CN113689458B - 2D shooting track path calculation method and device - Google Patents
2D shooting track path calculation method and device Download PDFInfo
- Publication number
- CN113689458B CN113689458B CN202111251929.8A CN202111251929A CN113689458B CN 113689458 B CN113689458 B CN 113689458B CN 202111251929 A CN202111251929 A CN 202111251929A CN 113689458 B CN113689458 B CN 113689458B
- Authority
- CN
- China
- Prior art keywords
- shooting
- picture
- moving
- track
- acquiring
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
Abstract
The invention discloses a method and a device for calculating a 2D shooting track path, wherein the method comprises the following steps: acquiring a photo set according to continuous overlapped shooting of cameras, inputting the photo set into a preset target detection model, and acquiring a target detection result of each photo; calculating a homography matrix between every two adjacent photos in the photo set by adopting a preset feature point matching algorithm; calculating a corresponding 2D shooting track according to the homography matrix between each two adjacent photos; and calculating the position of the target in the picture in the shooting scene according to the 2D shooting track and the target detection result. The method and the device are combined with the target detection model, the 2D shooting track in the large scene is calculated by adopting the homography matrix, the target in the large scene can be quickly positioned, the time cost and the labor cost are reduced, and the efficiency of searching the target is improved.
Description
Technical Field
The invention relates to the technical field of positioning, in particular to a 2D shooting track path calculation method and device.
Background
In actual life, people often need to identify and judge relative positions of a plurality of targets in a large scene so as to enable workers to quickly lock the positions of articles to be found, for example, for commodities on a long shelf with a large business surpass, if the whole shelf is continuously photographed in advance, and then the targets in the pictures are identified and positioned and the relative positions of the pictures are distinguished, accurate positioning of the commodities on the long shelf can be provided for the workers when the workers find certain commodities next time.
At present, target searching and positioning in a plurality of large scenes are searched by manual screening, and searching for a certain type of required target and the position thereof from a large number of static targets usually consumes a large amount of manpower and material resources, and the whole searching process consumes a long time and has a low efficiency.
Disclosure of Invention
The invention aims to provide a method and a device for calculating a 2D shooting track path, which are used for solving the problem of low target identification and positioning efficiency in a large scene in the prior art.
In order to achieve the above object, the present invention provides a 2D shooting trajectory path calculation method, including:
acquiring a photo set according to continuous overlapped shooting of cameras, inputting the photo set into a preset target detection model, and acquiring a target detection result of each photo;
calculating a homography matrix between every two adjacent photos in the photo set by adopting a preset feature point matching algorithm;
calculating a corresponding 2D shooting track according to the homography matrix between each two adjacent photos;
and calculating the position of the target in the picture in the shooting scene according to the 2D shooting track and the target detection result.
Preferably, the calculating the homography matrix between every two adjacent photos in the photo set by using a preset feature point matching algorithm includes:
extracting any two adjacent photos in the photo set, and acquiring the characteristics of the two adjacent photos, namely the first photo characteristic and the second photo characteristic;
and acquiring a homography matrix of the second photo in the two adjacent photos according to the preset feature matching algorithm, and sequentially calculating the homography matrix between each two adjacent photos in the photo set.
Preferably, the calculating the corresponding 2D shooting track according to the homography matrix between each two adjacent photos includes:
and taking the central point of the second picture in the two adjacent pictures as a reference point, acquiring the coordinate of a corresponding point in the first picture in the two adjacent pictures through the homography matrix and the reference point, acquiring the moving shooting track of the second picture relative to the first picture in a 2D plane according to the coordinate of the reference point of the second picture and the coordinate of the corresponding point in the first picture, and iteratively acquiring the shooting track of each picture.
The present invention also provides a 2D shooting trajectory path calculation apparatus, including:
the detection module is used for acquiring a photo set according to continuous overlapped shooting of the cameras, inputting the photo set into a preset target detection model and acquiring a target detection result of each photo;
the first calculation module is used for calculating a homography matrix between every two adjacent photos in the photo set by adopting a preset feature point matching algorithm;
the second calculation module is used for calculating a corresponding 2D shooting track according to the homography matrix between each two adjacent photos;
and the positioning module is used for calculating the position of the target in the picture in the shooting scene according to the 2D shooting track and the target detection result.
Preferably, the first computing module is further configured to:
extracting any two adjacent photos in the photo set, and acquiring the characteristics of the two adjacent photos, namely the first photo characteristic and the second photo characteristic;
and acquiring a homography matrix of the second photo in the two adjacent photos according to the preset feature matching algorithm, and sequentially calculating the homography matrix between each two adjacent photos in the photo set.
Preferably, the second computing module is further configured to:
and taking the central point of the second picture in the two adjacent pictures as a reference point, acquiring the coordinate of a corresponding point in the first picture in the two adjacent pictures through the homography matrix and the reference point, acquiring the moving shooting track of the second picture relative to the first picture in a 2D plane according to the coordinate of the reference point of the second picture and the coordinate of the corresponding point in the first picture, and iteratively acquiring the shooting track of each picture. The present invention also provides a terminal device, including:
one or more processors;
a memory coupled to the processor for storing one or more programs;
when the one or more programs are executed by the one or more processors, the one or more processors implement the 2D photographing trajectory path calculation method as any one of the above.
The present invention also provides a computer-readable storage medium having stored thereon a computer program which is executed by a processor to implement the 2D photographing trajectory path calculation method according to any one of the above.
Compared with the prior art, the invention has the beneficial effects that:
the method is based on the combination of the 2D shooting track path calculation with overlapped shooting and the target detection model in continuous movement, can quickly position the target in the large scene, reduces the time cost and the labor cost for a worker to manually search the target position for many times in the large scene, and improves the searching efficiency for quickly positioning the required target in the large scene.
Compared with the prior art that a large image covering the whole scene is searched through one image, the shooting mode that all targets of the whole large scene are covered through continuously shooting a plurality of local images of the large scene is easier to realize. In reality, it is often difficult to cover a picture of a whole scene by taking a large image because the scene is too large, and even if the large image of the whole scene is obtained in other ways, the number of pixels occupied by the target in the scene in the large image is much smaller than that in a single local small image, so that the appearance of the target in the large image is difficult to clearly show, and meanwhile, the target detection model is difficult to detect a tiny target in the large image because the number of pixels occupied by the target in the large image is small. Therefore, the shooting mode of covering all targets of the whole large scene together by continuously shooting the local photos of the large scene is simpler to operate and easy to realize, and meanwhile, the accuracy of the target detection model for detecting the targets in a single picture can be ensured.
Drawings
In order to more clearly illustrate the technical solution of the present invention, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a schematic flow chart of a 2D shooting trajectory path calculation method according to an embodiment of the present invention;
fig. 2 is a schematic structural diagram of a 2D shooting trajectory path calculation apparatus according to an embodiment of the present invention;
fig. 3 is a schematic structural diagram of a computer terminal device according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
It should be understood that the step numbers used herein are for convenience of description only and are not intended as limitations on the order in which the steps are performed.
It is to be understood that the terminology used in the description of the invention herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used in the specification of the present invention and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.
The terms "comprises" and "comprising" indicate the presence of the described features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.
The term "and/or" refers to and includes any and all possible combinations of one or more of the associated listed items.
Referring to fig. 1, the present invention provides a method for calculating a 2D shooting trajectory path, including the following steps:
s101, acquiring a photo set according to continuous overlapped shooting of cameras, inputting the photo set into a preset target detection model, and acquiring a target detection result of each photo.
Specifically, various pictures containing the targets are collected according to the targets to be recognized, the targets to be recognized in the pictures are labeled in rectangular frames, the collected pictures and corresponding labeled data are combined to form a data set of a target detection model, the target detection model used in the embodiment is Cascade R-CNN, the backbone network is ResNeXt101, and the target detection model is trained and tested on the prepared data set to obtain the trained target detection model.
And carrying out target detection and identification in a large scene by using a target detection model based on deep learning, wherein the commonly used target detection models comprise Faster R-CNN, Mask R-CNN, Cascade R-CNN, Yolo series, SSD and the like.
1) The method comprises the steps of collecting pictures containing targets to be identified, carrying out rectangular frame labeling on the targets in the collected pictures, and then using the collected pictures and corresponding labeling data thereof for training a target detection model.
2) And detecting a target rectangular frame for each photo of a group of continuous moving pictures, and finally obtaining all targets detected in the group of photos.
And S102, calculating a homography matrix between every two adjacent photos in the photo set by adopting a preset feature point matching algorithm.
Specifically, the camera is continuously moved in the same shooting plane by holding the shooting equipment or manually controlling the shooting camera to shoot, namely, the distance of the camera relative to a target scene is kept constant all the time during shooting, so as to ensure that the content shot by the camera at each time has the same depth of field, meanwhile, after the current picture is shot, when the next picture is shot, the camera needs to be moved from the current shooting position, the moving direction of the camera comprises four simple directions, namely horizontal left, horizontal right, vertical upward and vertical downward, and at least 30% of overlapping areas of two adjacent pictures need to be controlled during continuous shooting, so that 2D shooting track path calculation in subsequent steps is facilitated.
In the super long goods shelves district of large-scale merchant, handheld camera is just to focusing with the goods shelves, then carries out continuous movement from one end of goods shelves to the other end and has overlapping to shoot, the shooting requirement that needs satisfy at the in-process of shooing: the content of the current shot picture and the previous picture needs to have a 30% overlapping area, the moving direction during each shooting is limited to four simple directions including up, down, left and right, the distance between the camera and the goods shelf is kept unchanged during moving shooting, and the camera moves in the same 2D plane all the time.
A group of local photos of the long shelf shot in the superstore in the previous step are numbered in sequence, and if the group of photos has n photos, the numbers are p in sequence1、p2、p3...pn. Then, SIFT features of each picture are sequentially extracted, feature matching is sequentially carried out on the SIFT features of two adjacent pictures, the matching mode includes but is not limited to violence pattern matching, homography matrix calculation is carried out on feature points matched in pairs, and homography matrix H (p) between two adjacent pictures is sequentially obtained1,p2)、H(p2,p3)、H(p3,p4) ... H(pn-1,pn) N-1 homography matrices, where H (p)n-1,pn) Representing a photograph pnConversion to photograph pn-1A homography matrix of the coordinate space of (a). The general mathematical form of the homography matrix is as follows:
wherein h is11To h33A total of 9 parameters represent 9 degree-of-freedom variables of the homography matrix.
S103, calculating a corresponding 2D shooting track according to the homography matrix between each two adjacent photos.
The method comprises the steps of taking the center point of the second picture of two adjacent pictures as a reference point, obtaining the coordinate of a corresponding point in the first picture through a homography matrix and the reference point, obtaining the moving shooting track of the second picture relative to the first picture in a 2D plane according to the coordinate of the reference point of the second picture and the coordinate of the corresponding point in the first picture, and iteratively obtaining the shooting track of each picture.
In a group of taken photos, a homography matrix is obtained by utilizing a previous photo and a current photo through calculation, a central point of the current photo is taken as a reference point, coordinates of a corresponding point of the reference point in the previous photo are obtained through the homography matrix, a shooting track of the current photo relative to the previous photo in a 2D moving plane can be obtained according to a relative coordinate relation between the reference point and the corresponding point, in conclusion, a 2D shooting track path of the whole group of taken photos can be obtained by calculating a shooting track between every two adjacent photos in a group of continuously moved photos with overlapped shooting, and a specific 2D shooting track calculation step between the two adjacent photos is as follows:
A. extracting Feature points from two adjacent photos, wherein the calculation modes for extracting the Feature points include but are not limited to SIFT (Scale Invariant Feature Transform) Feature points and SURF (Speeded Up Robust Features) Feature points and the like;
B. carrying out feature matching on the feature points obtained by the previous step of calculation, and then calculating by using the matched feature points to obtain a homography matrix H;
C. and performing coordinate transformation on the reference point of the current picture by using the homography matrix H obtained by the previous step to obtain the corresponding point of the reference point in the previous picture, and then calculating the position relation of the reference point and the corresponding point in the 2D plane to obtain the 2D photographing track between two adjacent pictures.
And S104, calculating the position of the target in the picture in the shooting scene according to the 2D shooting track and the target detection result.
Specifically, the target to be searched in the large scene is accurately positioned according to the target detection result and the 2D shooting track path.
a) Finding photos containing the target to be searched in a group of photo recognition results of continuous moving photographing, and determining the position of the target to be searched in the photos;
b) according to the calculated 2D shooting movement track, the shooting position of the picture containing the target to be searched in the large scene can be obtained.
c) And finally, the position of the target to be searched in the large scene can be accurately positioned according to the position of the target in the picture and the shooting position of the picture in the large scene.
Because the pictures taken by the same camera have the same resolution, namely the width w and the height h of the pictures are the same, and the coordinates of the center point of each picture are (w/2, h/2), the reference point of each picture is (w/2, h/2), the center point of the current picture is taken as the reference point A (x, y), and then the corresponding point B (x ', y') of the reference point A in the previous picture is obtained according to the homography matrix of the current picture relative to the adjacent previous picture. The method for obtaining the corresponding point B (x ', y') by using the homography matrix is shown as the following two formulas:
by the calculation from the reference point to the corresponding point, n-1 corresponding points B [ p ] among n pictures can be calculated1,p2]、B[p2,p3]、B[p3,p4]...B[pn-1,pn]In which B [ p ]n-1,pn]Representing a picture pnIs in picture pn-1And then calculating the 2D shooting movement track of the group of n photos in the same plane according to the n-1 corresponding points, wherein the track calculation mode is as follows:
1) initializing a symbol i equal to 1 and coordinates A (w/2, h/2) of a reference point (center point) of each picture, and selecting the (i + 1) th picture to a corresponding point B (x) of the previous picture, namely the ith picturei, yi)。
2) Calculating a corresponding point B [ p ] between the picture i and the picture i +1i, pi+1]A moving distance d in the horizontal and vertical directions with respect to the reference point A (w/2, h/2)xAnd dy. The corresponding calculation formula is as follows:
3) calculating the shooting track between the current two pictures: judgment of dxAnd dyThe absolute value of between, if dx| is greater than | dyIf the photographing track is mainly moving in the horizontal direction, then x is the sameiWhen the coordinate is larger than 0, the picture is taken by moving the picture horizontally to the right, the coordinate of the 2D track is recorded as (1, 0), otherwise, the picture is taken by moving the picture horizontally to the left, the coordinate of the 2D track is recorded as (-1, 0), and if | Dx| less than | dyIf the photographing track is mainly moving in the vertical direction, then when y isiWhen the coordinate is larger than 0, the picture is taken by moving vertically downwards, the coordinate of the 2D track is recorded as (0, 1), otherwise, the picture is taken by moving vertically upwards, and the coordinate of the 2D track is recorded as (0, -1).
4) And (3) adding 1 to the digital symbol i each time, repeating the steps (2) and (3) until i is equal to n-1, finishing the calculation of the 2D shooting movement track among n pictures, and finally obtaining n-1 2D movement shooting track coordinates, wherein the 2D movement shooting track coordinates jointly form a 2D movement shooting track path of the whole large scene.
The method comprises the steps of positioning the position of a required target in a large scene, inputting a collected photo into a constructed target detection model to obtain a target detection result of each photo, searching the position of the required target in the photo from the recognition results of the photos, positioning the position of the photo in the large scene according to a 2D moving photographing track path of the photo containing the required target in the whole large scene, and finally positioning the required target in the large scene according to the position of the photo in the large scene and the position of the required target in the photo.
The 2D shooting track path calculation method based on continuous movement and overlapped shooting can judge the relative positions of targets in different pictures through homography transformation among continuous shooting pictures, can be widely applied to identification and relative position judgment of a plurality of targets in a large scene, so that the time cost and the labor cost of people for searching the relative positions of some targets in the large scene are saved, and the target screening and positioning efficiency in the large scene is improved.
The method is based on the combination of the 2D shooting track path calculation with overlapped shooting and the target detection model in continuous movement, can quickly position the target in the large scene, reduces the time cost and the labor cost for a worker to manually search the target position for many times in the large scene, and improves the searching efficiency for quickly positioning the required target in the large scene.
Compared with the method of shooting a large image covering the whole scene, the shooting mode of jointly covering all targets of the whole large scene by continuously shooting a plurality of local photos of the large scene is easier to realize. In reality, it is often difficult to cover a picture of a whole scene by taking a large image because the scene is too large, and even if the large image of the whole scene is obtained in other ways, the number of pixels occupied by the target in the scene in the large image is much smaller than that in a single local small image, so that the appearance of the target in the large image is difficult to clearly show, and meanwhile, the target detection model is difficult to detect a tiny target in the large image because the number of pixels occupied by the target in the large image is small. Therefore, the shooting mode of continuously shooting a plurality of local pictures of the large scene to jointly cover all targets of the whole large scene in the method is simpler to operate and easy to realize, and meanwhile, the accuracy of the target detection model in detecting the targets in a single picture can be ensured.
Referring to fig. 2, the present invention provides a 2D shooting trajectory path calculating device, including:
the detection module 11 is configured to obtain a photo set according to continuous overlapping shooting of the cameras, input the photo set into a preset target detection model, and obtain a target detection result of each photo.
The first calculating module 12 is configured to calculate a homography matrix between every two adjacent photos in the photo set by using a preset feature point matching algorithm.
And the second calculating module 13 is configured to calculate a corresponding 2D shooting track according to the homography matrix between each two adjacent photos.
And the positioning module 14 is used for calculating the position of the target in the picture in the shooting scene according to the 2D shooting track and the target detection result.
For specific limitations of the 2D shooting trajectory path calculation means, reference may be made to the above limitations of the 2D shooting trajectory path calculation method, which will not be described herein again. The respective modules in the 2D photographing trajectory path calculating apparatus may be wholly or partially implemented by software, hardware, and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
Referring to fig. 3, an embodiment of the present invention provides a terminal device, including:
one or more processors;
a memory coupled to the processor for storing one or more programs;
when the one or more programs are executed by the one or more processors, the one or more processors implement the 2D photographing trajectory path calculation method as described above.
The processor is used for controlling the overall operation of the computer terminal equipment so as to complete all or part of the steps of the 2D shooting track path calculation method. The memory is used to store various types of data to support the operation at the computer terminal device, which data may include, for example, instructions for any application or method operating on the computer terminal device, as well as application-related data. The Memory may be implemented by any type of volatile or non-volatile Memory device or combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic Memory, flash Memory, magnetic disk, or optical disk.
In an exemplary embodiment, the computer terminal Device may be implemented by one or more Application Specific 1 integrated circuits (AS 1C), a Digital Signal Processor (DSP), a Digital Signal Processing Device (DSPD), a Programmable Logic Device (PLD), a Field Programmable Gate Array (FPGA), a controller, a microcontroller, a microprocessor or other electronic components, and is configured to perform the 2D shooting trajectory path calculation method and achieve technical effects consistent with the above-mentioned methods.
In another exemplary embodiment, there is also provided a computer-readable storage medium including program instructions, which when executed by a processor, implement the steps of the 2D photographing trajectory path calculation method in any one of the above embodiments. For example, the computer-readable storage medium may be the above-mentioned memory including program instructions that are executable by a processor of a computer terminal device to perform the above-mentioned 2D photographing trajectory path calculation method, and achieve technical effects consistent with the above-mentioned method.
While the foregoing is directed to the preferred embodiment of the present invention, it will be understood by those skilled in the art that various changes and modifications may be made without departing from the spirit and scope of the invention.
Claims (4)
1. A2D shooting track path calculation method is characterized by comprising the following steps:
acquiring a photo set according to continuous overlapped shooting of cameras, inputting the photo set into a preset target detection model, and acquiring a target detection result of each photo;
calculating a homography matrix between every two adjacent photos in the photo set by adopting a preset feature point matching algorithm;
extracting any two adjacent photos in the photo set, and acquiring the characteristics of the two adjacent photos, namely the first photo characteristic and the second photo characteristic;
acquiring a homography matrix of a second photo in the two adjacent photos according to the preset feature matching algorithm, and sequentially calculating the homography matrix between each two adjacent photos in the photo set;
calculating a corresponding 2D shooting track according to the homography matrix between each two adjacent photos;
taking the center point of the second picture in the two adjacent pictures as a reference point, acquiring the coordinate of a corresponding point in the first picture in the two adjacent pictures through the homography matrix and the reference point, acquiring the moving shooting track of the second picture relative to the first picture in a 2D plane according to the coordinate of the reference point of the second picture and the coordinate of the corresponding point in the first picture, and iteratively acquiring the shooting track of each picture;
wherein the acquiring of the moving shooting track of the second picture relative to the first picture in the 2D plane comprises: calculating the corresponding points B (x) of the first and second photographsi,yi) Is horizontal with respect to the reference pointDistance d of movement in directionxAnd a moving distance d in the vertical directiony(ii) a Wherein x isi,yiRespectively representing the abscissa and the ordinate of the corresponding point on the coordinate axis;
if the moving distance d in the horizontal directionxA moving distance d greater than the vertical directionyThen the moving shooting track is moving in the horizontal direction, and if x isiIf the x is larger than 0, moving the shooting track to be horizontal shooting to the rightiIf the shooting track is less than 0, the shooting track is moved to be horizontal and leftward shooting;
if the moving distance d in the horizontal directionxA moving distance d smaller than the vertical directionyThen the moving shooting track is moved in the vertical direction, and if yiIf the motion is larger than 0, the motion shooting track is vertical and downward motion shooting, and if y is larger than 0, the motion shooting track is vertical and downward motion shootingiIf the moving shooting track is smaller than 0, the moving shooting track is vertically upward moving shooting;
and calculating the position of the target in the picture in the shooting scene according to the 2D shooting track and the target detection result.
2. A 2D shooting trajectory path calculation apparatus, comprising:
the detection module is used for acquiring a photo set according to continuous overlapped shooting of the cameras, inputting the photo set into a preset target detection model and acquiring a target detection result of each photo;
the first calculation module is used for calculating a homography matrix between every two adjacent photos in the photo set by adopting a preset feature point matching algorithm;
extracting any two adjacent photos in the photo set, and acquiring the characteristics of the two adjacent photos, namely the first photo characteristic and the second photo characteristic;
acquiring a homography matrix of a second photo in the two adjacent photos according to the preset feature matching algorithm, and sequentially calculating the homography matrix between each two adjacent photos in the photo set;
the second calculation module is used for calculating a corresponding 2D shooting track according to the homography matrix between each two adjacent photos;
taking the center point of the second picture in the two adjacent pictures as a reference point, acquiring the coordinate of a corresponding point in the first picture in the two adjacent pictures through the homography matrix and the reference point, acquiring the moving shooting track of the second picture relative to the first picture in a 2D plane according to the coordinate of the reference point of the second picture and the coordinate of the corresponding point in the first picture, and iteratively acquiring the shooting track of each picture;
wherein the acquiring of the moving shooting track of the second picture relative to the first picture in the 2D plane comprises: calculating the corresponding points B (x) of the first and second photographsi,yi) A moving distance d in a horizontal direction with respect to the reference pointxAnd a moving distance d in the vertical directiony(ii) a Wherein x isi,yiRespectively representing the abscissa and the ordinate of the corresponding point on the coordinate axis;
if the moving distance d in the horizontal directionxA moving distance d greater than the vertical directionyThen the moving shooting track is moving in the horizontal direction, and if x isiIf the x is larger than 0, moving the shooting track to be horizontal shooting to the rightiIf the shooting track is less than 0, the shooting track is moved to be horizontal and leftward shooting;
if the moving distance d in the horizontal directionxA moving distance d smaller than the vertical directionyThen the moving shooting track is moved in the vertical direction, and if yiIf the motion is larger than 0, the motion shooting track is vertical and downward motion shooting, and if y is larger than 0, the motion shooting track is vertical and downward motion shootingiIf the moving shooting track is smaller than 0, the moving shooting track is vertically upward moving shooting;
and the positioning module is used for calculating the position of the target in the picture in the shooting scene according to the 2D shooting track and the target detection result.
3. A terminal device, comprising:
one or more processors;
a memory coupled to the processor for storing one or more programs;
when executed by the one or more processors, cause the one or more processors to implement the 2D photographing trajectory path calculation method according to claim 1.
4. A computer-readable storage medium on which a computer program is stored, the computer program implementing the 2D photographing trajectory path calculation method according to claim 1 when executed by a processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111251929.8A CN113689458B (en) | 2021-10-27 | 2021-10-27 | 2D shooting track path calculation method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111251929.8A CN113689458B (en) | 2021-10-27 | 2021-10-27 | 2D shooting track path calculation method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113689458A CN113689458A (en) | 2021-11-23 |
| CN113689458B true CN113689458B (en) | 2022-03-29 |
Family
ID=78588210
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111251929.8A Active CN113689458B (en) | 2021-10-27 | 2021-10-27 | 2D shooting track path calculation method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113689458B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114040120B (en) * | 2022-01-06 | 2022-04-12 | 深圳思谋信息科技有限公司 | Shooting path determination method, device and equipment for panel element detection |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104866500A (en) * | 2014-02-25 | 2015-08-26 | 腾讯科技（深圳）有限公司 | Method and device for displaying pictures in classified manner |
| CN105758790A (en) * | 2016-04-08 | 2016-07-13 | 重庆交通大学 | Accelerating loading experimental system for highway pavement |
| CN105791705A (en) * | 2016-05-26 | 2016-07-20 | 厦门美图之家科技有限公司 | Video anti-shake method and system suitable for movable time-lapse photography and shooting terminal |
| CN109697420A (en) * | 2018-12-17 | 2019-04-30 | 长安大学 | A kind of Moving target detection and tracking towards urban transportation |
| CN112446363A (en) * | 2021-01-29 | 2021-03-05 | 广州市玄武无线科技股份有限公司 | Image splicing and de-duplication method and device based on video frame extraction |
| CN112950717A (en) * | 2019-11-26 | 2021-06-11 | 华为技术有限公司 | Space calibration method and system |
| CN113033353A (en) * | 2021-03-11 | 2021-06-25 | 北京文安智能技术股份有限公司 | Pedestrian trajectory generation method based on overlook image, storage medium and electronic device |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070076977A1 (en) * | 2005-10-05 | 2007-04-05 | Kuan-Wen Chen | Method for calibrating camera parameters |
| JP6531689B2 (en) * | 2016-03-22 | 2019-06-19 | 株式会社デンソー | Moving trajectory detection device, moving object detecting device, moving trajectory detection method |
| CN107566688B (en) * | 2017-08-30 | 2021-02-19 | 广州方硅信息技术有限公司 | Convolutional neural network-based video anti-shake method and device and image alignment device |
| CN111742550A (en) * | 2018-04-27 | 2020-10-02 | 深圳市柔宇科技股份有限公司 | 3D image shooting method, 3D shooting equipment and storage medium |
-
2021
- 2021-10-27 CN CN202111251929.8A patent/CN113689458B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104866500A (en) * | 2014-02-25 | 2015-08-26 | 腾讯科技（深圳）有限公司 | Method and device for displaying pictures in classified manner |
| CN105758790A (en) * | 2016-04-08 | 2016-07-13 | 重庆交通大学 | Accelerating loading experimental system for highway pavement |
| CN105791705A (en) * | 2016-05-26 | 2016-07-20 | 厦门美图之家科技有限公司 | Video anti-shake method and system suitable for movable time-lapse photography and shooting terminal |
| CN109697420A (en) * | 2018-12-17 | 2019-04-30 | 长安大学 | A kind of Moving target detection and tracking towards urban transportation |
| CN112950717A (en) * | 2019-11-26 | 2021-06-11 | 华为技术有限公司 | Space calibration method and system |
| CN112446363A (en) * | 2021-01-29 | 2021-03-05 | 广州市玄武无线科技股份有限公司 | Image splicing and de-duplication method and device based on video frame extraction |
| CN113033353A (en) * | 2021-03-11 | 2021-06-25 | 北京文安智能技术股份有限公司 | Pedestrian trajectory generation method based on overlook image, storage medium and electronic device |
Non-Patent Citations (3)
| Title |
|---|
| Image-Only Real-Time Incremental UAV Image Mosaic for Multi-Strip Flight;F. Zhang et al.;《IEEE Transactions on Multimedia》;20210423;第23卷;全文 * |
| 基于序列图像的运动轨迹检测方法;王照付 等;《辽宁石油化工大学学报》;20100930;第30卷(第3期);全文 * |
| 夜间城市交通监控中各类车辆轨迹的完整提取;汤春明 等;《计算机工程与应用》;20170419;第54卷(第6期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113689458A (en) | 2021-11-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11361459B2 (en) | Method, device and non-transitory computer storage medium for processing image | |
| CN108960211B (en) | Multi-target human body posture detection method and system | |
| KR101333871B1 (en) | Method and arrangement for multi-camera calibration | |
| CN109934847B (en) | Method and device for estimating posture of weak texture three-dimensional object | |
| CN110097586B (en) | Face detection tracking method and device | |
| CN111383252B (en) | Multi-camera target tracking method, system, device and storage medium | |
| CN112446363A (en) | Image splicing and de-duplication method and device based on video frame extraction | |
| CN109410211A (en) | The dividing method and device of target object in a kind of image | |
| EP3100177A1 (en) | Method for recognizing objects | |
| CN113689458B (en) | 2D shooting track path calculation method and device | |
| CN111263955A (en) | Method and device for determining movement track of target object | |
| CN112489088A (en) | Twin network visual tracking method based on memory unit | |
| CN109684953B (en) | Method and device for pig tracking based on target detection and particle filter algorithm | |
| CN113888425A (en) | Industrial quality inspection image character matching method and system based on multi-frame fusion | |
| Yetiş et al. | Adaptive vision based condition monitoring and fault detection method for multi robots at production lines in industrial systems | |
| CN102708559A (en) | Blur difference estimation using multi-kernel convolution | |
| CN116883897A (en) | Low-resolution target identification method | |
| CN117218633A (en) | Article detection method, device, equipment and storage medium | |
| CN113469216B (en) | Retail terminal poster identification and integrity judgment method, system and storage medium | |
| CN113378886B (en) | Method for automatically training shape matching model | |
| CN114004891A (en) | Distribution network line inspection method based on target tracking and related device | |
| CN114399432A (en) | Target identification method, device, equipment, medium and product | |
| CN112861652A (en) | Method and system for tracking and segmenting video target based on convolutional neural network | |
| WO2017186576A1 (en) | Method, apparatus and computer program product for tracking a reference image | |
| CN110705479A (en) | Model training method, target recognition method, device, equipment and medium |
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 | ||
| CP02 | Change in the address of a patent holder |
Address after: 510000 room 23bd, No. 109, TIYU West Road, Tianhe District, Guangzhou City, Guangdong Province Patentee after: GUANGZHOU XUANWU WIRELESS TECHNOLOGY Co.,Ltd. Address before: 32B, no.103b, TianLiHe Road, Guangzhou, 510000 Patentee before: GUANGZHOU XUANWU WIRELESS TECHNOLOGY Co.,Ltd. |
|
| CP02 | Change in the address of a patent holder |