CN113128302A - Image detection method and related product - Google Patents

Image detection method and related product Download PDF

Info

Publication number
CN113128302A
CN113128302A CN201911424558.1A CN201911424558A CN113128302A CN 113128302 A CN113128302 A CN 113128302A CN 201911424558 A CN201911424558 A CN 201911424558A CN 113128302 A CN113128302 A CN 113128302A
Authority
CN
China
Prior art keywords
model
detection
image
target
state
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
Application number
CN201911424558.1A
Other languages
Chinese (zh)
Inventor
黄德威
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shenzhen Intellifusion Technologies Co Ltd
Original Assignee
Shenzhen Intellifusion Technologies Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Shenzhen Intellifusion Technologies Co Ltd filed Critical Shenzhen Intellifusion Technologies Co Ltd
Priority to CN201911424558.1A priority Critical patent/CN113128302A/en
Publication of CN113128302A publication Critical patent/CN113128302A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/94—Hardware or software architectures specially adapted for image or video understanding

Landscapes

  • Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Theoretical Computer Science (AREA)
  • Image Analysis (AREA)

Abstract

The embodiment of the application discloses an image detection method and a related product, which are applied to an image detection system, wherein the image detection system comprises an image acquisition module, a first detection model, a second detection model, a correction model and a decision model, and the method comprises the following steps: when a model detection state from a current detection model is received through a decision model, determining a target detection model according to the model detection state and a historical model detection state, sending the target detection model to an image acquisition module, wherein the current detected image is any one of images except a first image, a second image and a last image in a target image sequence, the target detection model is one of a first detection model and a second detection model, then sending the current image to be detected to the target detection model through the image acquisition module, and finally, carrying out image detection through the target detection model according to the current image to be detected. The embodiment of the application improves the accuracy of image detection.

Description

Image detection method and related product
Technical Field
The present application relates to the field of electronic technologies, and in particular, to an image detection method and a related product.
Background
With the development of science and technology, people have higher and higher requirements on image detection, and at present, in order to detect a target object in a plurality of images, a detection and tracking method, namely frame skipping detection, is generally used, and frames which are not detected are tracked by using methods such as optical flow and the like, so that the image detection is inaccurate.
Disclosure of Invention
The embodiment of the application provides an image detection method and a related product, so as to improve the accuracy of image detection.
In a first aspect, an embodiment of the present application provides an image detection method, which is applied to an image detection system, where the image detection system includes an image acquisition module, a first detection model, a second detection model, a modification model, and a decision model, the image acquisition module is in traffic connection with the first detection model, the second detection model, and the decision model, the first detection model is in traffic connection with the modification model and the decision model, the second detection model is in traffic connection with the modification model and the decision model, and the modification model is in communication connection with the decision model, where the method includes:
when a model detection state from a current detection model is received through the decision model, determining a target detection model according to the model detection state and a historical model detection state, and sending the target detection model to the image acquisition module, wherein the current detected image is any image except a first image, a second image and a last image in a target image sequence, and the target detection model is one of the first detection model and the second detection model;
sending the current image to be detected to the target detection model through the image acquisition module;
and carrying out image detection according to the current image to be detected through the target detection model.
In a second aspect, an embodiment of the present application provides an image detection system, which includes an image acquisition module, a first detection model, a second detection model, a modification model, and a decision model, where the image acquisition module is in communication connection with the first detection model, the second detection model, and the decision model, respectively, the first detection model is in communication connection with the modification model and the decision model, respectively, the second detection model is in communication connection with the modification model and the decision model, respectively, and the modification model is in communication connection with the decision model, where,
the decision-making model is used for determining a target detection model according to the model detection state and the historical model detection state when receiving a model detection state from a current detection model, and sending the target detection model to the image acquisition module, wherein the current detected image is any image except a first image, a second image and a last image in a target image sequence, and the target detection model is one of the first detection model and the second detection model;
the image acquisition module is used for sending the current image to be detected to the target detection model;
and the target detection model is used for carrying out image detection according to the current image to be detected.
In a third aspect, the present application provides an image processing system, including a processor, a memory, a communication interface, and one or more programs, where the one or more programs are stored in the memory and configured to be executed by the processor, and the program includes instructions for executing the steps in any of the methods of the first aspect of the present application.
In a fourth aspect, the present application provides a computer-readable storage medium, where the computer-readable storage medium stores a computer program for electronic data exchange, where the computer program makes a computer perform part or all of the steps described in any one of the methods of the first aspect of the present application.
In a fifth aspect, the present application provides a computer program product, wherein the computer program product includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to cause a computer to perform some or all of the steps as described in any one of the methods of the first aspect of the embodiments of the present application. The computer program product may be a software installation package.
It can be seen that, in the embodiment of the present application, the method is applied to an image detection system, where the image detection system includes an image acquisition module, a first detection model, a second detection model, a modification model, and a decision model, and the method includes: when a model detection state from a current detection model is received through a decision model, determining a target detection model according to the model detection state and a historical model detection state, sending the target detection model to an image acquisition module, wherein the current detected image is any one of images except a first image, a second image and a last image in a target image sequence, the target detection model is one of a first detection model and a second detection model, then sending the current image to be detected to the target detection model through the image acquisition module, and finally, carrying out image detection through the target detection model according to the current image to be detected. Therefore, the image detection system of the embodiment of the application can analyze the model detection state of the most processed image of the detection models with different precisions through the decision model, and determine the detection model for processing the next image.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a schematic structural diagram of an image detection system provided in an embodiment of the present application;
fig. 2 is a schematic flowchart of an image detection method according to an embodiment of the present application;
FIG. 3 is a schematic structural diagram of another image inspection system provided in an embodiment of the present application;
fig. 4 is a block diagram illustrating functional modules of an image detection system according to an embodiment of the present disclosure.
Detailed Description
In order to make the technical solutions of the present application better understood, 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 terms "first," "second," and the like in the description and claims of the present application and in the above-described drawings are used for distinguishing between different objects and not for describing a particular order. Furthermore, the terms "include" and "have," as well as any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, system, article, or apparatus that comprises a list of steps or elements is not limited to only those steps or elements listed, but may alternatively include other steps or elements not listed, or inherent to such process, method, article, or apparatus.
Reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the application. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. It is explicitly and implicitly understood by one skilled in the art that the embodiments described herein can be combined with other embodiments.
The image detection system provided by the embodiment of the application can be electronic equipment presented in various software and hardware forms.
With the development of science and technology, people have higher and higher requirements on image detection, and at present, in order to detect a target object in a plurality of images, a detection and tracking method, namely frame skipping detection, is generally used, and frames which are not detected are tracked by using methods such as optical flow and the like, so that the image detection is inaccurate.
In view of the above problems, the present application provides an image detection method, and the following describes an embodiment of the present application in detail with reference to the accompanying drawings.
Referring to fig. 1, fig. 1 is a schematic structural diagram of an image detection system provided in an embodiment of the present application, and as shown in fig. 1, an image detection system 10 includes an image acquisition module 101, a first detection model 102, a second detection model 103, a modification model 104, and a decision model 105, where the image acquisition module 101 is in communication connection with the first detection model 102, the second detection model 103, and the decision model 105, respectively, the first detection model 102 is in communication connection with the modification model 104 and the decision model 105, respectively, the second detection model 103 is in communication connection with the modification model 104 and the decision model 105, respectively, and the modification model is in communication connection with the decision model.
The image obtaining module 101 can perform data interaction with the first detection model 102, the second detection model 103, and the decision model 105, respectively, the first detection model 102 can perform data interaction with the modification model 104 and the decision model 105, respectively, the second detection model 103 can perform data interaction with the modification model 104 and the decision model 105, respectively, and the modification model can perform data interaction with the decision model.
Referring to fig. 2, fig. 2 is a schematic flowchart of an image detection method according to an embodiment of the present application, applied to the image detection system shown in fig. 1, and shown in fig. 2, the image detection method includes:
s201, when receiving a model detection state from the current detection model through the decision model, the image detection system determines a target detection model according to the model detection state and a historical model detection state, and sends the target detection model to the image acquisition module, wherein the current detected image is any one image except a first image, a second image and a last image in a target image sequence, and the target detection model is one of the first detection model and the second detection model.
Wherein, the first detection model and the first detection model can detect a target object in the image, the target object can be any one of a human face, a specific thing, a specific animal and the like, the first detection model has higher accuracy than the second detection model, the first detection model can be a detection model with higher accuracy, such as resnet50, and the second detection model can be a light-weight detection model, such as mobilene. The target image sequence comprises at least 4 images to be processed.
Further, when the image detection system receives a model detection state from the current detection model through the decision model, an implementation manner of determining a target detection model according to the model detection state and a historical model detection state and sending the target detection model to the image acquisition module may be: if the current detection model is the first detection model, the image detection system obtains a first target model detection state in the historical model detection states through the decision model, and determines the target detection model according to the model detection state from the current detection model and the first target model detection state, wherein the first target model detection state is a last sent model detection state from the second detection model; and if the current detection model is the second detection model, the image detection system acquires a second target model detection state in the historical model detection states through the decision model, and determines the target detection model according to the model detection state from the current detection model and the second target model detection state, wherein the second target model detection state is the last sent model detection state from the first detection model.
For example, if the target image sequence includes 50 images, the current image to be detected is the 28 th image, among the first 27 images, the 1 st image, the 7 th image, the 15 th image and the 22 nd image are the images processed by the first detection model, the rest images are the images processed by the second detection model, and the current detection model is the first detection model, the model detection state from the last sending of the second detection model in the historical model detection state, that is, the model detection state of the 27 th image is obtained, and the target detection image corresponding to the 29 th image is determined according to the model detection state corresponding to the 27 th image and the model detection state corresponding to the 28 th image.
For example, if the target image sequence includes 30 images, the current image to be detected is the 20 th image, and among the first 19 images, the 1 st image, the 5 th image, and the 15 th image are the images processed by the first detection model, the rest images are the images processed by the second detection model, and the current detection model is the second detection model, the model detection state from the last sending of the first detection model in the historical model detection states, that is, the model detection state of the 15 th image, is obtained, and the target detection image corresponding to the 16 th image is determined according to the model detection state corresponding to the 20 th image and the model detection state corresponding to the 15 th image.
In this example, the image detection system can analyze the model detection state of the most processed image of the detection models with different accuracies through the decision model, determine the detection model for processing the next image, and improve the image detection effect.
S202, the image detection system sends the current image to be detected to the target detection model through the image acquisition module.
For example, if the current image to be detected is the fourth image in the target image sequence and the target detection model is the first detection model, the image acquisition module sends the fourth image in the target image sequence to the first detection model.
For example, if the current image to be detected is the 10 th image in the target image sequence and the target detection model is the second detection model, the image acquisition module sends the 10 th image in the target image sequence to the second detection model.
As can be seen, in this example, the image acquisition module in the image detection system can send the image to be processed to the target detection model determined by the decision module.
S203, the image detection system performs image detection according to the current image to be detected through the target detection model.
The image detection system performs image detection according to the current image to be detected through the target detection model, and may be: if the target detection model is the first detection model, the image detection system obtains a detection model state according to the current image to be detected through the first detection model; sending the detection model state to the decision model; if the target detection model is the second detection model, the image detection system obtains a detection model state according to the current image to be detected through the second detection model; sending the detection model state to the decision model.
Here, the implementation manner of obtaining the state of the detection model by the first detection model according to the current image to be detected may be: the first detection model detects the current image to be detected to obtain a detection result aiming at the current image to be detected; and analyzing the detection result aiming at the current image to be detected to obtain the state of the detection model.
It can be seen that, in the embodiment of the present application, the method is applied to an image detection system, where the image detection system includes an image acquisition module, a first detection model, a second detection model, a modification model, and a decision model, and the method includes: when a model detection state from a current detection model is received through a decision model, determining a target detection model according to the model detection state and a historical model detection state, sending the target detection model to an image acquisition module, wherein the current detected image is any one of images except a first image, a second image and a last image in a target image sequence, the target detection model is one of a first detection model and a second detection model, then sending the current image to be detected to the target detection model through the image acquisition module, and finally, carrying out image detection through the target detection model according to the current image to be detected. Therefore, the image detection system of the embodiment of the application can analyze the model detection state of the most processed image of the detection models with different precisions through the decision model, and determine the detection model for processing the next image.
When the image detection system receives a model detection state from the current detection model through the decision model, a target detection model is determined according to the model detection state and a historical model detection state, and before the target detection model is sent to the image acquisition module, the image detection method further comprises the following steps:
s301, the image detection system obtains an image detection request through the image obtaining module, wherein the image detection request comprises the target image sequence.
S302, the image detection system sends the first image in the target image sequence to the first detection model through the image acquisition module according to the image detection request.
The obtaining process of the image detection request may be: monitoring a target operation of a user; and generating the image detection request according to the target operation. The video recording function can be started by target operation of a user, and the target image sequence can be an image sequence corresponding to a video stream.
S303, the image detection system obtains a first detection model state according to the first image through the first detection model, and sends the first detection model state to the decision model.
And the first detection model state is a detection model state of the first detection model for detecting the first image.
Optionally, the implementation manner of the image detection system obtaining the state of the first detection model according to the first image through the first detection model may be that the image detection system performs the following operations through the first detection model: detecting the first image to obtain a detection result aiming at the first image; and analyzing the detection result aiming at the first image to obtain the state of the first detection model.
S304, the image detection system sends a second image in the target image sequence to the second detection model through the image acquisition module according to the image detection request.
S305, the image detection system obtains a second detection model state according to the second image through the second detection model, and sends the second detection model state to the decision model.
And the first detection model state is a detection model state of the second detection model for detecting the second image.
The state of the second detection model obtained by the second detection model according to the second image is similar to the state of the first detection model obtained by the first detection model according to the first image, please refer to the above description of obtaining the state of the first detection model by the first detection model according to the first image, and details are not repeated here.
S306, the image detection system determines a first target detection model according to the first detection model state and the second detection model state through the decision model, and sends the first target detection model to the image acquisition module.
The determining, by the decision model, an implementation manner of determining the target detection model according to the first detection model state and the second detection model state may be: the decision model obtains a current state value according to the first detection model state and the second detection model state; and determining the target detection model according to the current state value. For example, the range of the state value may be [0,1], when the current state value falls between [0,0.5], the corresponding target detection model is the second detection model, and when the current state value falls between (0.5,1], the corresponding target detection model is the first detection model.
It should be noted that the order of occurrence of S302, S303 and S304, S305 is not specifically limited, and S302, S303 may occur before S304, S305, or S302, S303 may occur after S304, S305, or S302, S303 and S304, S305 may be performed simultaneously, and is not specifically limited.
After S307, the method may further include: the image detection system sends a third image in the target image sequence to the first target detection model through the image acquisition module; and the image detection system carries out image detection according to the third image through the first target detection model.
In one possible example, before the image detection system determines, through the decision model, an object detection model according to the model detection state and a historical model detection state and sends the object detection model to the image acquisition module when receiving the model detection state from the current detection model, the method further includes the following steps:
s401, the image detection system obtains a first image detection result according to the first image through the first detection model, outputs the first image detection result, and simultaneously sends the first image detection result to the correction model;
s402, the image detection system obtains a first reference image detection result according to the second image through the second detection model, and sends the first reference image detection result to the correction model;
and S403, the image detection system obtains a second image detection result aiming at the second image according to the first image detection result and the first reference image detection result through the correction model, and outputs the second image detection result.
401 and 402 are not in sequence, 401 may occur before 402, 401 may occur after 402, and 401 may occur simultaneously with 402.
As can be seen, in this example, after the image detection is performed by the image detection system through the low-precision detection model, the image detection result obtained by the low-precision detection model can be corrected by the correction module according to the image detection result obtained by the high-precision detection model, so that the accuracy of the image detection is improved.
In a possible example, the performing, by the image detection system, image detection according to the current image to be detected through the target detection model further includes: if the target detection model is the first detection model, the image detection system obtains a third image detection result according to the current image to be detected through the first detection model, outputs the third image detection result, and simultaneously sends the third image detection result to the correction model; if the target detection model is the second detection model, the image detection system obtains a second reference image detection result according to the current image to be detected through the second detection model, and sends the second reference image detection result to the correction model; and the correction model obtains a fourth image detection result of the current image to be detected according to the third image detection result and a second reference image detection result, and outputs the fourth image detection result.
For example, if the target image sequence includes 21 images, the current image to be detected is the 16 th image, the 1 st, 4 th and 11 th images of the first 16 previous images are the images processed by the first detection model, the rest images are the images processed by the second detection model, and the current detection model is the first detection model, the first detection model detects the 16 th image to obtain an image detection result (i.e., a third image detection result) corresponding to the 16 th image, outputs the image detection result corresponding to the 16 th image, and sends the image detection result corresponding to the 16 th image to the correction model.
For example, if the target image sequence includes 30 images, the current image to be detected is the 20 th image, among the first 19 images, the 1 st, 5 th and 15 th images are the images processed by the first detection model, the rest images are the images processed by the second detection model, the current detection model is the second detection model, the second detection model detects the 20 th image to obtain the reference image detection result corresponding to the 20 th image (corresponding to the second reference image detection result), and sends the reference image detection result corresponding to the 20 th image to the correction model; the correction model obtains the image detection result corresponding to the 20 th image (corresponding to the fourth image detection result) according to the image detection result corresponding to the 15 th image and the reference image detection result corresponding to the 20 th image, and outputs the image detection result corresponding to the 20 th image.
Further, the output of the image detection result may be: and displaying the image detection result on a display device, and/or broadcasting the image detection result through a voice broadcaster.
In one possible example, the method further comprises: the image detection system receives the target detection model from the decision model through the image acquisition model,
the image detection system sends the current image to be detected to the target detection model through the image acquisition model, wherein the current image to be detected is the last image in the target image sequence; if the target detection model is the first detection model, the image detection system obtains an image detection result according to the current image to be detected through the first detection model, and outputs the image detection result; if the target detection model is the second detection model, the image detection system obtains a first reference image detection result according to the current image to be detected through the second detection model, and sends the first reference image detection result to the correction model; and the image detection system obtains an image detection result aiming at the current image to be detected according to the first reference image detection result and the image detection result of the first detection model which is sent to the correction model last time by the correction model, and outputs the image detection result.
For example, the target image sequence includes 10 images, if the current image to be detected is the 10 th image of the 10 images, and the target detection model is the first detection model, the image acquisition model receives the first detection model from the decision model, and then sends the 10 th image to the first detection model, and the first detection model obtains an image detection result according to the 10 th image, and outputs the image detection result.
For example, the target image sequence includes 20 images, if the current image to be detected is the 20 th image of the 20 images, and the target detection model is the second detection model, the image acquisition model receives the second detection model from the decision model, and then sends the 20 th image to the second detection model, and the second detection model obtains an image detection result according to the 20 th image and outputs the image detection result.
As can be seen, in this example, when the currently processed image is the last image in the target image sequence, the image detection system can send the last image to the target detection module determined according to the policy module through the image acquisition model to output the detection result, so as to improve the intelligence of image detection. .
Referring to fig. 3 in accordance with the embodiment shown in fig. 2, fig. 3 is a schematic structural diagram of another image inspection system provided in an embodiment of the present application, and as shown in fig. 3, an image inspection system 300 includes an application processor 310, a memory 320, a communication interface 330, and one or more programs 321, where the one or more programs 321 are stored in the memory 320 and configured to be executed by the application processor 310, and the one or more programs 321 include instructions for performing the following steps;
when a model detection state from a current detection model is received through the decision model, determining a target detection model according to the model detection state and a historical model detection state, and sending the target detection model to the image acquisition module, wherein the current detected image is any image except a first image, a second image and a last image in a target image sequence, and the target detection model is one of the first detection model and the second detection model;
sending the current image to be detected to the target detection model through the image acquisition module;
and carrying out image detection according to the current image to be detected through the target detection model.
It can be seen that, in the embodiment of the present application, the method is applied to an image detection system, where the image detection system includes an image acquisition module, a first detection model, a second detection model, a modification model, and a decision model, and the method includes: when a model detection state from a current detection model is received through a decision model, determining a target detection model according to the model detection state and a historical model detection state, sending the target detection model to an image acquisition module, wherein the current detected image is any one of images except a first image, a second image and a last image in a target image sequence, the target detection model is one of a first detection model and a second detection model, then sending the current image to be detected to the target detection model through the image acquisition module, and finally, carrying out image detection through the target detection model according to the current image to be detected. Therefore, the image detection system of the embodiment of the application can analyze the model detection state of the most processed image of the detection models with different precisions through the decision model, and determine the detection model for processing the next image.
In one possible example, the one or more programs 321 include instructions for performing the following steps; acquiring an image detection request through the image acquisition module, wherein the image detection request comprises the target image sequence; sending a first image in the target image sequence to the first detection model according to the image detection request through the image acquisition module; obtaining a first detection model state according to the first image through the first detection model, and sending the first detection model state to the decision model; sending a second image in the target image sequence to the second detection model according to the image detection request through the image acquisition module; obtaining a second detection model state according to the second image through the second detection model, and sending the second detection model state to the decision model; and determining a first target detection model according to the first detection model state and the second detection model state through the decision model, and sending the first target detection model to the image acquisition module.
In one possible example, in terms of the image detection by the object detection model according to the current image to be detected, the instructions in the one or more programs 321 are specifically configured to perform the following operations: if the target detection model is the first detection model, obtaining a detection model state according to the current image to be detected through the first detection model; sending the detection model state to the decision model; if the target detection model is the second detection model, obtaining a detection model state according to the current image to be detected through the second detection model; sending the detection model state to the decision model.
In one possible example, in the aspect that the determining, by the decision model, a model detection state from the current detection model is received, determining a target detection model according to the model detection state and a historical model detection state, and sending the target detection model to the image acquisition module, the instructions in the one or more programs 321 are specifically configured to perform the following operations: if the current detection model is the first detection model, obtaining a first target model detection state in the historical model detection states through the decision model, and determining the target detection model according to the model detection state from the current detection model and the first target model detection state, wherein the first target model detection state is a last sent model detection state from the second detection model; and if the current detection model is the second detection model, acquiring a second target model detection state in the historical model detection states through the decision model, and determining the target detection model according to the model detection state from the current detection model and the second target model detection state, wherein the second target model detection state is the last sent model detection state from the first detection model.
In one possible example, the one or more programs 321 further include instructions for performing the following steps; when the decision model receives a model detection state from the current detection model, determining a target detection model according to the model detection state and a historical model detection state, and before sending the target detection model to the image acquisition module, obtaining a first image detection result according to the first image through the first detection model, outputting the first image detection result, and sending the first image detection result to the correction model; obtaining a first reference image detection result according to the second image through the second detection model, and sending the first reference image detection result to the correction model; and obtaining a second image detection result aiming at the second image according to the first image detection result and the first reference image detection result through the correction model, and outputting the first image detection result.
In one possible example, in terms of the image detection by the object detection model according to the current image to be detected, the instructions in the one or more programs 321 are specifically configured to perform the following operations: if the target detection model is the first detection model, obtaining a third image detection result according to the current image to be detected through the first detection model, outputting the third image detection result, and simultaneously sending the third image detection result to the correction model; if the target detection model is the second detection model, obtaining a second reference image detection result according to the current image to be detected through the second detection model, and sending the second reference image detection result to the correction model; and the correction model obtains a fourth image detection result of the current image to be detected according to the third image detection result and a second reference image detection result, and outputs the fourth image detection result.
In one possible example, the one or more programs 321 further include instructions for performing the following steps; after receiving the target detection model from the decision model through the image acquisition model, sending the current image to be detected to the target detection model, wherein the current image to be detected is the last image in the target image sequence; if the target detection model is the first detection model, obtaining an image detection result according to the current image to be detected through the first detection model, and outputting the image detection result; if the target detection model is the second detection model, obtaining a first reference image detection result according to the current image to be detected through the second detection model, and sending the first reference image detection result to the correction model; and obtaining an image detection result aiming at the current image to be detected according to the first reference image detection result and the image detection result of the first detection model which is sent to the correction model last time through the correction model, and outputting the image detection result.
The above description has introduced the solution of the embodiment of the present application mainly from the perspective of the method-side implementation process. It is understood that the electronic device includes hardware structures and software modules for performing the respective functions in order to realize the functions. Those of skill in the art will readily appreciate that the present application is capable of being implemented in hardware or a combination of hardware and computer software for carrying out the various example modules and algorithm steps described in connection with the embodiments provided herein. Whether a function is performed as hardware or computer software drives hardware depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.
In the embodiment of the present application, the electronic device may be divided into the functional modules according to the method example, for example, each functional module may be divided corresponding to each function, or two or more functions may be integrated into one processing module. The integrated module can be realized in a hardware mode, and can also be realized in a software functional module mode. It should be noted that, in the embodiment of the present application, the division of the module is schematic, and is only one logic function division, and there may be another division manner in actual implementation.
Referring to fig. 4, in accordance with the embodiment shown in fig. 2, fig. 4 is a block diagram of functional modules of an image detection system provided in an embodiment of the present application, as shown in fig. 4, the image detection system includes a processing unit 401 and a communication unit 402, wherein,
the processing unit 401 is configured to, when a model detection state from a current detection model is received through the communication unit 402 through the decision model, determine a target detection model according to the model detection state and a historical model detection state, and send the target detection model to the image acquisition module, where the current detected image is any image in a target image sequence except for a first image, a second image, and a last image, and the target detection model is one of the first detection model and the second detection model; the image acquisition module is used for sending the current image to be detected to the target detection model; and the target detection model is used for carrying out image detection according to the current image to be detected.
The image detection apparatus 400 may further include a storage unit 403 for storing program codes and data of the electronic device. The processing unit 401 may be a processor, the communication unit 402 may be a touch display screen or a transceiver, and the storage unit 403 may be a memory.
It can be seen that, in the embodiment of the present application, the image detection system includes an image acquisition module, a first detection model, a second detection model, a modification model, and a decision model, and the method includes: the image detection system comprises the steps that firstly, when a decision-making model receives a model detection state from a current detection model, a target detection model is determined according to the model detection state and a historical model detection state, the target detection model is sent to an image acquisition module, the current detected image is any image except a first image, a second image and a last image in a target image sequence, the target detection model is one of a first detection model and a second detection model, then the current image to be detected is sent to the target detection model through the image acquisition module, and finally, image detection is carried out through the target detection model according to the current image to be detected. Therefore, the image detection system of the embodiment of the application can analyze the model detection state of the most processed image of the detection models with different precisions through the decision model, and determine the detection model for processing the next image.
In one possible example, the processing unit 401 is further configured to: when a model detection state from the current detection model is received through the decision model, determining a target detection model according to the model detection state and a historical model detection state, and before the target detection model is sent to the image acquisition module, acquiring an image detection request through the image acquisition module, wherein the image detection request comprises the target image sequence; sending a first image in the target image sequence to the first detection model through the image acquisition module according to the image detection request; obtaining a first detection model state according to the first image through the first detection model, and sending the first detection model state to the decision model; sending a second image in the target image sequence to the second detection model according to the image detection request through the image acquisition module; obtaining a second detection model state according to the second image through the second detection model, and sending the second detection model state to the decision model; and determining a first target detection model according to the first detection model state and the second detection model state through the decision model, and sending the first target detection model to the image acquisition module.
In a possible example, in terms of the image detection performed by the object detection model according to the current image to be detected, the processing unit 401 is specifically configured to: if the target detection model is the first detection model, obtaining a detection model state according to the current image to be detected through the first detection model and sending the detection model state to the decision model; and if the target detection model is the second detection model, obtaining a detection model state according to the current image to be detected through the second detection model and sending the detection model state to the decision model.
In a possible example, in the aspect that when the model detection state from the current detection model is received through the decision model, a target detection model is determined according to the model detection state and a historical model detection state, and the target detection model is sent to the image acquisition module, the processing unit 401 is specifically configured to: if the current detection model is the first detection model, acquiring a first target model detection state in the historical model detection states, and determining the target detection model according to the model detection state from the current detection model and the first target model detection state, wherein the first target model detection state is the last sent model detection state from the second detection model; and if the current detection model is the second detection model, acquiring a second target model detection state in the historical model detection states, and determining the target detection model according to the model detection state from the current detection model and the second target model detection state, wherein the second target model detection state is the last sent model detection state from the first detection model.
In one possible example, the processing unit 401 is further configured to: when a model detection state from the current detection model is received through the decision model, determining a target detection model according to the model detection state and a historical model detection state, and before the target detection model is sent to the image acquisition module, obtaining a first image detection result according to the first image through the first detection model, outputting the first image detection result, and simultaneously sending the first image detection result to the correction model; obtaining a first reference image detection result according to the second image through the second detection model, and sending the first reference image detection result to the correction model; and obtaining a second image detection result aiming at the second image according to the first image detection result and the first reference image detection result through the correction model, and outputting the first image detection result.
In a possible example, in terms of the image detection performed by the object detection model according to the current image to be detected, the processing unit 401 is specifically configured to: if the target detection model is the first detection model, obtaining a third image detection result according to the current image to be detected through the first detection model, outputting the third image detection result, and simultaneously sending the third image detection result to the correction model; if the target detection model is the second detection model, obtaining a second reference image detection result according to the current image to be detected through the second detection model, and sending the second reference image detection result to the correction model; and the correction model obtains a fourth image detection result of the current image to be detected according to the third image detection result and a second reference image detection result, and outputs the fourth image detection result.
In one possible example, the processing unit 401 is further configured to: after receiving the target detection model from the decision model through the image acquisition model, sending the current image to be detected to the target detection model, wherein the current image to be detected is the last image in the target image sequence; if the target detection model is the first detection model, obtaining an image detection result according to the current image to be detected through the first detection model, and outputting the image detection result; if the target detection model is the second detection model, obtaining a first reference image detection result according to the current image to be detected through the second detection model, and sending the first reference image detection result to the correction model; and obtaining an image detection result aiming at the current image to be detected according to the first reference image detection result and the image detection result of the first detection model which is sent to the correction model last time through the correction model, and outputting the image detection result.
Embodiments of the present application also provide a computer storage medium, wherein the computer storage medium stores a computer program for electronic data exchange, the computer program enabling a computer to execute part or all of the steps of any one of the methods as described in the above method embodiments, and the computer includes an image detection system.
Embodiments of the present application also provide a computer program product comprising a non-transitory computer readable storage medium storing a computer program operable to cause a computer to perform some or all of the steps of any of the methods as described in the above method embodiments. The computer program product may be a software installation package, said computer comprising the image detection system.
It should be noted that, for simplicity of description, the above-mentioned method embodiments are described as a series of acts or combination of acts, but those skilled in the art will recognize that the present application is not limited by the order of acts described, as some steps may occur in other orders or concurrently depending on the application. Further, those skilled in the art should also appreciate that the embodiments described in the specification are preferred embodiments and that the acts and modules referred to are not necessarily required in this application.
In the foregoing embodiments, the descriptions of the respective embodiments have respective emphasis, and for parts that are not described in detail in a certain embodiment, reference may be made to related descriptions of other embodiments.
In the embodiments provided in the present application, it should be understood that the disclosed apparatus may be implemented in other manners. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the above-described modules is merely a logical division, and other divisions may be realized in practice, for example, a plurality of modules 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 of devices or modules through some interfaces, and may be in an electrical or other form.
The modules described as separate parts may or may not be physically separate, and parts displayed as modules may or may not be physical modules, may be located in one place, or may be distributed on a plurality of model modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment.
In addition, functional modules in the embodiments of the present application may be integrated into one processing module, or each of the modules may exist alone physically, or two or more modules are integrated into one module. The integrated module can be realized in a hardware mode, and can also be realized in a software functional module mode.
The integrated modules, if implemented in the form of software functional modules and sold or used as separate products, may be stored in a computer readable memory. 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 the form of a software product, which is stored in a memory and includes instructions for causing a computer device (which may be a personal computer, a server, a model device, or the like) to execute all or part of the steps of the method described in the embodiments of the present application. And the aforementioned memory comprises: a U-disk, a Read-only memory (ROM), a Random Access Memory (RAM), a removable hard disk, a magnetic or optical disk, and other various media capable of storing program codes.
Those skilled in the art will appreciate that all or part of the steps in the methods of the above embodiments may be implemented by associated hardware instructed by a program, which may be stored in a computer-readable memory, which may include: flash memory disks, Read-only memories (ROMs), Random Access Memories (RAMs), magnetic or optical disks, and the like.
The foregoing detailed description of the embodiments of the present application has been presented to illustrate the principles and implementations of the present application, and the above description of the embodiments is only provided to help understand the method and the core concept of the present application; meanwhile, for a person skilled in the art, according to the idea of the present application, there may be variations in the specific embodiments and the application scope, and as described above, the content of the present specification should not be construed as a limitation to the present application.

Claims (10)

1. An image detection method is characterized in that the method is applied to an image detection system, the image detection system comprises an image acquisition module, a first detection model, a second detection model, a correction model and a decision model, the image acquisition module is respectively in communication connection with the first detection model, the second detection model and the decision model, the first detection model is respectively in communication connection with the correction model and the decision model, the second detection model is respectively in communication connection with the correction model and the decision model, and the correction model is in communication connection with the decision model, the method comprises the following steps:
when a model detection state from a current detection model is received through the decision model, determining a target detection model according to the model detection state and a historical model detection state, and sending the target detection model to the image acquisition module, wherein the current detected image is any image except a first image, a second image and a last image in a target image sequence, and the target detection model is one of the first detection model and the second detection model;
sending the current image to be detected to the target detection model through the image acquisition module;
and carrying out image detection according to the current image to be detected through the target detection model.
2. The method of claim 1, wherein before the determining, by the decision model, a target detection model based on the model detection state and a historical model detection state upon receiving a model detection state from the current detection model and sending the target detection model to the image acquisition module, the method further comprises:
acquiring an image detection request through the image acquisition module, wherein the image detection request comprises the target image sequence;
sending a first image in the target image sequence to the first detection model according to the image detection request through the image acquisition module;
obtaining a first detection model state according to the first image through the first detection model, and sending the first detection model state to the decision model;
sending a second image in the target image sequence to the second detection model according to the image detection request through the image acquisition module;
obtaining a second detection model state according to the second image through the second detection model, and sending the second detection model state to the decision model;
and determining a first target detection model according to the first detection model state and the second detection model state through the decision model, and sending the first target detection model to the image acquisition module.
3. The method according to claim 1 or 2, wherein the image detection according to the current image to be detected by the object detection model comprises:
if the target detection model is the first detection model, obtaining a detection model state according to the current image to be detected through the first detection model; sending the detection model state to the decision model;
if the target detection model is the second detection model, obtaining a detection model state according to the current image to be detected through the second detection model; sending the detection model state to the decision model.
4. The method of claim 1, wherein the determining, by the decision model, a target detection model according to the model detection state and a historical model detection state when receiving the model detection state from the current detection model and sending the target detection model to the image acquisition module comprises:
if the current detection model is the first detection model, obtaining a first target model detection state in the historical model detection states through the decision model, and determining the target detection model according to the model detection state from the current detection model and the first target model detection state, wherein the first target model detection state is a last sent model detection state from the second detection model;
and if the current detection model is the second detection model, acquiring a second target model detection state in the historical model detection states through the decision model, and determining the target detection model according to the model detection state from the current detection model and the second target model detection state, wherein the second target model detection state is the last sent model detection state from the first detection model.
5. The method of claim 3, wherein before the determining, by the decision model, a target detection model based on the model detection state and a historical model detection state upon receiving a model detection state from the current detection model and sending the target detection model to the image acquisition module, the method further comprises:
obtaining a first image detection result according to the first image through the first detection model, outputting the first image detection result, and simultaneously sending the first image detection result to the correction model;
obtaining a first reference image detection result according to the second image through the second detection model, and sending the first reference image detection result to the correction model;
and obtaining a second image detection result aiming at the second image according to the first image detection result and the first reference image detection result through the correction model, and outputting the first image detection result.
6. The method according to claim 4, wherein the image detection is performed by the object detection model according to the current image to be detected, further comprising:
if the target detection model is the first detection model, obtaining a third image detection result according to the current image to be detected through the first detection model, outputting the third image detection result, and simultaneously sending the third image detection result to the correction model;
if the target detection model is the second detection model, obtaining a second reference image detection result according to the current image to be detected through the second detection model, and sending the second reference image detection result to the correction model;
and obtaining a fourth image detection result of the current image to be detected according to the third image detection result and a second reference image detection result through the correction model, and outputting the fourth image detection result.
7. The method of claim 1, further comprising:
after receiving the target detection model from the decision model through the image acquisition model, sending the current image to be detected to the target detection model, wherein the current image to be detected is the last image in the target image sequence;
if the target detection model is the first detection model, obtaining an image detection result according to the current image to be detected through the first detection model, and outputting the image detection result;
if the target detection model is the second detection model, obtaining a first reference image detection result according to the current image to be detected through the second detection model, and sending the first reference image detection result to the correction model; and the correction model obtains an image detection result aiming at the current image to be detected according to the first reference image detection result and the image detection result which is sent by the first detection model to the correction model last time, and outputs the image detection result.
8. An image inspection system, comprising an image acquisition module, a first inspection model, a second inspection model, a modification model, and a decision model, wherein the image acquisition module is in communication with the first inspection model, the second inspection model, and the decision model, respectively, wherein the first inspection model is in communication with the modification model and the decision model, wherein the second inspection model is in communication with the modification model and the decision model, respectively, wherein the modification model is in communication with the decision model,
the decision-making model is used for determining a target detection model according to the model detection state and the historical model detection state when receiving the model detection state from the current detection model, and sending the target detection model to the image acquisition module, wherein the current detected image is any image except a first image, a second image and a last image in a target image sequence, and the target detection model is one of the first detection model and the second detection model;
the image acquisition module is used for sending the current image to be detected to the target detection model;
and the target detection model is used for carrying out image detection according to the current image to be detected.
9. An image detection system comprising a processor, a memory, a communication interface, and one or more programs stored in the memory and configured to be executed by the processor, the programs including instructions for performing the steps in the method of any of claims 1-7.
10. A computer-readable storage medium, characterized in that a computer program for electronic data exchange is stored, wherein the computer program causes a computer to perform the method according to any one of claims 1-7.
CN201911424558.1A 2019-12-30 2019-12-30 Image detection method and related product Pending CN113128302A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911424558.1A CN113128302A (en) 2019-12-30 2019-12-30 Image detection method and related product

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911424558.1A CN113128302A (en) 2019-12-30 2019-12-30 Image detection method and related product

Publications (1)

Publication Number Publication Date
CN113128302A true CN113128302A (en) 2021-07-16

Family

ID=76769829

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911424558.1A Pending CN113128302A (en) 2019-12-30 2019-12-30 Image detection method and related product

Country Status (1)

Country Link
CN (1) CN113128302A (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2018195797A1 (en) * 2017-04-26 2018-11-01 深圳配天智能技术研究院有限公司 Visual detection method, detection device, and robot
CN108764235A (en) * 2018-05-23 2018-11-06 中国民用航空总局第二研究所 Neural network model, object detection method, equipment and medium
CN110458866A (en) * 2019-08-13 2019-11-15 北京积加科技有限公司 Target tracking method and system
WO2019223582A1 (en) * 2018-05-24 2019-11-28 Beijing Didi Infinity Technology And Development Co., Ltd. Target detection method and system

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2018195797A1 (en) * 2017-04-26 2018-11-01 深圳配天智能技术研究院有限公司 Visual detection method, detection device, and robot
CN108764235A (en) * 2018-05-23 2018-11-06 中国民用航空总局第二研究所 Neural network model, object detection method, equipment and medium
WO2019223582A1 (en) * 2018-05-24 2019-11-28 Beijing Didi Infinity Technology And Development Co., Ltd. Target detection method and system
CN110458866A (en) * 2019-08-13 2019-11-15 北京积加科技有限公司 Target tracking method and system

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
牛文昊;王士林;: "多监控摄像头特定目标检索方法研究", 通信技术, no. 03 *
车宏;孙隆和;: "焦平面红外弱小目标检测新算法", 火力与指挥控制, no. 09 *

Similar Documents

Publication Publication Date Title
CN109583391B (en) Key point detection method, device, equipment and readable medium
CN112101305B (en) Multi-path image processing method and device and electronic equipment
CN113608641B (en) Method and device for adjusting display position of curved screen, intelligent sound box and storage medium
CN111222509B (en) Target detection method and device and electronic equipment
CN111327913B (en) Message processing method and device and electronic equipment
CN110363748B (en) Method, device, medium and electronic equipment for processing dithering of key points
US20220383637A1 (en) Live streaming sampling method and apparatus, and electronic device
KR102365431B1 (en) Electronic device for providing target video in sports play video and operating method thereof
CN115205925A (en) Expression coefficient determining method and device, electronic equipment and storage medium
CN110740315B (en) Camera correction method and device, electronic equipment and storage medium
CN111881740A (en) Face recognition method, face recognition device, electronic equipment and medium
CN112926083B (en) Interactive processing method based on building information model and related device
CN111191556A (en) Face recognition method and device and electronic equipment
CN114584836B (en) Method, device, system and medium for detecting using behavior of electronic product
CN104978731A (en) Information processing method and electronic equipment
CN110751120A (en) Detection method and device and electronic equipment
CN113128302A (en) Image detection method and related product
CN110933314A (en) Focus-following shooting method and related product
EP3629577A1 (en) Data transmission method, camera and electronic device
CN111832354A (en) Target object age identification method and device and electronic equipment
CN115474229A (en) Quality determination method and device for wireless network, electronic equipment and storage medium
CN112115740B (en) Method and apparatus for processing image
CN111199179B (en) Target object tracking method, terminal equipment and medium
CN112910875A (en) Display method and device
CN111832256A (en) Information marking method for examining picture and related device

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