people even many people, the present invention is implemented
Example provides a kind of human body detecting method and device.
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear, complete
Site preparation is described, it is clear that described embodiment is only a part of embodiment of the invention, rather than whole embodiments.It is based on
Embodiment in the present invention, it is every other that those of ordinary skill in the art are obtained under the premise of creative work is not made
Embodiment, belongs to the scope of protection of the invention.
Embodiment 1：
Fig. 1 is a kind of human testing process schematic provided in an embodiment of the present invention, and the process comprises the following steps：
S101：For the first cloud data cluster primarily determined that as human body to be detected, by the first cloud data cluster
The multiple first sub- cloud data clusters are divided into according to default rule.
Human body detecting method provided in an embodiment of the present invention can apply to electronic equipment, and the electronic equipment can be deep
Spend camera.The the first cloud data cluster primarily determined that as human body to be detected is preserved in the electronic equipment, and is preserved
The first cloud data cluster is divided into the default rule of the multiple first sub- cloud data clusters.
In embodiments of the present invention, electronic equipment preserve the first cloud data cluster in cloud data based on coordinate
In, be the plane constituted using X-axis and Y-axis as ground level i.e. horizontal plane, Z axis is built perpendicular to X/Y plane.For other implementations
The first cloud data cluster in example or cloud data in other cloud data clusters are also based on the such of this coordinate system.
The first cloud data cluster is divided into the multiple first sub- cloud data clusters by electronic equipment according to default rule.
For the first cloud data cluster of human body can be one or multiple for be detected the primarily determining that, each first point
Include multiple cloud datas in cloud aggregate of data.The first cloud data cluster is being divided into multiple according to default rule
Can first determine the corresponding scope of all first cloud data clusters in X-direction and Y direction during one sub- cloud data cluster
Each boundary point, according to each boundary point, determines regional extent, according to default length and width, will in the regional extent
All first cloud data clusters are divided into the multiple first sub- cloud data clusters.Then determine again in each first cloud data cluster
Comprising the first sub- cloud data cluster.The default length is division distance interval of the first cloud data cluster in X-axis,
The default width is division distance interval of the first cloud data cluster in Y-axis.The cubical length of general human body is big
In 30cm, width is more than 15cm.Arbitrarily walked about in space rotation in view of human body, it is possible that human body with depth camera not
Parallel situation, the width can also be greater than 30cm.The default length and width can be identical or not
Together, when the length and width is identical, for example, can be that the length is 10cm, the width is also 10cm.
S102：According to the peak in each first sub- cloud data cluster, each office in the first cloud data cluster is determined
Portion's peak, if the projector distance of two local peaks in the horizontal plane is less than the first threshold of setting, by two offices
The first sub- cloud data cluster in the corresponding region of the first sub- cloud data cluster where portion's peak merges into second point cloud number
According to cluster, if projection and any other local peak projector distances in the horizontal plane of the local peak in level are not
Less than the first threshold of setting, according to the local peak, it is determined that the second point cloud aggregate of data comprising the local peak.
In embodiments of the present invention, one or more points cloud data, Mei Gedian are included in each first sub- cloud data cluster
Cloud data have the information of its corresponding three-dimensional coordinate, and the plane that the cloud data is constituted using X-axis and Y-axis is ground level i.e. level
Face, the value on Z axis is the corresponding height value of the cloud data, recognizes that the peak in each first sub- cloud data cluster exists
The cloud data for maximum on Z axis, according to the peak in every sub- cloud data cluster, it is determined that each first cloud data
Each local peak in cluster.Local peak in each first cloud data cluster can be one or multiple,
Local peak may correspond to the head of human body.
First threshold is preset with electronic equipment, electronic equipment recognizes the throwing of two local peaks in the horizontal plane
Shadow distance, and judge whether the distance is less than the first threshold of setting, if it is, thinking this two local peaks pair
The the first sub- cloud data cluster answered belongs to the cloud data cluster of same human body, then by first where this two local peaks
The first sub- cloud data cluster in the corresponding region of sub- cloud data cluster merges into second point cloud aggregate of data.
If the projection of the local peak in the horizontal plane and other any projections of local peak in the horizontal plane
Distance all not less than setting first threshold, then according to the local peak, it is determined that including the second point of the local peak
Cloud aggregate of data.Each second point cloud aggregate of data is considered a human body.
Peak in the sub- cloud data cluster of the basis each first, determines each part in the first cloud data cluster
During peak, can be specifically, by the peak in each first sub- cloud data cluster, first cloud to be defined as by comparing
Each local peak in aggregate of data.Each local peak it will be projected to horizontal plane.
If the projection of the local peak in the horizontal plane and other any projections of local peak in the horizontal plane
Distance all not less than setting first threshold, it is described according to the local peak, it is determined that including the second of the local peak
Cloud data cluster, can be specifically according to the local peak and default length and width, cube to be determined, by position
It is defined as second point cloud aggregate of data in all cloud datas in the cube.The local peak can be located at the cube
Summit on.As it is assumed that each part peak is the head of human body, so preferably, the local peak is located at this cube
On the center of the upper surface of body.In each cube have cloud data to constitute second point cloud aggregate of data to be considered one more
Human body.The length of general human body is more than 30cm, and the length of human body mentioned here refers to left shoulder edge to the right shoulder side of human body
The distance of edge, the first threshold can be 30cm.
S103：For each second point cloud aggregate of data, the target HOG values of the second point cloud aggregate of data correspondence image are obtained,
According to the forecast model of training in advance, determine whether the corresponding image of the second point cloud aggregate of data is human body image.
For each second point cloud aggregate of data, the target HOG values of the second point cloud aggregate of data correspondence image are obtained, according to pre-
The forecast model first trained, determines whether the corresponding image of the second point cloud aggregate of data is human body image.Obtain this second
During the target HOG values of cloud data cluster correspondence image, the second point cloud aggregate of data can be done to YOZ planes or XOZ planes
Two-dimensional projection, obtains the target HOG values of the image after second point cloud aggregate of data projection.In the prediction mould according to training in advance
Type, can be specifically using HOG features instruction when whether determine the corresponding image of the second point cloud aggregate of data be human body image
Practice SVM classifier and carry out human bioequivalence, determine whether the image after the second point cloud aggregate of data projection is human body image.
Due in embodiments of the present invention, for the first cloud data cluster primarily determined that as human body to be detected, with
The projector distance of local peak in the horizontal plane in some cloud aggregates of data is clustered again, determines second point cloud aggregate of data.
By obtaining the target HOG values of second point cloud aggregate of data correspondence image, determine that the corresponding image of the second point cloud aggregate of data is
No is human body image.This mode for determining human body, can more distinguish the situation of parallel overlapping standing before and after two people even many people, carry
The high accuracy of human testing.
Embodiment 2：
On the basis of above-described embodiment, in embodiments of the present invention, determining the process of the first cloud data cluster includes：
Obtain the cloud data of image to be detected；
The cloud data of ground level is determined in the cloud data of described image using RANSAC algorithms, in described image
The cloud data of the ground level is deleted in cloud data；
Clustered in the cloud data after deleting the cloud data of ground level according to default cluster radius, after cluster
Obtained each classification is defined as the first cloud data cluster.
In embodiments of the present invention, electronic equipment obtains the cloud data of image to be detected, when obtaining described image
Cloud data after, determine the cloud data of ground level in the cloud data of described image using RANSAC algorithms, and
The cloud data of the ground level is deleted in the cloud data of described image.By the left point after the cloud data for deleting ground level
Cloud data, are clustered according to default cluster radius, and each classification obtained after cluster is defined as into the first cloud data cluster.Institute
Stating cluster radius can be relevant with the dense degree of a cloud, and point cloud is more intensive, and cluster radius is smaller, and point cloud is more sparse, cluster half
Footpath is bigger.
When electronic equipment is depth camera, electronic equipment is when obtaining the image and cloud data of scene to be detected, tool
Body can after depth camera is fixed, gather target scene cloud data.Depth camera collection can be a frame figure
Picture or one section of video.If one section of video, then depth camera can be obtained in each two field picture in this section of video
Cloud data.
When determining the cloud data of ground level in the cloud data of described image using RANSAC algorithms, specifically may be used
To be, in the cloud data of the image, the plane is considered ground level by estimation containing there is the most plane of the quantity of a cloud.
Can also be by choosing the cloud data on arbitrary 3 ground, and the three-dimensional coordinate included with this 3 cloud datas manually
Information, by formula AX+BY+CZ+D=0, calculate the equation parameter of 3 cloud data corresponding flats, that is, calculate A, B,
C and D value.For all cloud datas in image, each cloud data is calculated to European between the plane calculated
Distance, then each distance and the distance threshold of setting are compared, if in the corresponding scope of distance threshold, by the cloud
Data are classified as " intra-office point ", otherwise, are classified as " point not in the know ", and the ground level being made up of " intra-office point " is corrected, finally obtained
The cloud data of all ground levels.The distance threshold can be 15cm.
After the cloud data of ground level is determined, the cloud data of ground level is deleted in the cloud data of described image
Remove, the cloud data of deletion is not involved in each ensuing operation.
After the cloud data of the ground level is deleted, by the remaining cloud after the cloud data of ground level is deleted
Clustered in data according to default cluster radius, each classification obtained after cluster is defined as the first cloud data cluster.Tool
Body can be, according to default cluster radius, if between two adjacent cloud datas in remaining cloud data
Distance is less than the default cluster radius, and two cloud datas are defined as into same category, traveled through successively all remaining
All remaining cloud datas are carried out clustering processing, each classification obtained after cluster are defined as into first point by cloud data
Cloud aggregate of data.The cluster radius is relevant with the dense degree for putting cloud, and cloud data is more intensive, and cluster radius is smaller, puts cloud number
According to more sparse, cluster radius is bigger, generally, and the cluster radius can be 12cm.
When being clustered, any one in cloud data point cloud can be designated as pi, will be all in cloud data
Point cloud is deposited in cloud data cluster Q to be sorted, for the certain point cloud p in Qi, it is drawn off being stored to the C1 subclass gathered in C
In, identification and piThe distance between be less than cluster radius point cloud pi, by each piTake out and be stored in C1 subclass from Q；So
Another cloud is taken out in remaining cloud of Q again afterwards, is stored in the C2 subclass that C gathers, above-mentioned cluster half is repeated
The identification in footpath, is repeated in performing said process, until Q is sky, so that each subclass in sorted C set is obtained,
Each subclass is defined as each corresponding type, corresponding each first cloud data cluster is obtained.
Typically, cloud data, the cloud data of ground level of human body are included in the cloud data in described image to be detected
And other non-human cloud datas.The cloud data of ground level is deleted in cloud data in described image to be detected,
Clustered for remaining cloud data according to default cluster radius, human body is included in each first cloud data after cluster
Cloud data, but it is also possible that including non-human cloud data.One may be included in each first cloud data cluster
Or multiple human bodies.Therefore the first cloud data cluster after cluster is referred to as to primarily determine that the first cloud data cluster for human body, can
To carry out follow-up step for the first cloud data cluster, the second cloud data comprising human body is determined more accurately out
Cluster, and then detect human body.
Embodiment 3：
In order to avoid wall, the interference of the too high or too low object such as desk further improves the accuracy of human testing,
On the basis of the various embodiments described above, in embodiments of the present invention, it is described by the first cloud data cluster according to default rule
Then it is divided into before the multiple first sub- cloud data clusters, methods described also includes：
Recognize the highs and lows in each first cloud data cluster；
If peak threshold value of the peak higher than setting in the first cloud data cluster, or minimum point would be less than setting most
Low spot threshold value, then delete the first cloud data cluster.
In embodiments of the present invention, peak threshold value and minimum point threshold value are preserved in electronic equipment.For to be detected
The first cloud data cluster for human body is primarily determined that, the highs and lows in each first cloud data cluster are recognized, judged
Whether the peak in each first cloud data cluster is more than the peak threshold value of setting, if it is, the peak is corresponding
The first cloud data cluster delete.Judge whether the minimum point in each first cloud data cluster is less than the minimum point threshold of setting
Value, if it is, the corresponding first cloud data cluster of the minimum point is deleted.
The the first cloud data cluster deleted is not involved in each ensuing operation.The peak threshold value of the setting can be
Than larger numerical value, such as can be 2.3 meters, 2.5 meters, the minimum point threshold value of setting can be smaller numerical value, for example
It can be 1 meter, 1.3 meters etc., since it typically be considered that the height of human body is no more than peak threshold value, and be not less than minimum point threshold
Value.If the peak in the first cloud data cluster is higher than peak threshold value, such as 2.3 meters, then it is assumed that first cloud number
May be metope according to cluster, so being deleted.If the minimum point in the first cloud data cluster is less than minimum point threshold value, such as
1.3 meters, then it is assumed that the first cloud data cluster may be the first non-human cloud data cluster such as desk, chair, thus by its
Delete.
In order to improve the accuracy of human testing, for the first cloud data cluster primarily determined that as human body to be detected,
Can also delete the cloud data of the peak threshold value higher than setting in each first cloud data cluster, and less than setting
Minimum point threshold value cloud data delete.The cloud data of deletion is not involved in each ensuing operation.
Fig. 2 is a kind of human testing process schematic that the embodiment of the present invention 3 is provided, and the process comprises the following steps：
S201：Obtain the cloud data of image to be detected.
S202：The cloud data of ground level is determined in the cloud data of described image using RANSAC algorithms, described
The cloud data of the ground level is deleted in the cloud data of image.
S203：Clustered in the cloud data after deleting the cloud data of ground level according to default cluster radius, will
The each classification obtained after cluster is defined as the first cloud data cluster.
S204：For the first cloud data cluster primarily determined that as human body to be detected, each first cloud data is recognized
Highs and lows in cluster, if peak threshold value of the peak higher than setting in the first cloud data cluster, or it is minimum
Point then deletes the first cloud data cluster less than the minimum point threshold value of setting.
S205：The remaining first cloud data cluster is divided into the multiple first sub- cloud datas according to default rule
Cluster.
S206：According to the peak in each first sub- cloud data cluster, each office in the first cloud data cluster is determined
Portion's peak, if the projector distance of two local peaks in the horizontal plane is less than the first threshold of setting, by two offices
The first sub- cloud data cluster in the corresponding region of the first sub- cloud data cluster where portion's peak merges into second point cloud number
According to cluster, if the distance of the projection of local peak in the horizontal plane and other any projections of local peak in the horizontal plane
All not less than the first threshold of setting, according to the local peak, it is determined that the second point cloud aggregate of data comprising the local peak.
S207：For each second point cloud aggregate of data, the target direction ladder of the second point cloud aggregate of data correspondence image is obtained
Histogram HOG values are spent, according to the forecast model of training in advance, determine whether the corresponding image of the second point cloud aggregate of data is people
Body image.
Embodiment 4：
In order to further improve the accuracy of human testing, it is described by the first cloud data cluster according to default rule
Being divided into the multiple first sub- cloud data clusters includes：
By the first cloud data cluster according to default length and width by the first cloud data cluster according to default
Regular partition be the multiple first sub- cloud data clusters, wherein the default length and width be less than human body cube it is corresponding
Length and width.
In embodiments of the present invention, the first cloud data cluster is being divided into multiple first sons according to default rule
During cloud data cluster, can be specifically, by the first cloud data cluster according to default length and width by described first
Cloud data cluster is divided into the multiple first sub- cloud data clusters, wherein the default length and width is less than human body cube pair
The length and width answered.The default length is division distance interval of the first cloud data cluster in X-axis, described default
Width for the first cloud data cluster in Y-axis division distance interval.The cubical length of general human body is more than 30cm, wide
Degree is more than 15cm.The default length and width can be identical or difference, when the length and width is identical
When, for example can be that the length is 10cm, the width is also 10cm.
Embodiment 5：
In each part by the peak in each first sub- cloud data cluster, being defined as in the first cloud data cluster most
During high point, the inaccurate of follow-up human testing is likely to result in, in order to further improve the accuracy of human testing.In above-mentioned each reality
Apply on the basis of example, in embodiments of the present invention, the peak in the sub- cloud data cluster of the basis each first determines first
Each local peak in cloud data cluster includes：
For the peak in each first sub- cloud data cluster, judge that the peak in the first sub- cloud data cluster is
Peak in no the first sub- cloud data cluster for setting quantity adjacent thereto, if it is, by the first sub- cloud data
Peak in cluster is used as the local peak in the first cloud data cluster.
In embodiments of the present invention, for each cloud data cluster, the multiple first sub- points are divided into according to default rule
Cloud aggregate of data, the position of each first sub- cloud data cluster is clear, therefore is clear that and the first son point cloud
The sub- cloud data cluster of adjacent each first of aggregate of data.In the peak in each first sub- cloud data cluster, it is determined that
During each local peak in the first cloud data cluster, can be specifically, in each first sub- cloud data cluster
Peak, judge the peak in the first sub- cloud data cluster whether be predetermined number adjacent thereto first son point cloud number
According to the peak in cluster, if it is, regarding the peak in the first sub- cloud data cluster as the office in the first cloud data cluster
Portion's peak.The predetermined number can be 4, or 8.
Fig. 2A is a kind of first cloud data cluster schematic diagram provided in an embodiment of the present invention, it can be seen from Fig. 2A this
Some cloud aggregates of data XOY plane perspective view, comprising 12 the first sub- cloud data clusters in the first cloud data cluster, mark
Number be 1-12.For the 6th the first sub- cloud data cluster, the first son point cloud adjacent with the 6th the first sub- cloud data cluster
Aggregate of data has 8, and label is respectively 1,2,3,5,7,9,10,11.When the default quantity is 8, then the 6th the is judged
Whether peak in one sub- cloud data cluster is peak in the sub- cloud data cluster of all first adjacent thereto, if
It is to regard the peak in the 6th the first sub- cloud data cluster as the local peak in the first cloud data cluster.If institute
It is 4 to state predetermined number, then whether judge the peak in the 6th the first sub- cloud data cluster is adjacent thereto any 4
Peak in first sub- cloud data cluster, if it is, regarding the peak in the 6th the first sub- cloud data cluster as first
Local peak in cloud data cluster.Selecting the 4 first son point clouds adjacent with the 6th first sub- cloud data cluster
Can be any selection 4 in 1,2,3,5,7,9,10,11 during aggregate of data, preferably, selection and the 6th the first sub- point
First sub- cloud data cluster of the adjacent upper and lower, left and right of cloud aggregate of data, i.e., the first sub- cloud data marked as 2,5,7,10
Cluster.
Local peak in each first cloud data cluster can be one or multiple, each part highest
Point may correspond to the head of human body.
Embodiment 6：
On the basis of the various embodiments described above, in embodiments of the present invention, if the local peak is in the horizontal plane
Projection and other any projections of local peak in the horizontal plane distance all not less than the first threshold of setting, according to this
Local peak, it is determined that including comprising the local peak second point cloud aggregate of data：
According to the local peak and default cylinder radius, cylinder is determined, by the institute in the cylinder
There is cloud data to be defined as second point cloud aggregate of data.
In embodiments of the present invention, according to the local peak, it is determined that the second point cloud number comprising the local peak
It can be specifically, according to the local peak and default cylinder radius, to determine cylinder during according to cluster, the circle will be located at
All cloud datas in cylinder are defined as second point cloud aggregate of data.The diameter of the cylinder can be identical with first threshold,
Can also be different, when the diameter of cylinder is identical with first threshold, the first threshold is if 30cm, then described default
Cylinder radius is 15cm.The default cylinder diameter is more than default cluster radius.
Embodiment 7：
In order to further improve the accuracy of human testing, on the basis of the various embodiments described above, in the present invention is implemented,
Before the target HOG values for obtaining the second point cloud aggregate of data correspondence image, methods described also includes：
For each second point cloud aggregate of data, the quantity of the point cloud in the second point cloud aggregate of data is recognized；
The second point cloud aggregate of data that the quantity of a cloud is less than default Second Threshold is deleted.
In embodiments of the present invention, default Second Threshold is preserved in electronic equipment.When the second cloud data is determined
After cluster, each second point cloud aggregate of data can be directed to, the quantity of the point cloud in the second point cloud aggregate of data is recognized, judges the point
Whether the quantity of cloud is less than default Second Threshold, if it is, the quantity of a cloud to be less than to the second point of default Second Threshold
Cloud aggregate of data is deleted.
When the point cloud quantity in the second point cloud aggregate of data is less than default Second Threshold, people may be corresponded to and lifted
Arm, then delete the corresponding second point cloud aggregate of data of the arm, have no effect on the detection to human body.
Fig. 3 is a kind of human testing process schematic that the embodiment of the present invention 7 is provided, and the process comprises the following steps：
S301：For the first cloud data cluster primarily determined that as human body to be detected, each first cloud data is recognized
Highs and lows in cluster, if peak threshold value of the peak higher than setting in the first cloud data cluster, or it is minimum
Point then deletes the first cloud data cluster less than the minimum point threshold value of setting.
S302：The remaining first cloud data cluster is divided into the multiple first sub- cloud datas according to default rule
Cluster.
S303：According to the peak in each first sub- cloud data cluster, each office in the first cloud data cluster is determined
Portion's peak, if the projector distance of two local peaks in the horizontal plane is less than the first threshold of setting, by two offices
The first sub- cloud data cluster in the corresponding region of the first sub- cloud data cluster where portion's peak merges into second point cloud number
According to cluster, if the distance of the projection of local peak in the horizontal plane and other any projections of local peak in the horizontal plane
All not less than the first threshold of setting, according to the local peak, it is determined that the second point cloud aggregate of data comprising the local peak.
S304：For each second point cloud aggregate of data, the quantity of the point cloud in the second point cloud aggregate of data is recognized, cloud will be put
Quantity be less than default Second Threshold second point cloud aggregate of data delete.
S305：For remaining each second point cloud aggregate of data, the target of the second point cloud aggregate of data correspondence image is obtained
Histograms of oriented gradients HOG values, according to the forecast model of training in advance, determine that the corresponding image of the second point cloud aggregate of data is
No is human body image.
Embodiment 8：
In order to further improve the accuracy of human testing, on the basis of the various embodiments described above, in the present invention is implemented,
Described to be directed to each second point cloud aggregate of data, obtaining the target HOG values of the second point cloud aggregate of data correspondence image includes：
For each second point cloud aggregate of data, the peak in the second point cloud aggregate of data is recognized, according to the peak
And the proportionate relationship of default height and the width, determine the human region in the corresponding image of the second point cloud aggregate of data；
Obtain the target HOG values of the human region.
In embodiments of the present invention, after each second point cloud aggregate of data is determined, if X-axis and putting down that Y-axis is constituted
Face is ground level, to YOZ planes or XOZ planes the second point cloud aggregate of data can be done into two-dimensional projection, after identification projection
Peak in second point cloud aggregate of data, according to the peak and the proportionate relationship of default height and the width, it is determined that described
Human region in the corresponding image of second point cloud aggregate of data, obtains the target HOG values of the human region.Each peak
It is considered the head of people, peak is higher, then corresponding human region is bigger.
It is determined that can also be specifically that identification is every during human region in the corresponding image of the second point cloud aggregate of data
Peak in individual second point cloud aggregate of data, according to the peak and the proportionate relationship of default height and the width, determines square
Shape region, extends default length to depth camera direction by the rectangular area, forms 3-D view, by the 3-D view to
YOZ planes or XOZ planes do two-dimensional projection, and the two-dimensional projection is defined as into the corresponding image of the second point cloud aggregate of data
In human region.The default length is 10cm.
Fig. 4 is a kind of human detection result schematic diagram provided in an embodiment of the present invention, and each human body in the Fig. 4 is by it
Corresponding annulus is entangled, and illustrates that the human body is detected.
Fig. 5 is a kind of human body detection device structure chart that the embodiment of the present invention 1 is provided, and described device includes：
Division module 51, for for the first cloud data cluster primarily determined that as human body to be detected, by described first
Cloud data cluster is divided into the multiple first sub- cloud data clusters according to default rule；
Determining module 52, for the peak in each first sub- cloud data cluster, determines the first cloud data cluster
In each local peak, if the projector distance of two local peaks in the horizontal plane is less than the first threshold of setting,
The first sub- cloud data cluster in the corresponding region of the first sub- cloud data cluster where this two local peaks is merged into
Second point cloud aggregate of data, if the projection of local peak in the horizontal plane and other any local peaks are in the horizontal plane
The distance of projection is all not less than the first threshold of setting, according to the local peak, it is determined that including the second of the local peak
Cloud data cluster；
Detection module 53, for for each second point cloud aggregate of data, obtaining the second point cloud aggregate of data correspondence image
Target direction HOG values, according to the forecast model of training in advance, determine whether the corresponding image of the second point cloud aggregate of data is people
Body image.
The determining module 52, is additionally operable to obtain the cloud data of image to be detected；Using RANSAC algorithms described
The cloud data of ground level is determined in the cloud data of image, the point of the ground level is deleted in the cloud data of described image
Cloud data；Clustered in the cloud data after deleting the cloud data of ground level according to default cluster radius, after cluster
Obtained each classification is defined as the first cloud data cluster.
Fig. 6 is a kind of human body detection device structure chart provided in an embodiment of the present invention, on the basis of Fig. 5, described device
Also include：
First identification removing module 61, for recognizing the highs and lows in each first cloud data cluster；If
Peak threshold value of the peak higher than setting in first cloud data cluster, or minimum point are less than the minimum point threshold value of setting, then
The first cloud data cluster is deleted.
The division module 51, specifically for the first cloud data cluster is incited somebody to action into described according to default length and width
First cloud data cluster is divided into the multiple first sub- cloud data clusters according to default rule, wherein the default length and width
Degree is less than the corresponding length and width of human body cube.
The determining module 52, specifically for for the peak in each first sub- cloud data cluster, judge this first
Peak in sub- cloud data cluster whether be predetermined number adjacent thereto the first sub- cloud data cluster in peak, such as
Fruit is to regard the peak in the first sub- cloud data cluster as the local peak in the first cloud data cluster.
The determining module 52, specifically according to the local peak and default cylinder radius, determines cylinder, will
All cloud datas in the cylinder are defined as second point cloud aggregate of data.
Described device also includes：
Second identification removing module 62, is additionally operable to be directed to each second point cloud aggregate of data, recognizes the second point cloud aggregate of data
In point cloud quantity；The second point cloud aggregate of data that the quantity of a cloud is less than default Second Threshold is deleted.
The detection module 53, specifically for for each second point cloud aggregate of data, recognizing in the second point cloud aggregate of data
Peak, according to the peak and the proportionate relationship of default height and the width, determine the second point cloud aggregate of data pair
Human region in the image answered；Obtain the target HOG values of the human region.
The embodiment of the invention discloses a kind of human body detecting method and device, methods described includes：At the beginning of to be detected
Step is defined as the first cloud data cluster of human body, and the first cloud data cluster is divided into multiple first according to default rule
Sub- cloud data cluster；According to the peak in each first sub- cloud data cluster, each office in the first cloud data cluster is determined
Portion's peak, if the projector distance of two local peaks in the horizontal plane is less than the first threshold of setting, by two offices
The first sub- cloud data cluster in the corresponding region of the first sub- cloud data cluster where portion's peak merges into second point cloud number
According to cluster, if the distance of the projection of local peak in the horizontal plane and other any projections of local peak in the horizontal plane
All not less than the first threshold of setting, according to the local peak, it is determined that the second point cloud aggregate of data comprising the local peak；
For each second point cloud aggregate of data, the target direction histogram of gradients HOG values of the second point cloud aggregate of data correspondence image are obtained,
According to the forecast model of training in advance, determine whether the corresponding image of the second point cloud aggregate of data is human body image.Due to
In the embodiment of the present invention, for the first cloud data cluster primarily determined that as human body to be detected, with the first cloud data cluster
Local peak projector distance in the horizontal plane clustered again, determine second point cloud aggregate of data.By obtaining second point
The target HOG values of cloud aggregate of data correspondence image, determine whether the corresponding image of the second point cloud aggregate of data is human body image.This
The mode for determining human body is planted, the situation of parallel overlapping standing before and after two people even many people can be more distinguished, improve human testing
Accuracy.
For systems/devices embodiment, because it is substantially similar to embodiment of the method, so the comparison of description is simple
Single, the relevent part can refer to the partial explaination of embodiments of method.
It should be noted that herein, such as first and second or the like relational terms are used merely to a reality
Body or an operation make a distinction with another entity or another operation, and not necessarily require or imply these entities
Or there is any this actual relation or order between operation.
It should be understood by those skilled in the art that, embodiments herein can be provided as method, system or computer program
Product.Therefore, the application can use the reality of complete hardware embodiment, complete Application Example or connected applications and hardware aspect
Apply the form of example.Moreover, the application can be used in one or more computers for wherein including computer usable program code
The computer program production that usable storage medium is implemented on (including but is not limited to magnetic disk storage, CD-ROM, optical memory etc.)
The form of product.
The application is the flow with reference to method, equipment (system) and computer program product according to the embodiment of the present application
Figure and/or block diagram are described.It should be understood that can be by every first-class in computer program instructions implementation process figure and/or block diagram
Journey and/or the flow in square frame and flow chart and/or block diagram and/or the combination of square frame.These computer programs can be provided
The processor of all-purpose computer, special-purpose computer, Embedded Processor or other programmable data processing devices is instructed to produce
A raw machine so that produced by the instruction of computer or the computing device of other programmable data processing devices for real
The device for the function of being specified in present one flow of flow chart or one square frame of multiple flows and/or block diagram or multiple square frames.
These computer program instructions, which may be alternatively stored in, can guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works so that the instruction being stored in the computer-readable memory, which is produced, to be included referring to
Make the manufacture of device, the command device realize in one flow of flow chart or multiple flows and/or one square frame of block diagram or
The function of being specified in multiple square frames.
These computer program instructions can be also loaded into computer or other programmable data processing devices so that in meter
Series of operation steps is performed on calculation machine or other programmable devices to produce computer implemented processing, thus in computer or
The instruction performed on other programmable devices is provided for realizing in one flow of flow chart or multiple flows and/or block diagram one
The step of function of being specified in individual square frame or multiple square frames.
Although having been described for the preferred embodiment of the application, those skilled in the art once know basic creation
Property concept, then can make other change and modification to these embodiments.So, appended claims are intended to be construed to include excellent
Select embodiment and fall into having altered and changing for the application scope.
Obviously, those skilled in the art can carry out the essence of various changes and modification without departing from the application to the application
God and scope.So, if these modifications and variations of the application belong to the scope of the application claim and its equivalent technologies
Within, then the application is also intended to comprising including these changes and modification.
Claims (16)
1. a kind of human body detecting method, it is characterised in that methods described includes：
For the first cloud data cluster primarily determined that as human body to be detected, by the first cloud data cluster according to default
Regular partition is the multiple first sub- cloud data clusters；
According to the peak in each first sub- cloud data cluster, each local peak in the first cloud data cluster is determined,
If the projector distance of two local peaks in the horizontal plane is less than the first threshold of setting, by this two local peak institutes
The corresponding region of the first sub- cloud data cluster in the first sub- cloud data cluster merge into second point cloud aggregate of data, if office
The distance of the projection in the horizontal plane of portion's peak and other any projections of local peak in the horizontal plane is all not less than setting
Fixed first threshold, according to the local peak, it is determined that the second point cloud aggregate of data comprising the local peak；
For each second point cloud aggregate of data, the target direction histogram of gradients of the second point cloud aggregate of data correspondence image is obtained
HOG values, according to the forecast model of training in advance, determine whether the corresponding image of the second point cloud aggregate of data is human body image.
2. the method as described in claim 1, it is characterised in that determining the process of the first cloud data cluster includes：
Obtain the cloud data of image to be detected；
The cloud data of ground level is determined in the cloud data of described image using the consistent RANSAC algorithms of random sampling, in institute
The cloud data of the ground level is deleted in the cloud data for stating image；
Clustered, will be obtained after cluster according to default cluster radius in the cloud data after deleting the cloud data of ground level
Each classification be defined as the first cloud data cluster.
3. the method as described in claim 1, it is characterised in that it is described by the first cloud data cluster according to default rule
It is divided into before the multiple first sub- cloud data clusters, methods described also includes：
Recognize the highs and lows in each first cloud data cluster；
If peak threshold value of the peak higher than setting in the first cloud data cluster, or minimum point would be less than the minimum point of setting
Threshold value, then delete the first cloud data cluster.
4. the method as described in claim 1, it is characterised in that it is described by the first cloud data cluster according to default rule
Being divided into the multiple first sub- cloud data clusters includes：
By the first cloud data cluster according to default length and width by the first cloud data cluster according to default rule
The multiple first sub- cloud data clusters are then divided into, wherein the default length and width is less than the corresponding length of human body cube
And width.
5. the method as described in claim 1, it is characterised in that the highest in the sub- cloud data cluster of the basis each first
Point, determines that each local peak in the first cloud data cluster includes：
For the peak in each first sub- cloud data cluster, judge the peak in the first sub- cloud data cluster whether be
Peak in first sub- cloud data cluster of predetermined number adjacent thereto, if it is, by the first sub- cloud data cluster
Peak be used as the local peak in the first cloud data cluster.
6. the method as described in claim 1, it is characterised in that if the projection of the local peak in the horizontal plane is with appointing
Anticipate other projections of local peak in the horizontal plane distance all not less than setting first threshold, according to the local highest
Point, it is determined that including comprising the local peak second point cloud aggregate of data：
According to the local peak and default cylinder radius, cylinder is determined, by the institute in the cylinder a little
Cloud data are defined as second point cloud aggregate of data.
7. the method as described in claim 1 or 3, it is characterised in that the acquisition second point cloud aggregate of data correspondence image
Before target HOG values, methods described also includes：
For each second point cloud aggregate of data, the quantity of the point cloud in the second point cloud aggregate of data is recognized；
The second point cloud aggregate of data that the quantity of a cloud is less than default Second Threshold is deleted.
8. the method as described in claim 1, it is characterised in that described to be directed to each second point cloud aggregate of data, obtain this second
The target HOG values of cloud data cluster correspondence image include：
For each second point cloud aggregate of data, the peak in the second point cloud aggregate of data is recognized, according to the peak and in advance
If height and the width proportionate relationship, determine the human region in the corresponding image of the second point cloud aggregate of data；
Obtain the target HOG values of the human region.
9. a kind of human body detection device, it is characterised in that described device includes：
Division module, for for the first cloud data cluster primarily determined that as human body to be detected, by first cloud number
According to cluster the multiple first sub- cloud data clusters are divided into according to default rule；
Determining module, for the peak in each first sub- cloud data cluster, is determined every in the first cloud data cluster
Individual local peak, if the projector distance of two local peaks in the horizontal plane is less than the first threshold of setting, by this two
The first sub- cloud data cluster in the corresponding region of the first sub- cloud data cluster where individual local peak merges into second point
Cloud aggregate of data, if the projection of local peak in the horizontal plane and other any projections of local peak in the horizontal plane
Distance is all not less than the first threshold of setting, according to the local peak, it is determined that the second point cloud number comprising the local peak
According to cluster；
Detection module, for for each second point cloud aggregate of data, obtaining the target side of the second point cloud aggregate of data correspondence image
To histogram of gradients HOG values, according to the forecast model of training in advance, whether the corresponding image of the second point cloud aggregate of data is determined
For human body image.
10. device as claimed in claim 9, it is characterised in that the determining module, is additionally operable to obtain image to be detected
Cloud data；The point cloud number of ground level is determined in the cloud data of described image using the consistent RANSAC algorithms of random sampling
According to deleting the cloud data of the ground level in the cloud data of described image；After the cloud data of ground level is deleted
Clustered in cloud data according to default cluster radius, each classification obtained after cluster is defined as the first cloud data
Cluster.
11. device as claimed in claim 9, it is characterised in that described device also includes：
First identification removing module, for recognizing the highs and lows in each first cloud data cluster；If first point
Peak in cloud aggregate of data is higher than the peak threshold value of setting, or minimum point is less than the minimum point threshold value of setting, then by this
Some cloud aggregates of data are deleted.
12. device as claimed in claim 9, it is characterised in that the division module, specifically for by first cloud number
The first cloud data cluster is divided into the multiple first sub- points according to default rule according to default length and width according to cluster
Cloud aggregate of data, wherein the default length and width is less than the corresponding length and width of human body cube.
13. device as claimed in claim 9, it is characterised in that the determining module, specifically for for each first sub- point
Peak in cloud aggregate of data, judges whether the peak in the first sub- cloud data cluster is predetermined number adjacent thereto
Peak in first sub- cloud data cluster, if it is, regarding the peak in the first sub- cloud data cluster as first cloud
Local peak in aggregate of data.
14. device as claimed in claim 9, it is characterised in that the determining module, specifically according to the local peak and
Default cylinder radius, determines cylinder, and all cloud datas in the cylinder are defined as into second point cloud number
According to cluster.
15. the device as described in claim 9 or 11, it is characterised in that described device also includes：
Second identification removing module, for for each second point cloud aggregate of data, recognizing the point cloud in the second point cloud aggregate of data
Quantity；The second point cloud aggregate of data that the quantity of a cloud is less than default Second Threshold is deleted.
16. device as claimed in claim 9, it is characterised in that the detection module, specifically for for each second point cloud
Aggregate of data, recognizes the peak in the second point cloud aggregate of data, according to the peak and the ratio of default height and the width
Relation, determines the human region in the corresponding image of the second point cloud aggregate of data；Obtain the target HOG of the human region
Value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710359529.6A CN107247926B (en) | 2017-05-19 | 2017-05-19 | A kind of human body detecting method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710359529.6A CN107247926B (en) | 2017-05-19 | 2017-05-19 | A kind of human body detecting method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107247926A true CN107247926A (en) | 2017-10-13 |
| CN107247926B CN107247926B (en) | 2019-09-17 |
Family
ID=60017347
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710359529.6A Active CN107247926B (en) | 2017-05-19 | 2017-05-19 | A kind of human body detecting method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107247926B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109214348A (en) * | 2018-09-19 | 2019-01-15 | 北京极智嘉科技有限公司 | A kind of obstacle detection method, device, equipment and storage medium |
| CN109740628A (en) * | 2018-12-03 | 2019-05-10 | 深圳市华讯方舟太赫兹科技有限公司 | Point cloud clustering method, image processing equipment and the device with store function |
| CN112183381A (en) * | 2020-09-30 | 2021-01-05 | 深兰人工智能(深圳)有限公司 | Method and device for detecting driving area of vehicle |
| CN112651266A (en) * | 2019-10-11 | 2021-04-13 | 北京百度网讯科技有限公司 | Pedestrian detection method and device |
| WO2022057804A1 (en) * | 2020-09-16 | 2022-03-24 | 青岛维感科技有限公司 | Distance monitoring method, device, and system, and storage medium |
| US11295133B2 (en) | 2019-08-28 | 2022-04-05 | Industrial Technology Research Institute | Interaction display method and interaction display system |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102930246A (en) * | 2012-10-16 | 2013-02-13 | 同济大学 | Indoor scene identifying method based on point cloud fragment division |
| EP2562688A2 (en) * | 2011-08-22 | 2013-02-27 | Samsung Electronics Co., Ltd. | Method of Separating Object in Three Dimensional Point Cloud |
| CN104298998A (en) * | 2014-09-28 | 2015-01-21 | 北京理工大学 | 3D point cloud data processing method |
| CN104408445A (en) * | 2014-12-19 | 2015-03-11 | 吉林纪元时空动漫游戏科技股份有限公司 | Automatic real-time human body detecting method |
| CN104751146A (en) * | 2015-04-13 | 2015-07-01 | 中国科学技术大学 | Indoor human body detection method based on 3D (three-dimensional) point cloud image |
| CN105674991A (en) * | 2016-03-29 | 2016-06-15 | 深圳市华讯方舟科技有限公司 | Robot positioning method and device |
-
2017
- 2017-05-19 CN CN201710359529.6A patent/CN107247926B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2562688A2 (en) * | 2011-08-22 | 2013-02-27 | Samsung Electronics Co., Ltd. | Method of Separating Object in Three Dimensional Point Cloud |
| CN102930246A (en) * | 2012-10-16 | 2013-02-13 | 同济大学 | Indoor scene identifying method based on point cloud fragment division |
| CN104298998A (en) * | 2014-09-28 | 2015-01-21 | 北京理工大学 | 3D point cloud data processing method |
| CN104408445A (en) * | 2014-12-19 | 2015-03-11 | 吉林纪元时空动漫游戏科技股份有限公司 | Automatic real-time human body detecting method |
| CN104751146A (en) * | 2015-04-13 | 2015-07-01 | 中国科学技术大学 | Indoor human body detection method based on 3D (three-dimensional) point cloud image |
| CN105674991A (en) * | 2016-03-29 | 2016-06-15 | 深圳市华讯方舟科技有限公司 | Robot positioning method and device |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109214348A (en) * | 2018-09-19 | 2019-01-15 | 北京极智嘉科技有限公司 | A kind of obstacle detection method, device, equipment and storage medium |
| CN109740628A (en) * | 2018-12-03 | 2019-05-10 | 深圳市华讯方舟太赫兹科技有限公司 | Point cloud clustering method, image processing equipment and the device with store function |
| US11295133B2 (en) | 2019-08-28 | 2022-04-05 | Industrial Technology Research Institute | Interaction display method and interaction display system |
| CN112651266A (en) * | 2019-10-11 | 2021-04-13 | 北京百度网讯科技有限公司 | Pedestrian detection method and device |
| WO2022057804A1 (en) * | 2020-09-16 | 2022-03-24 | 青岛维感科技有限公司 | Distance monitoring method, device, and system, and storage medium |
| CN112183381A (en) * | 2020-09-30 | 2021-01-05 | 深兰人工智能(深圳)有限公司 | Method and device for detecting driving area of vehicle |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107247926B (en) | 2019-09-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107247926B (en) | A kind of human body detecting method and device | |
| CN106650640B (en) | Negative obstacle detection method based on laser radar point cloud local structure characteristics | |
| CN106204547B (en) | The method that rod-shaped atural object spatial position is automatically extracted from Vehicle-borne Laser Scanning point cloud | |
| Zhang et al. | Filtering airborne LiDAR data by embedding smoothness-constrained segmentation in progressive TIN densification | |
| US9754160B2 (en) | Method and device for detecting gathering of objects based on stereo vision as well as non-transitory computer-readable medium | |
| CN112070769B (en) | Layered point cloud segmentation method based on DBSCAN | |
| CN104715471B (en) | Target locating method and its device | |
| CN103258203B (en) | The center line of road extraction method of remote sensing image | |
| CN104657968B (en) | Automatic vehicle-mounted three-dimensional laser point cloud facade classification and outline extraction method | |
| CN109035330A (en) | Cabinet approximating method, equipment and computer readable storage medium | |
| CN105068918B (en) | A kind of page method of testing and device | |
| CN104050709B (en) | A kind of three dimensional image processing method and electronic equipment | |
| CN105205866A (en) | Dense-point-cloud-based rapid construction method of urban three-dimensional model | |
| CN108416785A (en) | Topological dividing method towards enclosure space and device | |
| JP7156515B2 (en) | Point cloud annotation device, method and program | |
| CN108920765A (en) | A kind of hypothesis planar fit method based on building three-dimensional line segment model | |
| CN103632167B (en) | Monocular vision space recognition method under class ground gravitational field environment | |
| CN105701798A (en) | Point cloud extraction method and device for columnar object | |
| CN108225334A (en) | A kind of localization method and device based on three-dimensional live-action data | |
| CN112785596B (en) | Dot cloud picture bolt segmentation and height measurement method based on DBSCAN clustering | |
| CN112085675B (en) | Depth image denoising method, foreground segmentation method and human motion monitoring method | |
| CN106022266A (en) | Target tracking method and target tracking apparatus | |
| Arachchige et al. | Automatic processing of mobile laser scanner point clouds for building facade detection | |
| CN111870952A (en) | Height map generation method, device, equipment and storage medium | |
| KR101092250B1 (en) | Apparatus and method for object segmentation from range image |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |