CN111652111A - Target detection method and related device - Google Patents
Target detection method and related device Download PDFInfo
- Publication number
- CN111652111A CN111652111A CN202010475378.2A CN202010475378A CN111652111A CN 111652111 A CN111652111 A CN 111652111A CN 202010475378 A CN202010475378 A CN 202010475378A CN 111652111 A CN111652111 A CN 111652111A
- Authority
- CN
- China
- Prior art keywords
- dynamic
- target
- algorithm
- splicing
- picture
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 98
- 238000010586 diagram Methods 0.000 claims abstract description 26
- 238000000034 method Methods 0.000 claims abstract description 22
- 238000004422 calculation algorithm Methods 0.000 claims description 85
- 238000000605 extraction Methods 0.000 claims description 14
- 238000007781 pre-processing Methods 0.000 claims description 10
- 238000000638 solvent extraction Methods 0.000 claims description 8
- 230000006870 function Effects 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000011156 evaluation Methods 0.000 claims description 6
- 238000001914 filtration Methods 0.000 claims description 4
- 239000000203 mixture Substances 0.000 claims description 4
- 230000011218 segmentation Effects 0.000 claims description 2
- 230000007547 defect Effects 0.000 abstract description 8
- 230000008569 process Effects 0.000 description 5
- 238000012545 processing Methods 0.000 description 4
- 230000004044 response Effects 0.000 description 4
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000005192 partition Methods 0.000 description 2
- 230000003068 static effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 230000006386 memory function Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/20—Scenes; Scene-specific elements in augmented reality scenes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
- G06T3/4038—Image mosaicing, e.g. composing plane images from plane sub-images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/07—Target detection
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a target detection method and a related device, wherein the target detection method comprises the following steps: acquiring a picture to be detected; extracting a dynamic region related to the target from the picture to be detected; splicing the extracted dynamic regions to form a dynamic region splicing diagram; and detecting the target according to the dynamic area splicing diagram. The method overcomes the defects of low detection rate, easy target loss, inaccurate position and the like of small target detection under ultrahigh resolution.
Description
Technical Field
The invention relates to the technical field of image detection, in particular to a target detection method and a related device.
Background
The existing object detection and tracking mainly includes that a detection algorithm is used for calculating an overall image to obtain a target position, and the target position is used as a reference to continuously track a target through a tracking algorithm. However, for the ultrahigh resolution picture, the size of the target to be detected and tracked is very different from the size of the whole picture, so that most of dynamic targets cannot be detected by the detection algorithm, and the detection and tracking algorithm can only detect and track large targets (such as targets of airplanes, engineering vehicles, trains and the like) in an open scene under the above conditions, and the universality is extremely poor.
The existing perimeter algorithm for detecting the targets of the ultrahigh-resolution images can only detect local areas near the perimeter lines of the pictures, however, due to local detection, the pictures can lose most of picture dynamic object information, and therefore certain constraints exist when the perimeter function lines are drawn, the perimeter algorithm is inconvenient for users to use, the detection efficiency is extremely low, and the detection of the targets is not facilitated.
Disclosure of Invention
The invention provides a target detection method, which overcomes the defects of low detection rate, easy target loss, inaccurate position and the like of small target detection under ultrahigh resolution.
In order to solve the above technical problems, a first technical solution provided by the present invention is: provided is a target detection method including: acquiring a picture to be detected; extracting a dynamic region related to the target from the picture to be detected; splicing the extracted dynamic regions to form a dynamic region splicing diagram; and detecting the target according to the dynamic area splicing diagram.
Wherein, still include after obtaining the picture to be detected: zooming the picture to be detected to a preset ratio; preprocessing the picture to be detected; the preprocessing method comprises one or any combination of a Gaussian filtering algorithm, a fuzzy algorithm and a histogram equalization algorithm.
Wherein the step of extracting the dynamic region related to the target from the picture to be detected comprises: marking dynamic points related to the target in the picture to be detected; partitioning the marked points to form a dynamic area; wherein, the dynamic area is one or at least two.
Wherein the marking of the dynamic point related to the target in the picture to be detected comprises: acquiring dynamic points related to the target by adopting a background modeling algorithm and marking the dynamic points; the background modeling algorithm comprises one or any combination of a Gaussian model, a Gaussian mixture model, a foreground detection algorithm and an ABM algorithm; the partitioning the marked points into dynamic regions comprises: partitioning the marked points by adopting an image connected domain extraction algorithm to form a dynamic region; the image connected domain extraction algorithm comprises one or any combination of depth-first search and breadth-first search.
Wherein, before the stitching the extracted dynamic regions to form a dynamic region stitching map, the method further comprises: fixing the central point of the dynamic area with the area not conforming to the first preset value, and scaling the size of the dynamic area; and/or calculating the overlapped dynamic areas through a detection evaluation function, and combining the dynamic areas of which the calculation results are greater than a second preset value; the stitching the extracted dynamic regions to form a dynamic region stitching graph includes: splicing the dynamic regions according to the sizes of the pictures of the dynamic regions by adopting a filling algorithm to form a splicing picture of the dynamic regions; wherein the filling algorithm comprises a two-dimensional image arrangement algorithm.
Wherein, after the splicing the extracted dynamic regions to form a dynamic region splicing map, the method further comprises: judging whether the dynamic area mosaic meets preset requirements or not; if not, segmenting the dynamic region splicing map, and splicing again according to the size of the segmented picture by adopting a filling algorithm to form a new dynamic region splicing map, wherein the size of the new dynamic region splicing map is smaller than that of the dynamic region splicing map before segmentation; if so, performing the following steps: and detecting the target according to the dynamic area splicing diagram.
Wherein the detecting the target according to the dynamic region splicing map comprises: adopting a target detection algorithm to splice the images in the dynamic area to detect the target; wherein, the target detection algorithm comprises one or any combination of fast-RCNN, Yolov3, Yolov2, Yolov1 and SSD algorithm.
In order to solve the above technical problems, a second technical solution provided by the present invention is: provided is an object detection device including: the image acquisition module is used for acquiring an image to be detected; the dynamic region extraction module is used for extracting a dynamic region related to the target from the picture to be detected; the splicing module is used for splicing the extracted dynamic regions to form a dynamic region splicing diagram; and the detection module is used for detecting the target according to the dynamic area splicing diagram.
In order to solve the above technical problems, a third technical solution provided by the present invention is: there is provided an object detection apparatus comprising a memory and a processor, wherein the memory stores program instructions, and the processor retrieves the program instructions from the memory to perform any of the object detection methods described above.
In order to solve the above technical problems, a fourth technical solution provided by the present invention is: there is provided a computer readable storage medium storing an ordered file, the program file being executable to implement the object detection method of any one of the above.
The invention has the beneficial effects that: different from the situation of the prior art, the method extracts the dynamic area related to the target from the picture to be detected, splices the extracted dynamic area to form a dynamic area splicing map, and detects the target according to the dynamic splicing map so as to overcome the defects of low detection rate, easy target loss, inaccurate position and the like of small target detection under ultrahigh resolution.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the description of the embodiments are briefly introduced below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without inventive efforts, wherein:
FIG. 1 is a schematic flow chart diagram of a first embodiment of the object detection method of the present invention;
FIG. 2 is a schematic flow chart diagram of a second embodiment of the object detection method of the present invention;
FIG. 3 is a flowchart illustrating an embodiment of step S12 in FIG. 1;
FIG. 4 is a flowchart illustrating an embodiment of step S13 in FIG. 1;
FIG. 5 is a schematic flow chart diagram of a third embodiment of the object detection method of the present invention;
FIG. 6 is a schematic structural diagram of a first embodiment of the object detecting device of the present invention;
FIG. 7 is a schematic structural diagram of a second embodiment of the object detecting device of the present invention;
fig. 8 is a schematic structural diagram of the computer-readable storage medium of the present invention.
Detailed Description
The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, 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 application.
The present invention will be described in detail below with reference to the accompanying drawings and examples.
Fig. 1 is a schematic flow chart of a target detection method according to a first embodiment of the present invention. The method comprises the following steps:
step S11: and acquiring the picture to be detected.
Specifically, the picture to be detected includes a target to be detected, and the picture to be detected is a single picture, or the picture to be detected is a frame of picture captured from a section of video. The target to be detected can be a large target, such as a tall building, a tree, a tower and the like, and can also be a small target, such as a human, a vehicle, an animal and the like. In one embodiment, the target to be detected is a dynamic target, such as a walking person, car, dog, etc.; in other embodiments, the target to be detected may be a static target, such as a tree, a tall building, etc. It is understood that the dynamic target and the static target are relative to different references and are not limited herein.
Step S12: and extracting a dynamic area related to the target from the picture to be detected.
Specifically, the picture to be detected contains a plurality of objects, and in order to simplify an algorithm in target detection, the target to be detected in the picture to be detected is extracted. For example, if the target to be detected is a person, and the picture to be detected includes a tree, a car, a person, and the like, the region related to the target to be detected is extracted, that is, the region of a bag or an article carried by the person or the like is extracted as a plurality of small regions.
Specifically, referring to fig. 3, step S12 specifically includes:
step S31: and marking the dynamic point related to the target in the picture to be detected.
Specifically, a background modeling algorithm is adopted to detect and mark dynamic points related to a target; in a specific embodiment, the background modeling algorithm includes one or any combination of a gaussian model (BGModel), a Gaussian Mixture Model (GMM), a foreground detection algorithm (ViBE), and an ABM algorithm.
In one embodiment, assuming that the target to be detected is a person, a background modeling algorithm is adopted to detect dynamic points such as arms, feet, legs, shoulders and the like of the person for marking. If the image to be detected is a frame of image intercepted from a section of video, the position of the dynamic point can be determined according to the comparison with a plurality of frames or a frame of image before the frame of image.
Step S32: and partitioning the marked points to form a dynamic area.
In one embodiment, the number of marked points is several, and the marked points are partitioned to form the dynamic area. Specifically, all marked points may form one dynamic region, or a plurality of marked points may be divided to form a plurality of dynamic regions. For example, the dynamic regions may be divided according to the positions of arms, legs, and shoulders.
Specifically, an image connected domain extraction algorithm is adopted to partition the marked points to form a dynamic region. The image connected domain extraction algorithm comprises one or any combination of depth-first search and breadth-first search.
Step S13: and splicing the extracted dynamic regions to form a dynamic region splicing diagram.
Specifically, a plurality of extracted dynamic regions related to the target to be detected are spliced to form a dynamic region splicing diagram. In an embodiment, the dynamic regions may be stitched according to the picture sizes of the dynamic regions by using a filling algorithm to form a dynamic region stitching map. Wherein the filling algorithm comprises a two-dimensional image arrangement algorithm.
Referring to fig. 4, in an embodiment, before splicing the extracted dynamic regions, preprocessing the dynamic regions is further required, which specifically includes:
step S41: and fixing the central point of the dynamic region with the area not meeting the first preset value, and scaling the size of the dynamic region.
After the plurality of dynamic regions are obtained in the above manner, the areas of the plurality of dynamic regions are different, and in order to facilitate splicing of the plurality of dynamic regions, the dynamic regions with areas not meeting the first preset value are scaled by using the central point as a reference, so that the dynamic regions meet the preset requirement.
Step S42: and calculating the overlapped dynamic areas through a detection evaluation function, and combining the dynamic areas of which the calculation results are greater than a second preset value.
After the plurality of dynamic regions are obtained in the above manner, the areas of the plurality of dynamic regions are partially overlapped, so that the plurality of dynamic regions are conveniently spliced, the overlapped dynamic regions are calculated through a detection evaluation function, and the dynamic regions with the calculation results larger than the second preset value are merged.
Specifically, in an embodiment, assuming that the second preset value is 0.3, the overlapped dynamic regions are calculated by detecting the evaluation function, and if the obtained calculation result is greater than 0.3, the multiple dynamic regions with the overlapped regions are combined to form one dynamic region.
Step S14: and detecting the target according to the dynamic area splicing diagram.
The position of the target is detected in the dynamic region mosaic formed after mosaic, and in one embodiment, the position of the target can be determined by detecting the pixel points of the target. Specifically, the target can be detected in the dynamic region mosaic formed after mosaic by a target detection algorithm. Target detection algorithms include Faster-RCNN, Yolov3, Yolov2, Yolov1, SSD, etc.
According to the target detection method provided by the invention, a background modeling algorithm is used for obtaining dynamic areas, a plurality of dynamic areas are spliced to obtain a spliced graph, and then a target position is obtained through a target detection algorithm. The method has the advantages of real-time response, simple application, wide application scene and the like, can well apply the definition of the regional scene of the high-resolution picture, overcomes the defects of low detection rate of the traditional high-resolution picture on small objects, easy loss of targets, inaccurate position and the like, and can perform close-up detection and tracking on dynamic objects in the ultra-wide high-resolution scene in real time.
The target detection method provided by the embodiment of the invention is used for detecting the tiny dynamic objects under the ultrahigh resolution, can use the definition of the regional scene of the high-resolution picture, overcomes the defects of low detection rate of the traditional high-resolution picture on the small objects, easy target loss, inaccurate position and the like, and can perform close-up detection and tracking on the dynamic objects under the ultra-wide high-resolution scene in real time.
Referring to fig. 2, a flowchart of a second embodiment of the target detection method of the present invention is shown, wherein steps S21, S24, S25 and S26 are the same as steps S11, S12, S13 and S14 in the first embodiment shown in fig. 1, except that the present embodiment further includes, after step S21:
step S22: and zooming the picture to be detected to a preset ratio.
Specifically, if the picture to be detected is too large, the picture to be detected is reduced according to a certain proportion for the convenience of detection. If the picture to be detected is too small, the picture to be detected is amplified according to a certain proportion for the convenience of detection.
Step S23: and preprocessing the picture to be detected.
Specifically, after the image to be detected is reduced or enlarged, the reduced or enlarged image to be detected is preprocessed. Wherein, the pretreatment process comprises the following steps: and (3) performing one or any combination processing of a Gaussian filtering algorithm, a fuzzy algorithm and a histogram equalization algorithm on the picture to be detected, thereby obtaining a clearer picture which is easier to identify and detect.
The target detection method provided by the invention comprises the steps of carrying out scaling and preprocessing on a picture to be detected, obtaining a dynamic region through the picture to be detected by using a background modeling algorithm, splicing a plurality of ding-tai regions to obtain a spliced picture, and then obtaining a target position through a target detection algorithm. The method has the advantages of real-time response, simple application, wide application scene and the like, can well apply the definition of the regional scene of the high-resolution picture, overcomes the defects of low detection rate of the traditional high-resolution picture on small objects, easy loss of targets, inaccurate position and the like, and can perform close-up detection and tracking on dynamic objects in the ultra-wide high-resolution scene in real time.
Fig. 5 is a schematic flow chart of a target detection method according to a third embodiment of the present invention. Wherein, steps S51, S52, S53 and S56 are the same as steps S11, S12, S13 and S14 in the first embodiment shown in fig. 1, except that the embodiment further includes, after step S53:
step S54: and judging whether the dynamic area splicing map meets the preset requirement or not.
Specifically, after the dynamic regions are spliced to form a dynamic region splicing map, whether the size of the dynamic region splicing map meets a preset requirement is judged, and the preset requirement meets the requirement of target detection. In other embodiments, whether the definition and the like of the dynamic region splicing map meet the requirements can be detected, which is not limited to the size and the definition, and is not particularly limited.
If the dynamic region mosaic does not meet the preset requirement, performing step S55: and (4) segmenting the dynamic region splicing image, and splicing again according to the size of the segmented image by adopting a filling algorithm to form a new dynamic region splicing image.
Specifically, the filling algorithm includes a two-dimensional image arrangement algorithm. And returning to the step S54 after obtaining the new dynamic region splicing map, continuously judging whether the dynamic region splicing map meets the preset requirement, and if the dynamic region splicing map meets the preset requirement, executing the step S56: and detecting the target according to the dynamic area splicing diagram.
Specifically, the position of the target is detected in the dynamic region mosaic formed after mosaic, and in one embodiment, the position of the target can be determined by detecting the pixel points of the target. Specifically, the target can be detected in the dynamic region mosaic formed after mosaic by a target detection algorithm. Target detection algorithms include Faster-RCNN, Yolov3, Yolov2, Yolov1, SSD, etc.
Specifically, the split pictures are re-spliced by using a filling algorithm, and the size of the formed dynamic region splicing map is smaller than that of the dynamic region splicing map before the split pictures are formed, that is, the splitting process is a process of compressing the size of the original dynamic region splicing map. Therefore, in step S54, it may be preferentially determined whether the size of the dynamic region mosaic meets the preset size in the preset requirement, and if not, the size of the dynamic region mosaic is compressed, where the size compression specifically includes: and segmenting and splicing the dynamic region splicing map. Therefore, the dynamic region splicing diagram with the minimum area can be obtained, and the target detection is convenient to carry out subsequently.
The target detection method provided by the invention comprises the steps of zooming and preprocessing a picture to be detected, obtaining a dynamic area through the picture to be detected by using a background modeling algorithm, splicing a plurality of dynamic areas to obtain a spliced picture, and then obtaining a target position through a target detection algorithm. The method has the advantages of real-time response, simple application, wide application scene and the like, can well apply the definition of the regional scene of the high-resolution picture, overcomes the defects of low detection rate of the traditional high-resolution picture on small objects, easy loss of targets, inaccurate position and the like, and can perform close-up detection and tracking on dynamic objects in the ultra-wide high-resolution scene in real time.
Fig. 6 is a schematic structural diagram of a target detection device according to a first embodiment of the present invention. The target detection device comprises a picture acquisition module 61, a dynamic region extraction module 62, a splicing module 63 and a detection module 64.
The image obtaining module 61 is configured to obtain an image to be detected, and in an embodiment, the image obtaining module 61 is further configured to scale the image to be detected to a preset ratio and pre-process the image to be detected; the preprocessing method comprises one or any combination of a Gaussian filtering algorithm, a fuzzy algorithm and a histogram equalization algorithm.
The dynamic region extracting module 62 is configured to extract a dynamic region related to the target from the picture to be detected. In an embodiment, the dynamic region extracting module 62 is further configured to mark a dynamic point related to the target in the picture to be detected; partitioning the marked points to form a dynamic area; wherein, the dynamic area is one or at least two. Specifically, the dynamic region extraction module 62 is further configured to obtain a dynamic point related to the target by using a background modeling algorithm and perform labeling; the background modeling algorithm comprises one or any combination of a Gaussian model, a Gaussian mixture model, a foreground detection algorithm and an ABM algorithm; the dynamic region extraction module 62 is further configured to partition the marked points by using an image connected domain extraction algorithm to form a dynamic region; the image connected domain extraction algorithm comprises one or any combination of depth-first search and breadth-first search. In an embodiment, the dynamic region extraction module 62 is further configured to fix a central point of the dynamic region having an area that does not meet the first preset value, and scale the size of the dynamic region; and/or calculating the overlapped dynamic areas through a detection evaluation function, and combining the dynamic areas of which the calculation results are greater than a second preset value.
The splicing module 63 is configured to splice the extracted dynamic regions to form a dynamic region splicing map. Specifically, in an embodiment, the splicing module 63 is further configured to splice the dynamic regions according to the picture sizes of the dynamic regions by using a filling algorithm to form a dynamic region splicing map; wherein the filling algorithm comprises a two-dimensional image arrangement algorithm. The splicing module 63 judges whether the dynamic region splicing map meets preset requirements; if not, the dynamic region splicing image is segmented, and a filling algorithm is adopted to re-splice according to the size of the segmented image so as to form a new dynamic region splicing image.
The detection module 64 is configured to detect the target according to the dynamic region splicing map. Specifically, in an embodiment, the detection module 64 is further configured to adopt a target detection algorithm to stitch the target in the dynamic region; wherein, the target detection algorithm comprises one or any combination of fast-RCNN, Yolov3, Yolov2, Yolov1 and SSD algorithm.
The target detection device provided by the invention can realize zooming and preprocessing of the picture to be detected, obtains a dynamic region through the picture to be detected by using a background modeling algorithm, obtains a spliced picture by splicing a plurality of ding ai regions, and then obtains a target position through a target detection algorithm. The method has the advantages of real-time response, simple application, wide application scene and the like, can well apply the definition of the regional scene of the high-resolution picture, overcomes the defects of low detection rate of the traditional high-resolution picture on small objects, easy loss of targets, inaccurate position and the like, and can perform close-up detection and tracking on dynamic objects in the ultra-wide high-resolution scene in real time.
Fig. 7 is a schematic structural diagram of a target detection device according to a second embodiment of the present invention. The smart device comprises a memory 72 and a processor 71 connected to each other.
The memory 72 is used to store program instructions for implementing the object detection method of any of the above.
The processor 71 may also be referred to as a CPU (Central Processing Unit). The processor 71 may be an integrated circuit chip having signal processing capabilities. The processor 71 may also be a general purpose processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
The memory 72 may be a memory bank, a TF card, etc. and may store all the information in the target detection apparatus, including the input raw data, the computer program, the intermediate operation results and the final operation results. It stores and retrieves information based on the location specified by the controller. With the memory, the target detection device has a memory function, and normal operation can be guaranteed. The memory of the object detection device may be classified into a main memory (internal memory) and an auxiliary memory (external memory) according to the purpose of use, and also into an external memory and an internal memory. The external memory is usually a magnetic medium, an optical disk, or the like, and can store information for a long period of time. The memory refers to a storage component on the main board, which is used for storing data and programs currently being executed, but is only used for temporarily storing the programs and the data, and the data is lost when the power is turned off or the power is cut off.
In the several embodiments provided in the present application, it should be understood that the disclosed method and apparatus may be implemented in other ways. For example, the above-described apparatus embodiments are merely illustrative, and for example, a division of a module or a unit is merely a logical division, and an actual implementation may have another division, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.
Units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the embodiment.
In addition, functional units in the embodiments of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.
The integrated unit, if implemented in the form of a software functional unit and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present application may be substantially implemented or contributed to by the prior art, or all or part of the technical solution may be embodied in a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a system server, a network device, or the like) or a processor (processor) to execute all or part of the steps of the method of the embodiments of the present application.
Please refer to fig. 8, which is a schematic structural diagram of a computer-readable storage medium according to the present invention. The storage medium of the present application stores a program file 81 capable of implementing all the above object detection methods, where the program file 81 may be stored in the storage medium in the form of a software product, and includes several instructions to enable a computer device (which may be a personal computer, a server, or a network device) or a processor (processor) to execute all or part of the steps of the methods according to the embodiments of the present application. The aforementioned storage device includes: various media capable of storing program codes, such as a usb disk, a mobile hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, or terminal devices, such as a computer, a server, a mobile phone, and a tablet.
The above description is only an embodiment of the present invention, and not intended to limit the scope of the present invention, and all modifications of equivalent structures and equivalent processes, which are made by using the contents of the present specification and the accompanying drawings, or directly or indirectly applied to other related technical fields, are included in the scope of the present invention.
Claims (10)
1. A method of object detection, comprising:
acquiring a picture to be detected;
extracting a dynamic region related to the target from the picture to be detected;
splicing the extracted dynamic regions to form a dynamic region splicing diagram;
and detecting the target according to the dynamic area splicing diagram.
2. The object detection method according to claim 1, wherein after the obtaining the picture to be detected, the method further comprises:
zooming the picture to be detected to a preset ratio;
preprocessing the picture to be detected;
the preprocessing method comprises one or any combination of a Gaussian filtering algorithm, a fuzzy algorithm and a histogram equalization algorithm.
3. The object detection method according to claim 1, wherein the step of extracting the dynamic region related to the object from the picture to be detected comprises:
marking dynamic points related to the target in the picture to be detected;
partitioning the marked points to form a dynamic area;
wherein, the dynamic area is one or at least two.
4. The target detection algorithm of claim 3, wherein the marking of the dynamic point related to the target in the picture to be detected comprises:
acquiring dynamic points related to the target by adopting a background modeling algorithm and marking the dynamic points;
the background modeling algorithm comprises one or any combination of a Gaussian model, a Gaussian mixture model, a foreground detection algorithm and an ABM algorithm;
the partitioning the marked points into dynamic regions comprises:
partitioning the marked points by adopting an image connected domain extraction algorithm to form a dynamic region;
the image connected domain extraction algorithm comprises one or any combination of depth-first search and breadth-first search.
5. The object detection method of claim 1, wherein the stitching the extracted dynamic regions to form a dynamic region stitching map further comprises:
fixing the central point of the dynamic area with the area not conforming to the first preset value, and scaling the size of the dynamic area; and/or
Calculating the overlapped dynamic areas through a detection evaluation function, and combining the dynamic areas of which the calculation results are greater than a second preset value;
the stitching the extracted dynamic regions to form a dynamic region stitching graph includes:
splicing the dynamic regions according to the sizes of the pictures of the dynamic regions by adopting a filling algorithm to form a splicing picture of the dynamic regions;
wherein the filling algorithm comprises a two-dimensional image arrangement algorithm.
6. The object detection method of claim 1, wherein the stitching the extracted dynamic regions to form a dynamic region stitching map further comprises:
judging whether the dynamic area mosaic meets preset requirements or not;
if not, segmenting the dynamic region splicing map, and splicing again according to the size of the segmented picture by adopting a filling algorithm to form a new dynamic region splicing map, wherein the size of the new dynamic region splicing map is smaller than that of the dynamic region splicing map before segmentation;
if so, performing the following steps: and detecting the target according to the dynamic area splicing diagram.
7. The object detection method of claim 1, wherein the detecting the object according to the dynamic region splicing map comprises:
adopting a target detection algorithm to splice the images in the dynamic area to detect the target;
wherein, the target detection algorithm comprises one or any combination of fast-RCNN, Yolov3, Yolov2, Yolov1 and SSD algorithm.
8. An object detection device, comprising:
the image acquisition module is used for acquiring an image to be detected;
the dynamic region extraction module is used for extracting a dynamic region related to the target from the picture to be detected;
the splicing module is used for splicing the extracted dynamic regions to form a dynamic region splicing diagram;
and the detection module is used for detecting the target according to the dynamic area splicing diagram.
9. An object detection device, comprising: a memory storing program instructions and a processor retrieving the program instructions from the memory to perform the object detection method of any one of claims 1-7.
10. A computer-readable storage medium, characterized in that a program file is stored, which can be executed to implement the object detection method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010475378.2A CN111652111A (en) | 2020-05-29 | 2020-05-29 | Target detection method and related device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010475378.2A CN111652111A (en) | 2020-05-29 | 2020-05-29 | Target detection method and related device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111652111A true CN111652111A (en) | 2020-09-11 |
Family
ID=72349705
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010475378.2A Pending CN111652111A (en) | 2020-05-29 | 2020-05-29 | Target detection method and related device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111652111A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113344957A (en) * | 2021-07-19 | 2021-09-03 | 北京城市网邻信息技术有限公司 | Image processing method, image processing apparatus, and non-transitory storage medium |
| CN113487513A (en) * | 2021-07-20 | 2021-10-08 | 浙江大华技术股份有限公司 | Picture brightness adjusting method and adjusting device and storage medium thereof |
| CN115100646A (en) * | 2022-06-27 | 2022-09-23 | 武汉兰丁智能医学股份有限公司 | Cell image high-definition rapid splicing identification marking method |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104766344A (en) * | 2015-03-31 | 2015-07-08 | 华南理工大学 | Vehicle detecting method based on moving edge extractor |
| CN106570824A (en) * | 2016-10-12 | 2017-04-19 | 网易（杭州）网络有限公司 | Splicing method and device for scattered graphs |
| CN109035292A (en) * | 2018-08-31 | 2018-12-18 | 北京智芯原动科技有限公司 | Moving target detecting method and device based on deep learning |
| CN110378218A (en) * | 2019-06-13 | 2019-10-25 | 大亚湾核电运营管理有限责任公司 | A kind of image processing method, device and terminal device |
| CN110751678A (en) * | 2018-12-12 | 2020-02-04 | 北京嘀嘀无限科技发展有限公司 | Moving object detection method and device and electronic equipment |
| CN110991274A (en) * | 2019-11-18 | 2020-04-10 | 杭州电子科技大学 | Pedestrian tumbling detection method based on Gaussian mixture model and neural network |
-
2020
- 2020-05-29 CN CN202010475378.2A patent/CN111652111A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104766344A (en) * | 2015-03-31 | 2015-07-08 | 华南理工大学 | Vehicle detecting method based on moving edge extractor |
| CN106570824A (en) * | 2016-10-12 | 2017-04-19 | 网易（杭州）网络有限公司 | Splicing method and device for scattered graphs |
| CN109035292A (en) * | 2018-08-31 | 2018-12-18 | 北京智芯原动科技有限公司 | Moving target detecting method and device based on deep learning |
| CN110751678A (en) * | 2018-12-12 | 2020-02-04 | 北京嘀嘀无限科技发展有限公司 | Moving object detection method and device and electronic equipment |
| CN110378218A (en) * | 2019-06-13 | 2019-10-25 | 大亚湾核电运营管理有限责任公司 | A kind of image processing method, device and terminal device |
| CN110991274A (en) * | 2019-11-18 | 2020-04-10 | 杭州电子科技大学 | Pedestrian tumbling detection method based on Gaussian mixture model and neural network |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113344957A (en) * | 2021-07-19 | 2021-09-03 | 北京城市网邻信息技术有限公司 | Image processing method, image processing apparatus, and non-transitory storage medium |
| CN113487513A (en) * | 2021-07-20 | 2021-10-08 | 浙江大华技术股份有限公司 | Picture brightness adjusting method and adjusting device and storage medium thereof |
| CN115100646A (en) * | 2022-06-27 | 2022-09-23 | 武汉兰丁智能医学股份有限公司 | Cell image high-definition rapid splicing identification marking method |
| CN115100646B (en) * | 2022-06-27 | 2023-01-31 | 武汉兰丁智能医学股份有限公司 | Cell image high-definition rapid splicing identification marking method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Liu et al. | Region enhanced scale-invariant saliency detection | |
| CN111652111A (en) | Target detection method and related device | |
| US8355079B2 (en) | Temporally consistent caption detection on videos using a 3D spatiotemporal method | |
| CN108564579B (en) | Concrete crack detection method and detection device based on time-space correlation | |
| CN111524145A (en) | Intelligent picture clipping method and system, computer equipment and storage medium | |
| CN113496208B (en) | Video scene classification method and device, storage medium and terminal | |
| CN103488966A (en) | Intelligent mobile phone capable of identifying real-name ticket information | |
| CN113392681A (en) | Human body falling detection method and device and terminal equipment | |
| CN113326773A (en) | Recognition model training method, recognition method, device, equipment and storage medium | |
| CN107770487B (en) | Feature extraction and optimization method, system and terminal equipment | |
| WO2021196925A1 (en) | Method and apparatus for detecting and tracking moving object | |
| CN110348353B (en) | Image processing method and device | |
| Fang et al. | 1-D barcode localization in complex background | |
| CN107274382A (en) | A kind of state identification method, device and the electronic equipment of hard pressing plate | |
| JP6403207B2 (en) | Information terminal equipment | |
| CN112101134B (en) | Object detection method and device, electronic equipment and storage medium | |
| US9798932B2 (en) | Video extraction method and device | |
| CN111177449B (en) | Multi-dimensional information integration method based on picture and related equipment | |
| CN113516609A (en) | Split screen video detection method and device, computer equipment and storage medium | |
| CN112183277A (en) | Detection method and device for abandoned object and lost object, terminal equipment and storage medium | |
| CN113743219B (en) | Moving object detection method and device, electronic equipment and storage medium | |
| CN116798056B (en) | Form image positioning method, apparatus, device and computer readable storage medium | |
| CN110580444B (en) | Human body detection method and device | |
| CN113365027B (en) | Video processing method and device, electronic equipment and storage medium | |
| WO2015162028A2 (en) | Method, device, user equipment and computer program for object extraction from multimedia content |
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 |