CN114445768A - Target identification method and device, electronic equipment and storage medium - Google Patents
Target identification method and device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN114445768A CN114445768A CN202111648725.8A CN202111648725A CN114445768A CN 114445768 A CN114445768 A CN 114445768A CN 202111648725 A CN202111648725 A CN 202111648725A CN 114445768 A CN114445768 A CN 114445768A
- Authority
- CN
- China
- Prior art keywords
- target
- image
- images
- detection frame
- training
- 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
- 238000000034 method Methods 0.000 title claims abstract description 73
- 238000003860 storage Methods 0.000 title claims abstract description 16
- 238000001514 detection method Methods 0.000 claims abstract description 138
- 238000012545 processing Methods 0.000 claims abstract description 43
- 238000007781 pre-processing Methods 0.000 claims abstract description 16
- 230000011218 segmentation Effects 0.000 claims abstract description 13
- 238000012549 training Methods 0.000 claims description 77
- 238000012795 verification Methods 0.000 claims description 39
- 238000004590 computer program Methods 0.000 claims description 21
- 238000003062 neural network model Methods 0.000 claims description 18
- 238000005520 cutting process Methods 0.000 claims description 15
- 238000013145 classification model Methods 0.000 claims description 14
- 238000001914 filtration Methods 0.000 claims description 11
- 230000002708 enhancing effect Effects 0.000 claims description 7
- 238000011156 evaluation Methods 0.000 abstract description 7
- 238000010586 diagram Methods 0.000 description 15
- 230000006870 function Effects 0.000 description 6
- 238000012544 monitoring process Methods 0.000 description 6
- 230000000694 effects Effects 0.000 description 5
- 238000012986 modification Methods 0.000 description 5
- 230000004048 modification Effects 0.000 description 5
- 230000008569 process Effects 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 238000003709 image segmentation Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000012552 review Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Image Analysis (AREA)
Abstract
The embodiment of the application provides a target identification method, a target identification device, electronic equipment and a storage medium, wherein the method comprises the following steps: by acquiring a target image; preprocessing a target image to obtain a preprocessed target image; carrying out segmentation and scaling processing on the preprocessed target image to obtain a plurality of local images with the same size; respectively carrying out target detection on the plurality of local images to obtain a plurality of target detection frame images corresponding to a plurality of targets; the method comprises the steps of classifying a plurality of target detection frame images to obtain the categories of target objects contained in the plurality of detection frame images, splitting complex scenes by dividing the target images in and out, improving the detection capability of small target objects and the generalization performance of the small target objects to different scenes, improving the application space of the target identification method, and based on the combination of target detection and classification, carrying out secondary result evaluation on detection results, reducing false reports and comprehensively improving the accuracy of target identification.
Description
Technical Field
The present application relates to the field of target identification technologies, and in particular, to a target identification method and apparatus, an electronic device, and a storage medium.
Background
Based on city safety, more and more cameras are used in urban area, and in video monitoring, the image picture is along with light, the time difference, and the change of environment such as weather produces huge difference, and then leads to the data of gathering through the control picture to produce very big interference to the discernment ability of target, how to counter light, weather interference, effectively show the technical problem that target detail characteristic becomes present and await solution. In addition, at present, for the detection of a plurality of targets in a monitoring picture, under a scene with a larger monitoring picture, the detection effect is poor, the objects are easily interfered by surrounding similar objects or are hidden by light rays to cause false alarm, or the objects are missed because the objects cannot be detected, so that the effective characteristics of the plurality of targets are difficult to accurately and effectively identify at one time.
Disclosure of Invention
The embodiment of the application provides a target identification method, a target identification device, an electronic device and a storage medium, and based on the combination of target detection and classification, secondary result evaluation can be performed on a detection result, false alarm is reduced, and the accuracy of target identification can be improved.
A first aspect of an embodiment of the present application provides a target identification method, where the method includes:
acquiring a target image;
preprocessing the target image to obtain a preprocessed target image;
carrying out segmentation and scaling processing on the preprocessed target image to obtain a plurality of local images with the same size;
respectively carrying out target detection on the plurality of local images to obtain a plurality of target detection frame images corresponding to a plurality of targets;
and classifying the multiple target detection frame images to obtain the categories of target objects contained in the multiple detection frame images.
A second aspect of the embodiments of the present application provides an object recognition apparatus, including:
an acquisition unit configured to acquire a target image;
the processing unit is used for preprocessing the target image to obtain a preprocessed target image;
the processing unit is further configured to perform segmentation and scaling on the preprocessed target image to obtain a plurality of local images with the same size;
the detection unit is used for respectively carrying out target detection on the plurality of local images to obtain a plurality of target detection frame images corresponding to a plurality of targets;
and the classification unit is used for classifying the target detection frame images to obtain the classes of the target objects contained in the detection frame images.
A third aspect of the present application provides an electronic device comprising: a processor and a memory; and one or more programs stored in the memory and configured to be executed by the processor, the programs including instructions for some or all of the steps as described in the first aspect.
In a fourth aspect, the present application provides a computer-readable storage medium, where the computer-readable storage medium is used to store a computer program, where the computer program is used to make a computer execute some or all of the steps described in the first aspect of the present application.
In a fifth aspect, embodiments of the present application provide a computer program product, where the computer program product comprises a non-transitory computer-readable storage medium storing a computer program, the computer program being operable to cause a computer to perform some or all of the steps as described in the first aspect of embodiments of the present application. The computer program product may be a software installation package.
The embodiment of the application has the following beneficial effects:
it can be seen that, by the target identification method, apparatus, electronic device and storage medium described in the embodiments of the present application, a target image is obtained; preprocessing a target image to obtain a preprocessed target image; carrying out segmentation and scaling processing on the preprocessed target image to obtain a plurality of local images with the same size; respectively carrying out target detection on the plurality of local images to obtain a plurality of target detection frame images corresponding to a plurality of targets; the method comprises the steps of classifying a plurality of target detection frame images to obtain the categories of target objects contained in the plurality of detection frame images, splitting complex scenes by dividing the target images in and out, improving the detection capability of small target objects and the generalization performance of the small target objects to different scenes, improving the application space of the target identification method, and based on the combination of target detection and classification, carrying out secondary result evaluation on detection results, reducing false reports and comprehensively improving the accuracy of target identification.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.
Fig. 1 is a schematic flowchart of an embodiment of a target identification method provided in an embodiment of the present application;
FIG. 2 is a schematic diagram illustrating a segmentation of a preprocessed target image according to an embodiment of the present disclosure;
FIG. 3 is a schematic diagram of cropping and scaling a training sample image and a verification sample image according to an embodiment of the present application;
FIG. 4 is a flowchart illustrating an embodiment of another object recognition method provided in an embodiment of the present application;
FIG. 5 is a flowchart illustrating an embodiment of another object recognition method provided in an embodiment of the present application;
FIG. 6 is a flowchart illustrating an embodiment of another object recognition method provided in an embodiment of the present application;
fig. 7 is a schematic structural diagram of an embodiment of an electronic device according to an embodiment of the present application;
fig. 8 is a schematic structural diagram of an embodiment of an object recognition device according to an embodiment of the present application.
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 some, but not all, embodiments of the present application. 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," "third," and "fourth," etc. in the description and claims of this application and in the accompanying 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.
Please refer to fig. 1, which is a flowchart illustrating an embodiment of a target identification method according to an embodiment of the present application. The target identification method described in this embodiment includes the following steps:
101. and acquiring a target image.
The target image is an image acquired by monitoring a target in a target identification scene. Specifically, in each area of a city, a target image acquired by a camera for a monitoring area can be acquired.
102. And preprocessing the target image to obtain a preprocessed target image.
Wherein the pretreatment may include at least one of: filtering processing, scaling processing, image enhancement processing, and the like.
Optionally, in the step 102, the preprocessing the target image to obtain a preprocessed target image may include the following steps:
21. carrying out low-pass filtering processing on the target image to obtain a first partial image characteristic;
22. acquiring a second partial image feature except the first partial image feature in the target image;
23. enhancing the second partial image features to obtain enhanced second partial image features;
24. and recombining the first partial image characteristic and the enhanced second partial image characteristic to obtain an enhanced target image.
The image features of the target image can be divided into two parts, namely a low-frequency part obtained by image low-pass filtering (smooth blurring) and a high-frequency part obtained by subtracting the low-frequency part from the original target image. Specifically, the target image is subjected to low-pass filtering processing, the obtained first partial image feature is a low-frequency partial image feature, the second partial image feature is a high-frequency partial image feature, the high-frequency partial image feature contains more image details, and the second partial image feature can be subjected to Enhancement processing, for example, an Adaptive Contrast Enhancement (ACE) algorithm can be adopted, so that the Contrast of the target image can be enhanced, the influence of changes of environments such as light, time difference and weather on the image quality can be reduced, the quality of the target image can be improved, the feature information of the target object can be highlighted, and the feature extraction capability and the resolution effect can be improved.
Optionally, in step 23, performing enhancement processing on the second partial image feature to obtain an enhanced second partial image feature, where the method includes:
multiplying the image characteristic value of the second partial image characteristic by the gain value to obtain a gained image characteristic value;
and obtaining the enhanced second partial image characteristic based on the gained image characteristic value.
Wherein the gain value may be set to some fixed value, or a quantity related to the variance.
103. And carrying out segmentation and scaling processing on the preprocessed target image to obtain a plurality of local images with the same size.
As shown in fig. 2, the preprocessed target image may be segmented, and the target image may be segmented into a plurality of partial images, that is, a complex scene may be segmented into a plurality of simple parts, and specifically, the target image may be directly segmented into a plurality of partial images with equal size; alternatively, the target image is divided into a plurality of partial images, and then the partial images are enlarged or reduced to the same size to perform target detection on the plurality of partial images of the same size.
In the specific implementation, when the size of a picture transmitted by monitoring is too large, the scene is complex, the detection effect on small targets is poor, and the detection is easy to miss, so that the small targets in a larger scene can be effectively detected only by a larger resolution and a relatively simple environment.
Optionally, in step 103, the segmenting and scaling the preprocessed target image to obtain a plurality of local images with the same size includes:
dividing the preprocessed target image into a plurality of local images;
and amplifying the plurality of segmented local images to the same size by adopting a linear interpolation method to obtain a plurality of local images with the same size.
In specific implementation, the target image may be divided into a plurality of local images according to image details of different sizes in the target image, the plurality of local images may be the same or different, and then, a linear interpolation method is used to amplify, according to a first preset size, the local image different from the first preset size in an in-out manner, so as to obtain the local image the same as the first preset size, thereby obtaining the plurality of local images the same in size.
104. And respectively carrying out target detection on the plurality of local images to obtain a plurality of target detection frame images corresponding to the plurality of targets.
The multiple local images with the same size may be input to a detection model, and a detection result may be output, where the detection model may be, for example, a yolo (young Only Look once) model, and may output detected targets, and each target may be marked in a corresponding local image to obtain multiple target detection frame images.
105. And classifying the plurality of target detection frame images to obtain the types of target objects contained in the plurality of detection frame images.
After the target detection is carried out on the local images, in order to ensure the accuracy of target identification, the images of the target detection frames can be continuously classified by adopting a classification model so as to carry out secondary evaluation on the target.
In the specific implementation, under complex scenes such as streets, parks and communities, a plurality of targets are recognized at one time, and false recognition is easily caused by mutual interference among the targets and influence of environmental transformation.
Optionally, in step 105, the classifying the multiple targets to obtain the classes of the target objects included in the multiple detection frame images includes:
for each target detection frame image, zooming and cutting the target detection frame image to obtain a processed image corresponding to the target detection frame image;
extracting target image features in the processed image corresponding to each target detection frame image; and inputting the target image characteristics into a classification model, and outputting the categories of the target objects contained in the target detection frame images to obtain the categories of the target objects contained in the plurality of detection frame images.
In a specific implementation, the backbone model may be used as a classification model to classify a plurality of target objects, specifically, each target detection frame image may be scaled and cropped to obtain a processed image corresponding to the target detection frame image, specifically, the target detection frame image may be scaled first to obtain a reduced or enlarged image, then the reduced or enlarged image is cropped according to a second preset size to obtain an image of the second preset size, further, target image features in the cropped image may be extracted, the target image features may be input into the classification model to output a class of the target object included in the target detection frame image, and if the class is the class corresponding to the specified target object, it may be determined that the specified target object is detected, and the target object may be reported, and similarly, the target image features of the plurality of images may be classified, determining the types of target objects contained in the plurality of detection frame images, and judging whether each target object is a specified target object.
Optionally, the embodiment of the present application further includes a step of training and verifying the classification model, which is specifically as follows:
a training sample set and a verification sample set are obtained in advance, the training sample set comprises training sample images to be trained, and the verification sample set comprises verification sample images to be verified;
randomly cutting the training sample images in the training sample set to obtain cut regional training images; carrying out zooming processing on the region image to obtain a region training image with a target size;
amplifying the verification sample images in the verification sample set to obtain amplified images; cutting the central area of the amplified image to obtain an area verification image with the same size as the target;
and carrying out classification training on the preset neural network model according to the region training image with the target size, and verifying the preset neural network model according to the region verification image with the target size to obtain a classification model.
In the embodiment of the application, a preset neural network model can be trained and verified in advance, and specifically, training sample images in a training sample set can be cut randomly to obtain cut regional training images; and carrying out zooming processing on the region images to obtain region training images with target sizes, so that sizes of training sample images in a training sample set for training are consistent, and training differences caused by different sizes of objects are avoided. Amplifying the verification sample images in the verification sample set to obtain an amplified image; and cutting the central area of the amplified image to obtain an area verification image with the same size as the target size, so that the sizes of the verification sample images in the verification sample set for verification are consistent, and the verification result difference caused by different sizes of objects is avoided. Specifically, as shown in fig. 3, a region of the training sample image in the training sample set is randomly cut out, and then the training sample image is scaled to 224 × 224, the verification sample image in the verification sample set is scaled up (for example, scaled up to 256 × 256 or 384 × 384), and then an image block (224 × 224) with a fixed size is cut out in the central region, so that the input size of the target object is similar during training and verification of the neural network model, and a learning error caused by a large difference in object resolution is avoided.
Optionally, the classification training of the preset neural network model according to the region training image of the target size may include the following steps:
extracting a region training characteristic sample in the region training image;
and training a preset neural network model according to the regional training characteristic samples and the corresponding result samples.
It can be seen that, by the target identification method provided by the embodiment of the application, the target image is obtained; preprocessing a target image to obtain a preprocessed target image; carrying out segmentation and scaling processing on the preprocessed target image to obtain a plurality of local images with the same size; respectively carrying out target detection on the plurality of local images to obtain a plurality of target detection frame images corresponding to a plurality of targets; the method comprises the steps of classifying a plurality of target detection frame images to obtain the categories of target objects contained in the plurality of detection frame images, splitting complex scenes by dividing the target images in and out, improving the detection capability of small target objects and the generalization performance of the small target objects to different scenes, improving the application space of the target identification method, and based on the combination of target detection and classification, carrying out secondary result evaluation on detection results, reducing false reports and comprehensively improving the accuracy of target identification.
In accordance with the above, please refer to fig. 4, which is a flowchart illustrating an embodiment of a target identification method according to an embodiment of the present application. The target identification method described in this embodiment includes the following steps:
201. and acquiring a target image.
202. And carrying out low-pass filtering processing on the target image to obtain a first partial image characteristic.
203. A second partial image feature other than the first partial image feature in the target image is acquired.
204. And performing enhancement processing on the second partial image features to obtain enhanced second partial image features.
205. And recombining the first partial image characteristic and the enhanced second partial image characteristic to obtain an enhanced target image.
206. And dividing the enhanced target image into a plurality of local images.
207. And amplifying the plurality of segmented local images to the same size by adopting a linear interpolation method to obtain a plurality of local images with the same size.
208. And respectively carrying out target detection on the plurality of local images to obtain a plurality of target detection frame images corresponding to the plurality of targets.
209. And classifying the plurality of target detection frame images to obtain the types of target objects contained in the plurality of detection frame images.
The detailed description of steps 201 to 209 may refer to the corresponding steps from step 101 to step 105 of the target identification method described in fig. 1, and is not repeated herein.
In specific implementation, as shown in fig. 5, fig. 5 is a schematic flow diagram of an embodiment of another target identification method, where a target image may be obtained, and the target image is preprocessed by an ACE algorithm to obtain a preprocessed target image; then, carrying out image segmentation and scaling processing to obtain a plurality of local images with the same size; respectively inputting each local image into a detection model for target detection, if a specified target object is detected, extracting a target detection frame image, and zooming and cutting the target detection frame image to obtain a processed image; and inputting the processed image into a classification model to obtain the classification of the target object included in the processed image, and reporting if the classification of the specified target object is determined according to the classification of the target object.
It can be seen that, by the target identification method provided by the embodiment of the application, the target image is obtained; carrying out low-pass filtering processing on the target image to obtain a first partial image characteristic; acquiring a second partial image feature except the first partial image feature in the target image; enhancing the second partial image features to obtain enhanced second partial image features; recombining the first partial image characteristic and the enhanced second partial image characteristic to obtain an enhanced target image; therefore, the contrast of the target image can be enhanced, the influence of the change of the environments such as light, time difference, weather and the like on the image quality is reduced, the quality of the target image is improved, the characteristic information of the target object is highlighted, and the characteristic extraction capability and the distinguishing effect are improved. Dividing the enhanced target image into a plurality of local images; amplifying the segmented local images to the same size by adopting a linear interpolation method to obtain a plurality of local images with the same size; therefore, by adopting image segmentation, the small target resolution in the target image can be improved, the complex scene can be split, the small target object in the complex scene can be accurately detected, the generalization performance of target identification to different scenes can be improved, and the application space of the target identification can be improved. Respectively carrying out target detection on the plurality of local images to obtain a plurality of target detection frame images corresponding to a plurality of targets; and classifying the plurality of target detection frame images to obtain the categories of target objects contained in the plurality of detection frame images, so that the detected targets can be evaluated again to ensure the accuracy of target identification.
In accordance with the above, please refer to fig. 6, which is a flowchart illustrating an embodiment of a target identification method according to an embodiment of the present application. The target identification method described in this embodiment includes the following steps:
301. and acquiring a target image.
302. And carrying out low-pass filtering processing on the target image to obtain the first partial image characteristics.
303. A second partial image feature other than the first partial image feature in the target image is acquired.
304. And performing enhancement processing on the second partial image features to obtain enhanced second partial image features.
305. And recombining the first partial image characteristic and the enhanced second partial image characteristic to obtain an enhanced target image.
306. And dividing the enhanced target image into a plurality of local images.
307. And amplifying the plurality of segmented local images to the same size by adopting a linear interpolation method to obtain a plurality of local images with the same size.
308. And respectively carrying out target detection on the plurality of local images to obtain a plurality of target detection frame images corresponding to the plurality of targets.
309. And for each target detection frame image, carrying out zooming and cutting processing on the target detection frame image to obtain a processed image corresponding to the target detection frame image.
310. Extracting target image features in the processed image corresponding to each target detection frame image; and inputting the characteristics of the target images into the classification model, and outputting the categories of the target objects contained in the target detection frame images to obtain the categories of the target objects contained in the plurality of detection frame images.
The detailed description of the steps 301 to 310 may refer to the corresponding steps from step 101 to step 105 of the target identification method described in fig. 1, and will not be described herein again.
It can be seen that, by the target identification method provided by the embodiment of the application, the target image is obtained; carrying out low-pass filtering processing on the target image to obtain a first partial image characteristic; acquiring a second partial image characteristic except the first partial image characteristic in the target image; enhancing the second partial image features to obtain enhanced second partial image features; recombining the first partial image characteristic and the enhanced second partial image characteristic to obtain an enhanced target image; therefore, the contrast of the target image can be enhanced, the influence of the change of environments such as light, time difference, weather and the like on the image quality is reduced, the quality of the target image is improved, the characteristic information of the target object is highlighted, and the characteristic extraction capability and the distinguishing effect are improved. Dividing the enhanced target image into a plurality of local images; amplifying the plurality of segmented local images to the same size by adopting a linear interpolation method to obtain a plurality of local images with the same size; therefore, by adopting image segmentation, the small target resolution in the target image can be improved, the complex scene can be split, the small target object in the complex scene can be accurately detected, the generalization performance of target identification to different scenes can be improved, and the application space of the target identification can be improved. Respectively carrying out target detection on the plurality of local images to obtain a plurality of target detection frame images corresponding to a plurality of targets; and classifying the plurality of target detection frame images to obtain the categories of target objects contained in the plurality of detection frame images, so that the detected targets can be evaluated again to ensure the accuracy of target identification.
In accordance with the above, the following is a device for implementing the object recognition method, specifically as follows:
fig. 7 is a schematic structural diagram of an embodiment of an electronic device according to an embodiment of the present disclosure. The electronic device 400 described in this embodiment includes: at least one input device 1000; at least one output device 2000; at least one processor 3000, e.g., a CPU; and a memory 4000, the input device 1000, the output device 2000, the processor 3000, and the memory 4000 being connected by a bus 5000.
The input device 1000 may be a touch panel, a physical button, or a mouse.
The output device 2000 may be a display screen.
The memory 4000 may be a high-speed RAM memory or a non-volatile memory (e.g., a magnetic disk memory). The memory 4000 is used for storing a set of program codes, and the input device 1000, the output device 2000 and the processor 3000 are used for calling the program codes stored in the memory 4000 to execute the following operations:
the processor 3000 is configured to:
acquiring a target image;
preprocessing the target image to obtain a preprocessed target image;
carrying out segmentation and scaling processing on the preprocessed target image to obtain a plurality of local images with the same size;
respectively carrying out target detection on the plurality of local images to obtain a plurality of target detection frame images corresponding to a plurality of targets;
and classifying the multiple target detection frame images to obtain the categories of target objects contained in the multiple detection frame images.
In one possible example, in the aspect of preprocessing the target image to obtain a preprocessed target image, the processor 3000 is specifically configured to:
performing low-pass filtering processing on the target image to obtain a first partial image characteristic;
acquiring a second partial image feature except the first partial image feature in the target image;
enhancing the second partial image features to obtain enhanced second partial image features;
and recombining the first partial image characteristic and the enhanced second partial image characteristic to obtain an enhanced target image.
In one possible example, in the aspect of performing enhancement processing on the second partial image feature to obtain an enhanced second partial image feature, the processor 3000 is specifically configured to:
multiplying the image characteristic value of the second partial image characteristic by a gain value to obtain a gained image characteristic value;
and obtaining the enhanced second partial image characteristic based on the gained image characteristic value.
In a possible example, in the aspect of performing segmentation and scaling on the preprocessed target image to obtain a plurality of local images with the same size, the processor 3000 is specifically configured to:
dividing the preprocessed target image into a plurality of local images;
and amplifying the plurality of segmented local images to the same size by adopting a linear interpolation method to obtain a plurality of local images with the same size.
In one possible example, in the classifying the multiple target detection frame images to obtain the classes of the target objects included in the multiple detection frame images, the processor 3000 is specifically configured to:
for each target detection frame image, zooming and cutting the target detection frame image to obtain a processed image corresponding to the target detection frame image;
extracting target image features in the processed image corresponding to each target detection frame image; and inputting the target image characteristics into a classification model, and outputting the categories of the target objects contained in the target detection frame images to obtain the categories of the target objects contained in the plurality of detection frame images.
In one possible example, the processor 3000 is further specifically configured to:
the method comprises the steps of obtaining a training sample set and a verification sample set in advance, wherein the training sample set comprises training sample images to be trained, and the verification sample set comprises verification sample images to be verified;
randomly cutting the training sample images in the training sample set to obtain cut regional training images; carrying out zooming processing on the region image to obtain a region training image with a target size;
amplifying the verification sample images in the verification sample set to obtain amplified images; cutting the central area of the amplified image to obtain an area verification image with the same size as the target;
and carrying out classification training on a preset neural network model according to the region training image with the target size, and verifying the preset neural network model according to the region verification image with the target size to obtain the classification model.
In one possible example, in the aspect of performing classification training on a preset neural network model according to the region training image with the target size, the processor 3000 is specifically configured to:
extracting a region training feature sample in the region training image;
and training the preset neural network model according to the regional training feature samples and the corresponding result samples.
It can be seen that, with the electronic device described in the embodiments of the present application, by acquiring a target image; preprocessing a target image to obtain a preprocessed target image; carrying out segmentation and scaling processing on the preprocessed target image to obtain a plurality of local images with the same size; respectively carrying out target detection on the plurality of local images to obtain a plurality of target detection frame images corresponding to a plurality of targets; the method comprises the steps of classifying a plurality of target detection frame images to obtain the categories of target objects contained in the plurality of detection frame images, splitting complex scenes by dividing the target images in and out, improving the detection capability of small target objects and the generalization performance of the small target objects to different scenes, improving the application space of the target identification method, and based on the combination of target detection and classification, carrying out secondary result evaluation on detection results, reducing false reports and comprehensively improving the accuracy of target identification.
Please refer to fig. 8, which is a schematic structural diagram of an embodiment of a target identification apparatus according to an embodiment of the present application. The object recognition apparatus 500 described in the present embodiment includes: the acquiring unit 501, the processing unit 502, the detecting unit 503 and the classifying unit 504 are specifically as follows:
an acquisition unit 501 for acquiring a target image;
a processing unit 502, configured to perform preprocessing on the target image to obtain a preprocessed target image;
the processing unit 502 is further configured to perform segmentation and scaling on the preprocessed target image to obtain a plurality of local images with the same size;
a detecting unit 503, configured to perform target detection on the multiple local images respectively, so as to obtain multiple target detection frame images corresponding to multiple targets;
a classifying unit 504, configured to classify the multiple target detection frame images to obtain categories of target objects included in the multiple detection frame images.
Optionally, in the aspect of preprocessing the target image to obtain a preprocessed target image, the processing unit 502 is specifically configured to:
performing low-pass filtering processing on the target image to obtain a first partial image characteristic;
acquiring a second partial image feature except the first partial image feature in the target image;
enhancing the second partial image features to obtain enhanced second partial image features;
and recombining the first partial image characteristic and the enhanced second partial image characteristic to obtain an enhanced target image.
Optionally, in terms of performing enhancement processing on the second partial image features to obtain enhanced second partial image features, the processing unit 502 is specifically configured to:
multiplying the image characteristic value of the second partial image characteristic by a gain value to obtain a gained image characteristic value;
and obtaining the enhanced second partial image characteristic based on the gained image characteristic value.
Optionally, in respect that the preprocessed target image is segmented and scaled to obtain a plurality of local images with the same size, the processing unit 502 is specifically configured to:
dividing the preprocessed target image into a plurality of local images;
and amplifying the plurality of segmented local images to the same size by adopting a linear interpolation method to obtain a plurality of local images with the same size.
Optionally, in the aspect of classifying the multiple target detection frame images to obtain the classes of the target objects included in the multiple detection frame images, the classifying unit 504 is specifically configured to:
for each target detection frame image, zooming and cutting the target detection frame image to obtain a processed image corresponding to the target detection frame image;
extracting target image features in the processed image corresponding to each target detection frame image; and inputting the target image characteristics into a classification model, and outputting the categories of the target objects contained in the target detection frame images to obtain the categories of the target objects contained in the plurality of detection frame images.
Optionally, the apparatus further includes a training unit, where the training unit is specifically configured to:
the method comprises the steps of obtaining a training sample set and a verification sample set in advance, wherein the training sample set comprises training sample images to be trained, and the verification sample set comprises verification sample images to be verified;
randomly cutting the training sample images in the training sample set to obtain cut regional training images; carrying out zooming processing on the region image to obtain a region training image with a target size;
amplifying the verification sample images in the verification sample set to obtain amplified images; cutting the central area of the amplified image to obtain an area verification image with the same size as the target;
and carrying out classification training on a preset neural network model according to the region training image with the target size, and verifying the preset neural network model according to the region verification image with the target size to obtain the classification model.
Optionally, in the aspect of performing classification training on a preset neural network model according to the region training image of the target size, the training unit is specifically configured to:
extracting a region training feature sample in the region training image;
and training the preset neural network model according to the regional training feature samples and the corresponding result samples.
It can be seen that, with the target recognition apparatus described in the embodiments of the present application, by acquiring a target image; preprocessing a target image to obtain a preprocessed target image; segmenting and scaling the preprocessed target image to obtain a plurality of local images with the same size; respectively carrying out target detection on the plurality of local images to obtain a plurality of target detection frame images corresponding to a plurality of targets; the method comprises the steps of classifying a plurality of target detection frame images to obtain the categories of target objects contained in the plurality of detection frame images, splitting complex scenes by dividing the target images in and out, improving the detection capability of small target objects and the generalization performance of the small target objects to different scenes, improving the application space of the target identification method, and based on the combination of target detection and classification, carrying out secondary result evaluation on detection results, reducing false reports and comprehensively improving the accuracy of target identification.
It can be understood that the functions of each program module of the object recognition apparatus in this embodiment may be specifically implemented according to the method in the foregoing method embodiment, and the specific implementation process may refer to the related description of the foregoing method embodiment, which is not described herein again.
The present application further provides a computer storage medium, where the computer storage medium may store a program, and the program includes some or all of the steps of any one of the object recognition methods described in the above method embodiments when executed.
Embodiments of the present application provide a computer program product, wherein the computer program product comprises a non-transitory computer-readable storage medium storing a computer program operable to cause a computer to perform some or all of the steps described in any one of the object identification methods described in the embodiments of the present application. The computer program product may be a software installation package.
While the present application has been described in connection with various embodiments, other variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed application, from a review of the drawings, the disclosure, and the appended claims. In the claims, the word "comprising" does not exclude other elements or steps, and the word "a" or "an" does not exclude a plurality. A single processor or other unit may fulfill the functions of several items recited in the claims. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, apparatus (device), or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein. A computer program stored/distributed on a suitable medium supplied together with or as part of other hardware, may also take other distributed forms, such as via the Internet or other wired or wireless telecommunication systems.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (devices) and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a general purpose computer, special purpose computer, embedded processor, or other programmable processor to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable statistical population of devices, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable human vehicle trajectory analysis device to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable human vehicle trajectory analysis device to cause a series of operational steps to be performed on the computer or other programmable device to produce a computer implemented process such that the instructions which execute on the computer or other programmable device provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
Although the present application has been described in conjunction with specific features and embodiments thereof, it will be evident that various modifications and combinations can be made thereto without departing from the spirit and scope of the application. Accordingly, the specification and figures are merely exemplary of the present application as defined in the appended claims and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of the present application. It will be apparent to those skilled in the art that various changes and modifications may be made in the present application without departing from the spirit and scope of the application. Thus, if such modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is intended to include such modifications and variations as well.
Claims (10)
1. A method of object recognition, the method comprising:
acquiring a target image;
preprocessing the target image to obtain a preprocessed target image;
carrying out segmentation and scaling processing on the preprocessed target image to obtain a plurality of local images with the same size;
respectively carrying out target detection on the plurality of local images to obtain a plurality of target detection frame images corresponding to a plurality of targets;
and classifying the multiple target detection frame images to obtain the categories of target objects contained in the multiple detection frame images.
2. The method of claim 1, wherein the pre-processing the target image to obtain a pre-processed target image comprises:
performing low-pass filtering processing on the target image to obtain a first partial image characteristic;
acquiring a second partial image feature except the first partial image feature in the target image;
enhancing the second partial image features to obtain enhanced second partial image features;
and recombining the first partial image characteristic and the enhanced second partial image characteristic to obtain an enhanced target image.
3. The method according to claim 2, wherein the enhancing the second partial image feature to obtain an enhanced second partial image feature comprises:
multiplying the image characteristic value of the second partial image characteristic by a gain value to obtain a gained image characteristic value;
and obtaining the enhanced second partial image characteristic based on the gained image characteristic value.
4. The method according to claim 1, wherein the segmenting and scaling the preprocessed target image to obtain a plurality of local images with the same size comprises:
dividing the preprocessed target image into a plurality of local images;
and amplifying the plurality of segmented local images to the same size by adopting a linear interpolation method to obtain a plurality of local images with the same size.
5. The method according to claim 1, wherein the classifying the plurality of target detection frame images to obtain the categories of the target objects included in the plurality of detection frame images comprises:
for each target detection frame image, zooming and cutting the target detection frame image to obtain a processed image corresponding to the target detection frame image;
extracting target image features in the processed image corresponding to each target detection frame image; and inputting the target image characteristics into a classification model, and outputting the categories of the target objects contained in the target detection frame images to obtain the categories of the target objects contained in the plurality of detection frame images.
6. The method of claim 5, further comprising:
the method comprises the steps of obtaining a training sample set and a verification sample set in advance, wherein the training sample set comprises training sample images to be trained, and the verification sample set comprises verification sample images to be verified;
randomly cutting the training sample images in the training sample set to obtain cut regional training images; carrying out zooming processing on the region image to obtain a region training image with a target size;
amplifying the verification sample images in the verification sample set to obtain amplified images; cutting the central area of the amplified image to obtain an area verification image with the same size as the target;
and carrying out classification training on a preset neural network model according to the region training image with the target size, and verifying the preset neural network model according to the region verification image with the target size to obtain the classification model.
7. The method according to claim 6, wherein the classifying training of the preset neural network model according to the region training image of the target size comprises:
extracting a region training feature sample in the region training image;
and training the preset neural network model according to the regional training characteristic sample.
8. An object recognition apparatus, characterized in that the apparatus comprises:
an acquisition unit configured to acquire a target image;
the processing unit is used for preprocessing the target image to obtain a preprocessed target image;
the processing unit is further configured to perform segmentation and scaling on the preprocessed target image to obtain a plurality of local images with the same size;
the detection unit is used for respectively carrying out target detection on the plurality of local images to obtain a plurality of target detection frame images corresponding to a plurality of targets;
and the classification unit is used for classifying the target detection frame images to obtain the classes of the target objects contained in the detection frame images.
9. An electronic device comprising a processor, a memory for storing one or more programs and configured for execution by the processor, the programs comprising instructions for performing the steps of 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.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111648725.8A CN114445768A (en) | 2021-12-30 | 2021-12-30 | Target identification method and device, electronic equipment and storage medium |
| PCT/CN2022/100385 WO2023123924A1 (en) | 2021-12-30 | 2022-06-22 | Target recognition method and apparatus, and electronic device and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111648725.8A CN114445768A (en) | 2021-12-30 | 2021-12-30 | Target identification method and device, electronic equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114445768A true CN114445768A (en) | 2022-05-06 |
Family
ID=81365801
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111648725.8A Pending CN114445768A (en) | 2021-12-30 | 2021-12-30 | Target identification method and device, electronic equipment and storage medium |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN114445768A (en) |
| WO (1) | WO2023123924A1 (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114973204A (en) * | 2022-06-21 | 2022-08-30 | 国汽智控(北京)科技有限公司 | Target detection method, device, equipment, storage medium and product |
| WO2023123924A1 (en) * | 2021-12-30 | 2023-07-06 | 深圳云天励飞技术股份有限公司 | Target recognition method and apparatus, and electronic device and storage medium |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116909169B (en) * | 2023-09-14 | 2023-12-19 | 光轮智能(北京)科技有限公司 | Training method of operation control model, operation control method, equipment and medium |
| CN117440104B (en) * | 2023-12-21 | 2024-03-29 | 北京遥感设备研究所 | Data compression reconstruction method based on target significance characteristics |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110084284A (en) * | 2019-04-04 | 2019-08-02 | 苏州千视通视觉科技股份有限公司 | Target detection and secondary classification algorithm and device based on region convolutional neural networks |
| CN111598796A (en) * | 2020-04-27 | 2020-08-28 | Oppo广东移动通信有限公司 | Image processing method and device, electronic device and storage medium |
| CN111898523A (en) * | 2020-07-29 | 2020-11-06 | 电子科技大学 | A target detection method for special vehicles in remote sensing images based on transfer learning |
| CN112132002A (en) * | 2020-08-18 | 2020-12-25 | 欧必翼太赫兹科技(北京)有限公司 | Method and device for detecting foreign matter in three-dimensional image data |
| CN113256634A (en) * | 2021-07-13 | 2021-08-13 | 杭州医策科技有限公司 | Cervical carcinoma TCT slice vagina arranging method and system based on deep learning |
| CN113763302A (en) * | 2021-09-30 | 2021-12-07 | 青岛海尔科技有限公司 | Method and device for determining image detection result |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101751871B1 (en) * | 2015-11-20 | 2017-06-28 | 고려대학교 산학협력단 | Method and device for extracting feature of image data |
| CN109272459B (en) * | 2018-08-20 | 2020-12-01 | Oppo广东移动通信有限公司 | Image processing method, image processing device, storage medium and electronic equipment |
| CN110533023B (en) * | 2019-07-08 | 2021-08-03 | 天津商业大学 | A method and device for detecting and identifying foreign bodies in railway freight cars |
| CN113139896A (en) * | 2020-01-17 | 2021-07-20 | 波音公司 | Target detection system and method based on super-resolution reconstruction |
| CN114445768A (en) * | 2021-12-30 | 2022-05-06 | 深圳云天励飞技术股份有限公司 | Target identification method and device, electronic equipment and storage medium |
-
2021
- 2021-12-30 CN CN202111648725.8A patent/CN114445768A/en active Pending
-
2022
- 2022-06-22 WO PCT/CN2022/100385 patent/WO2023123924A1/en active Application Filing
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110084284A (en) * | 2019-04-04 | 2019-08-02 | 苏州千视通视觉科技股份有限公司 | Target detection and secondary classification algorithm and device based on region convolutional neural networks |
| CN111598796A (en) * | 2020-04-27 | 2020-08-28 | Oppo广东移动通信有限公司 | Image processing method and device, electronic device and storage medium |
| CN111898523A (en) * | 2020-07-29 | 2020-11-06 | 电子科技大学 | A target detection method for special vehicles in remote sensing images based on transfer learning |
| CN112132002A (en) * | 2020-08-18 | 2020-12-25 | 欧必翼太赫兹科技(北京)有限公司 | Method and device for detecting foreign matter in three-dimensional image data |
| CN113256634A (en) * | 2021-07-13 | 2021-08-13 | 杭州医策科技有限公司 | Cervical carcinoma TCT slice vagina arranging method and system based on deep learning |
| CN113763302A (en) * | 2021-09-30 | 2021-12-07 | 青岛海尔科技有限公司 | Method and device for determining image detection result |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2023123924A1 (en) * | 2021-12-30 | 2023-07-06 | 深圳云天励飞技术股份有限公司 | Target recognition method and apparatus, and electronic device and storage medium |
| CN114973204A (en) * | 2022-06-21 | 2022-08-30 | 国汽智控(北京)科技有限公司 | Target detection method, device, equipment, storage medium and product |
| CN114973204B (en) * | 2022-06-21 | 2024-04-05 | 国汽智控(北京)科技有限公司 | Target detection method, device, equipment, storage medium and product |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2023123924A1 (en) | 2023-07-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108875676B (en) | Living body detection method, device and system | |
| CN114445768A (en) | Target identification method and device, electronic equipment and storage medium | |
| CN109766779B (en) | Loitering person identification method and related product | |
| JP6240199B2 (en) | Method and apparatus for identifying object in image | |
| US9008365B2 (en) | Systems and methods for pedestrian detection in images | |
| CN109727275B (en) | Object detection method, device, system and computer readable storage medium | |
| CN112492343A (en) | Video live broadcast monitoring method and related device | |
| CN112733666A (en) | Method, equipment and storage medium for collecting difficult images and training models | |
| CN111563398A (en) | Method and device for determining information of target object | |
| CN113920585A (en) | Behavior recognition method and device, equipment and storage medium | |
| CN110866428B (en) | Target tracking method, device, electronic equipment and storage medium | |
| CN113469099A (en) | Training method, detection method, device, equipment and medium of target detection model | |
| CN112052702A (en) | Method and device for identifying two-dimensional code | |
| CN115457466A (en) | Inspection video-based hidden danger detection method and system and electronic equipment | |
| CN113947744A (en) | Fire image detection method, system, equipment and storage medium based on video | |
| US20200394802A1 (en) | Real-time object detection method for multiple camera images using frame segmentation and intelligent detection pool | |
| CN112907206A (en) | Service auditing method, device and equipment based on video object identification | |
| CN115984780A (en) | Industrial solid waste warehouse-in and warehouse-out distinguishing method and device, electronic equipment and medium | |
| CN113095272B (en) | Living body detection method, living body detection device, living body detection medium and computer program product | |
| CN114821978A (en) | Method, device and medium for eliminating false alarm | |
| CN110659384B (en) | Video structured analysis method and device | |
| CN110751065B (en) | Training data acquisition method and device | |
| CN114694209A (en) | Video processing method and device, electronic equipment and computer storage medium | |
| CN114550049A (en) | Behavior recognition method, device, equipment and storage medium | |
| CN108875770B (en) | Pedestrian detection false alarm data labeling method, device, system and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |