CN111047624A - Image dim target detection method, device, equipment and storage medium - Google Patents
Image dim target detection method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN111047624A CN111047624A CN201911382402.1A CN201911382402A CN111047624A CN 111047624 A CN111047624 A CN 111047624A CN 201911382402 A CN201911382402 A CN 201911382402A CN 111047624 A CN111047624 A CN 111047624A
- Authority
- CN
- China
- Prior art keywords
- gray
- image
- value
- target
- pixel point
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 49
- 238000012545 processing Methods 0.000 claims abstract description 84
- 238000000034 method Methods 0.000 claims abstract description 50
- 238000001914 filtration Methods 0.000 claims abstract description 48
- 230000000877 morphologic effect Effects 0.000 claims abstract description 43
- 238000010606 normalization Methods 0.000 claims description 19
- 238000004590 computer program Methods 0.000 claims description 11
- 238000004458 analytical method Methods 0.000 claims description 8
- 238000005516 engineering process Methods 0.000 abstract description 7
- 238000007621 cluster analysis Methods 0.000 abstract description 5
- 238000011161 development Methods 0.000 abstract description 4
- 230000009286 beneficial effect Effects 0.000 abstract description 3
- 238000005286 illumination Methods 0.000 description 6
- 230000002159 abnormal effect Effects 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 3
- 238000003384 imaging method Methods 0.000 description 3
- 230000006641 stabilisation Effects 0.000 description 3
- 238000011105 stabilization Methods 0.000 description 3
- 238000009825 accumulation Methods 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 230000002349 favourable effect Effects 0.000 description 2
- 238000009432 framing Methods 0.000 description 2
- 238000003672 processing method Methods 0.000 description 2
- 230000000087 stabilizing effect Effects 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000003331 infrared imaging Methods 0.000 description 1
- 230000002401 inhibitory effect Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000009827 uniform distribution Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/254—Analysis of motion involving subtraction of images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration using local operators
- G06T5/30—Erosion or dilatation, e.g. thinning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a method for detecting a small and weak target of an image, which comprises the steps of obtaining a gray image of a current frame; performing top-hat morphological filtering processing according to the gray value of each pixel point of the gray image; performing background modeling according to the gray value of each pixel point of the gray image subjected to top hat morphological filtering, and identifying foreground points and background points in each pixel point of the divided gray image; and carrying out cluster analysis processing on each background point, and dividing the foreground points belonging to the same class into target frames of the same foreground moving target. According to the method, the background noise in the gray level image is suppressed through top-hat morphological filtering, the influence of complex background noise is solved, then background modeling is used for extracting target pixel points, the false alarm rate is reduced, the detection rate of weak and small targets is improved under the condition of low signal-to-noise ratio, and the development and application of an image tracking technology are facilitated. The application also provides a dim target detection device, equipment and a computer readable storage medium, which have the beneficial effects.
Description
Technical Field
The present invention relates to the field of image processing technologies, and in particular, to a method, an apparatus, a device, and a computer-readable storage medium for detecting a small and weak object in an image.
Background
The infrared weak and small target detection is widely applied to the fields of accurate striking, clearance security and the like of a weapon system and is a key technology for infrared guidance and target tracking. Because infrared possesses passive formation of image and characteristics such as imaging distance is far away, the target imaging pixel ratio is less in the long-distance reconnaissance monitoring, and the contrast is low, lacks characteristics such as texture profile, leads to its detection degree of difficulty to increase. In addition, secondary noise caused by long-distance transmission air turbulence and the like is superposed on output noise of the infrared imaging array, so that the signal-to-noise ratio of a weak target is low. Under the imaging scenes, the method has important application value for improving the detection rate of infrared dim targets and reducing the false alarm rate caused by noise.
However, in the current technology for processing the target image, the target in the image is mostly detected and identified by an inter-frame difference method, a maximum and minimum filtering method, a background statistical method, a background modeling method, and the like. The existing method has high detection rate in advance of strong image target signal-to-noise ratio. But it is difficult to achieve better detection effect in the detection of low signal-to-noise ratio and low contrast target.
Disclosure of Invention
The invention aims to provide a method, a device and equipment for detecting weak and small targets in an image and a computer readable storage medium, which improve the detection effect of the weak and small targets in the image.
In order to solve the above technical problem, the present invention provides a method for detecting a small and weak image target, comprising:
obtaining a gray level image of a current frame;
carrying out top-hat morphological filtering processing on the gray level image;
performing background modeling according to the gray value of each pixel point of the gray image subjected to top hat morphological filtering, and detecting foreground points in each pixel point of the gray image;
and performing clustering analysis processing on each foreground point, and dividing the foreground points belonging to the same class into target frames of the same foreground moving target.
In an optional embodiment of the present application, after obtaining the grayscale image of the current frame, before performing top-hat morphological filtering processing on the grayscale image, the method further includes:
acquiring gray maximum values and gray minimum values corresponding to the historical frame gray images of the current frame in a preset number;
and performing linear normalization processing on the gray value of each pixel point of the gray image by taking the maximum gray value and the minimum gray value as boundaries.
In an optional embodiment of the present application, after performing linear normalization processing on the gray scale values of the respective pixel points of the gray scale image, the method further includes:
performing platform processing on a gray value histogram corresponding to the gray value of each pixel point of the gray image to obtain a platform histogram of the gray image;
stretching the gray level image based on the platform histogram; and then according to the gray value of each pixel point of the gray image subjected to stretching processing, performing top-hat morphological filtering processing on the gray image.
In an optional embodiment of the present application, after obtaining the grayscale image of the current frame, the method further includes:
determining a maximum gray boundary value and a minimum gray boundary value of a gray value according to a preset pixel gray value proportion;
changing the gray value larger than the maximum gray boundary value into the maximum gray boundary value and changing the gray value smaller than the minimum gray boundary value into the minimum gray boundary value in each pixel point of the gray image; and then, the operation of obtaining the maximum gray value and the minimum gray value corresponding to the preset number of historical frame gray images of the gray image of the current frame is executed.
In an optional embodiment of the present application, after dividing foreground points belonging to the same class into target frames of the same foreground moving target, the method further includes:
and eliminating the target frames which do not meet the prior condition in each target frame.
The application also provides a device for detecting the weak and small target of the image, which comprises:
the image acquisition module is used for acquiring a gray level image of the current frame;
the filtering processing module is used for carrying out top-hat morphological filtering processing on the gray level image;
the background modeling module is used for carrying out background modeling according to the gray value of each pixel point of the gray image subjected to top hat morphological filtering and detecting foreground points in each pixel point of the gray image;
and the target classification module is used for carrying out clustering analysis processing on each foreground point and dividing the foreground points belonging to the same class into target frames of the same foreground moving target.
In an optional embodiment of the present application, the image processing apparatus further includes a normalization processing module, configured to, after obtaining a grayscale image of a current frame, obtain a maximum grayscale value and a minimum grayscale value corresponding to a preset number of historical frame grayscale images of the grayscale image of the current frame before performing top-hat morphological filtering processing on the grayscale image; and performing linear normalization processing on the gray value of each pixel point of the gray image by taking the maximum gray value and the minimum gray value as boundaries.
In an optional embodiment of the present application, the system further includes a histogram processing module, configured to perform a linear normalization process on the gray scale value of each pixel of the gray scale image, and then perform a platformization process on the gray scale value histogram corresponding to the gray scale value of each pixel of the gray scale image to obtain a platform histogram of the gray scale image; stretching the gray level image based on the platform histogram; and then according to the gray value of each pixel point of the gray image subjected to stretching processing, performing top-hat morphological filtering processing on the gray image.
The application also provides a weak little target detection equipment of image, includes:
a memory for storing a computer program;
a processor for executing the computer program to implement the operations of the steps of the image weak small object detection method as described in any one of the above.
The present application further provides a computer-readable storage medium, which stores a computer program, and the computer program, when executed by a processor, implements the steps of the image weak small object detection method according to any one of the above.
The invention provides a method for detecting a small and weak target of an image, which comprises the steps of obtaining a gray image of a current frame; performing top-hat morphological filtering processing according to the gray value of each pixel point of the gray image; performing background modeling according to the gray value of each pixel point of the gray image subjected to top hat morphological filtering, and identifying foreground points and background points in each pixel point of the gray image; and carrying out cluster analysis processing on each background point, and dividing the foreground points belonging to the same class into target frames of the same foreground moving target.
When weak and small targets in an image are detected and identified, background noise in a gray image is firstly suppressed through top hat morphological filtering, a significant foreground image is extracted, the influence of complex background noise is solved, target pixel point extraction is carried out on the significant foreground image in a time domain sequence image through background modeling, false targets detected on a single space foreground image can be eliminated, the detection rate is improved, the false alarm rate is reduced, and the detection efficiency can meet the real-time requirement. The method and the device can improve the detection rate of the weak and small targets under the condition of low signal-to-noise ratio, and are favorable for development and application of an image tracking technology.
The application also provides a dim target detection device, equipment and a computer readable storage medium, which have the beneficial effects.
Drawings
In order to more clearly illustrate the embodiments or technical solutions of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained based on these drawings without creative efforts.
Fig. 1 is a schematic flowchart of a method for detecting a weak and small image target according to an embodiment of the present application;
fig. 2 is a schematic flowchart of a method for detecting a weak and small image target according to another embodiment of the present application;
fig. 3 is a block diagram of a structure of an image weak and small object detection apparatus according to an embodiment of the present invention.
Detailed Description
In order that those skilled in the art will better understand the disclosure, the invention will be described in further detail with reference to the accompanying drawings and specific embodiments. It is to be understood that the described embodiments are merely exemplary of the invention, and not restrictive of the full scope of the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
As shown in fig. 1, fig. 1 is a schematic flowchart of a method for detecting a weak and small image target according to an embodiment of the present application, where the method may include:
step S11: and obtaining a gray level image of the current frame.
For an infrared image shot by an infrared camera, the infrared image is also a gray image, and for an ordinary camera, a gray image is obtained by performing gray processing on a current frame image shot and imaged.
Step S12: and carrying out top-hat morphological filtering processing on the gray level image.
Specifically, the top-hat morphological filtering is a commonly used image processing method, can effectively suppress the background, highlights areas brighter than areas around the original image outline, and is beneficial to the accuracy of subsequent background extraction.
It should be noted that the pixel ratio of the weak target in the image is very low, and the gray contrast between the target pixel and the background pixel is very weak. After the top cap morphological filtering, the significance of the target pixel point can be obviously improved, an obvious significant foreground image of the target pixel point is obtained, and the subsequent identification and detection of the target pixel point are facilitated.
Step S13: and performing background modeling according to the gray value of each pixel point of the gray image subjected to top hat morphological filtering, and detecting foreground points in each pixel point of the gray image.
Specifically, the rapid background modeling on the time domain sequence is adopted, and the rapid background modeling can be realized through modes such as mixed Gaussian background modeling or VIBE background modeling. On the basis of the background model, the gray value of each pixel point is judged according to a set gray threshold, background points which are smaller than the preset gray threshold belong to, and foreground points which are larger than the preset gray threshold belong to. The foreground point is also a pixel point that may belong to the target object.
Step S14: and performing clustering analysis processing on each foreground point, and dividing the foreground points belonging to the same class into target frames of the same foreground moving target.
On the basis of the obtained foreground points, clustering can be carried out on the discrete foreground points by adopting a mean shift clustering method, but not limited to the mean shift clustering method, so that class labels can be given to the discrete foreground points, and the discrete points meeting the requirement that the same class label can physically correspond to the same foreground moving target.
And acquiring the pixel coordinates of the foreground points of the same class number, namely, the upper, lower, left and right pixel coordinates of the class, so as to establish a surrounding frame corresponding to the class, wherein the surrounding frame can be regarded as an area of a target, the outside of the frame corresponds to a background area, then, the surrounding frames are combined, the crossed surrounding frames are combined, and the target frame is output.
When the method is used for detecting the weak and small target in the current frame image, the background noise in the gray level image of the current frame is firstly inhibited through top-hat morphological filtering, the significance of the weak and small target is improved to a certain extent, a significant foreground image is obtained, and the influence of complex background noise is solved; and then, a background modeling mode is used for extracting target pixel points of the obvious foreground image in a time domain sequence image, so that a false target detected on the single space foreground image can be eliminated, the background points and foreground points are distinguished for each pixel point, the influence of high image signal to noise ratio on the distinguishing accuracy of the background points and the foreground points is effectively avoided, the identification precision and the detection rate of the foreground points in the image are improved, the false alarm rate is reduced, and the detection efficiency can meet the real-time requirement. After foreground points are detected, classification and framing are carried out on the foreground points belonging to the same target through a clustering analysis algorithm, and therefore identification and delineation of small and weak targets are achieved. The method and the device can improve the detection rate of the weak and small targets under the condition of low signal-to-noise ratio, and are favorable for development and application of an image tracking technology.
Based on the above embodiment, in the above embodiment, the top-hat morphological filtering process is performed on the grayscale image before the background modeling is performed, so that the influence of the background noise is reduced to some extent. The application also provides an optional embodiment, before the top-hat morphological filtering processing is carried out, the preprocessing process is carried out on the gray-scale image, and the accuracy of subsequent target identification and detection can be effectively improved. Specifically, after the grayscale image of the current frame is obtained in the step S11, and before the step S12 of performing the top-hat morphological filtering process, the grayscale image may be processed as follows:
determining a maximum gray boundary value and a minimum gray boundary value of a gray value according to a preset pixel gray value proportion;
and changing the gray value which is greater than the maximum gray boundary value into the maximum gray boundary value and changing the gray value which is less than the minimum gray boundary value into the minimum gray boundary value in each pixel point of the gray image.
Specifically, the number of pixel points included in each gray value is counted, and then a preset pixel gray value ratio excluding the maximum value is selected, for example, the preset pixel gray value ratio may be 2%, if the gray value of 98% of the pixel points in the image is 10-240, the minimum gray boundary value of the image is 10, and the maximum gray boundary value is 240; then changing the gray value of the pixel point with the gray value smaller than 10 in the image into 10, and changing the gray value of the pixel point with the gray value larger than 240 into 240, thereby eliminating the noise interference of the abnormal pixel point with the gray value being too high or too low to the target identification. For the ratio of the gray value of the preset pixel excluding the maximum value, the ratio can be set empirically or adjusted according to actual needs, and there is no inevitable requirement in the present application.
In this embodiment, before the step S12, the gray value of each pixel point of the gray image is subjected to the most significant removal processing to eliminate the interference of the abnormal point, so as to reduce the influence fluctuation of the new target on the background and improve the accuracy of identifying foreground points and background points by the background modeling.
Of course, in addition to processing the grayscale image in a manner of removing the most value from the grayscale image, the present application also provides another alternative embodiment, and after the grayscale image of the current frame is obtained in the step S11, before the step S12 of performing the top-hat morphological filtering processing, the light stabilization processing may be performed on the grayscale image, and the specific process is as follows:
acquiring gray maximum values and gray minimum values corresponding to the historical frame gray images of the preset number in front of the gray image of the current frame;
and performing linear normalization processing on the gray value of each pixel point of the gray image by taking the maximum gray value and the minimum gray value as boundaries.
On the premise of extracting the gray value maximum value of a preset number of historical frames in the current frame, counting the range of the maximum value of the historical frames, respectively obtaining the gray minimum value of the historical frames and then the median value of the gray maximum value, taking the gray minimum value and the gray maximum value as the background gray normalization boundary of the current frame, playing a role of being consistent with the historical background gray, and physically realizing illumination stabilization processing.
Specifically, for a pixel point of a gray image, assuming that the gray value is a, after normalization processing, the gray value of the pixel point is changed into
Wherein M is the maximum gray value of the historical frame, and N is the minimum gray value of the historical frame.
Because the weak and small targets are easily submerged in the fluctuation of light, after normalization processing is carried out, illumination stabilization processing is physically realized, a more stable and accurate background model can be obtained in subsequent background modeling of the stabilized gray level image, foreground points can be detected more accurately, and false alarms are fewer.
It should be noted that, in the above embodiment, before the top-hat morphological filtering, both the removing of the maxima and the stabilizing illumination processing are performed on the grayscale image to improve the accuracy of the subsequent foreground point detection and identification, in practical application, only one of the processing may be performed on the grayscale image, or the removing of the maxima may be performed first and then the stabilizing illumination processing is performed, so as to improve the accuracy of the foreground point identification to the maximum extent and reduce the interference of the background noise.
In addition to the above two processing methods for the grayscale image, in another embodiment of the present application, after the grayscale image of the current frame is obtained in the step S11, the following processing may be performed on the grayscale image before the step S12 of performing the top-hat morphological filtering processing:
performing platform processing on a gray value histogram corresponding to the gray value of each pixel point of the gray image to obtain a platform histogram of the gray image;
stretching the gray level image based on the platform histogram; and then according to the gray value of each pixel point of the gray image subjected to stretching processing.
It should be noted that the grayscale histogram in this embodiment is a histogram drawn with grayscale values as abscissa and the number of pixels corresponding to each grayscale value in the grayscale image as ordinate. And (3) performing platform histogram conversion on the gray level histogram, specifically setting a certain platform threshold value or a self-adaptive given platform threshold value, then truncating the histogram exceeding the threshold value, and averaging the truncated histogram to raise the histogram reference so as to complete the platform histogram. In particular, reference may be made to the formula:
After the histogram platform conversion is realized, the histogram normalization of the platform is accumulated to make a gray mapping curve, the stretching processing of the platform histogram is completed,performing accumulation calculation on the platform histogram to obtain an accumulation function:
wherein k is 0,1, 2. Then, the gray value of the original image is changed into a new gray value through equalization:
in this embodiment, the histogram of the grayscale image is subjected to the flat-to-flat conversion, and the flat-to-flat histogram is stretched, which is a processing manner of the equalization of the histogram, and is to change the grayscale histogram of the original image from a certain grayscale interval in a relatively concentrated manner to a uniform distribution in the entire grayscale range, so that the number of pixels in a certain grayscale range is substantially the same.
Based on the foregoing embodiment, in another specific optional embodiment of the present application, as shown in fig. 2, fig. 2 is a schematic flow chart of a method for detecting a weak and small image target according to another specific embodiment of the present application, where the method may include:
step S201: and obtaining a gray level image of the current frame.
Step S202: determining a maximum gray boundary value and a minimum gray boundary value of a gray value according to a preset pixel gray value proportion;
step S203: and changing the gray value which is greater than the maximum gray boundary value into the maximum gray boundary value and changing the gray value which is less than the minimum gray boundary value into the minimum gray boundary value in each pixel point of the gray image.
Step S204: and acquiring the gray maximum value and the gray minimum value corresponding to the historical frame gray images of the preset number in front of the gray image of the current frame.
Step S205: and performing linear normalization processing on the gray value of each pixel point of the gray image by taking the maximum gray value and the minimum gray value as boundaries.
Step S206: performing platform processing on a gray value histogram corresponding to the gray value of each pixel point of the gray image to obtain a platform histogram of the gray image;
step S207: and stretching the gray level image based on the platform histogram.
Step S208: and performing top-hat morphological filtering processing on the stretched gray level image.
Step S209: and performing background modeling according to the gray value of each pixel point of the gray image subjected to top hat morphological filtering, and detecting foreground points in each pixel point of the gray image.
Step S210: and performing clustering analysis processing on each foreground point, and dividing the foreground points belonging to the same class into target frames of the same foreground moving target.
Step S211: and eliminating the target frames which do not meet the prior condition in each target frame.
Specifically, the prior condition is determined according to the detection scene and the characteristics of the target, such as a human target, the aspect ratio of the corresponding target frame does not exceed a certain value, and too long or too wide can not be the target that we need to detect, so that the corresponding class can be removed and not be used as an output. The prior conditions can be selected as the length-width ratio, the density ratio, the pixel number and the like of a class coordinate range according to an actual scene and a target, the extension contrast of the target and the like, can be used as the prior conditions of related detection, and other prior conditions can be expanded to meet the purpose of eliminating false alarms of the corresponding detection scene.
In the embodiment, when identifying and detecting a weak target in an image, the maximum gray value and the minimum gray value in the gray image of the current frame are removed, and noise interference on target identification caused by an abnormal pixel point with an excessively high or excessively low gray value is eliminated; after the gray scale minimum value is removed, the gray scale value of the gray scale image is normalized, so that the weak and small targets are prevented from being submerged in the illumination fluctuation, and the illumination stability processing of the gray scale image is realized; performing platform processing on the gray level image, and stretching the obtained platform histogram to realize gray level equalization of the target area and the background and improve the signal-to-noise ratio of the target; and further performing top hat morphological filtering, background modeling and cluster analysis on the gray level image to finally obtain a target frame framing the weak and small target, and rejecting the target frame which does not meet the prior condition according to the characteristics of the target to be detected, thereby further ensuring the accuracy of target detection and identification.
The following describes how the image weak and small object detection apparatus and the image weak and small object detection method described above may be referred to in correspondence.
Fig. 3 is a block diagram of an image weak and small object detection apparatus according to an embodiment of the present invention, and with reference to fig. 3, the image weak and small object detection apparatus may include:
an image acquisition module 100, configured to obtain a grayscale image of a current frame;
a filtering processing module 200, configured to perform top-hat morphological filtering processing on the grayscale image;
the background modeling module 300 is configured to perform background modeling according to the gray value of each pixel point of the gray image subjected to top hat morphological filtering, and detect a foreground point in each pixel point of the gray image;
and the target classification module 400 is configured to perform cluster analysis on each foreground point, and classify foreground points belonging to the same class into target frames of the same foreground moving target.
Optionally, in another specific embodiment of the present application, the method may further include:
the normalization processing module is used for acquiring the gray maximum value and the gray minimum value corresponding to the historical frame gray images of the current frame in the preset number before the gray images of the current frame are obtained and before the top-hat morphological filtering processing is carried out on the gray images; and performing linear normalization processing on the gray value of each pixel point of the gray image by taking the maximum gray value and the minimum gray value as boundaries.
Optionally, in another specific embodiment of the present application, the method may further include:
the histogram processing module is used for performing linear normalization processing on the gray value of each pixel point of the gray image, and then performing platform processing on the gray value histogram corresponding to the gray value of each pixel point of the gray image to obtain a platform histogram of the gray image; stretching the gray level image based on the platform histogram; and then according to the gray value of each pixel point of the gray image subjected to stretching processing, performing top-hat morphological filtering processing on the gray image.
Optionally, in another specific embodiment of the present application, the method may further include:
the gray scale minimum value removing module is used for determining the maximum gray scale boundary value and the minimum gray scale boundary value of the gray scale value according to the preset pixel gray scale value proportion after the gray scale image of the current frame is obtained; changing the gray value larger than the maximum gray boundary value into the maximum gray boundary value and changing the gray value smaller than the minimum gray boundary value into the minimum gray boundary value in each pixel point of the gray image; and then, the operation of obtaining the maximum gray value and the minimum gray value corresponding to the preset number of historical frame gray images of the gray image of the current frame is executed.
Optionally, in another specific embodiment of the present application, the method may further include:
and the elimination module is used for eliminating the target frames which do not meet the prior condition in each target frame after dividing the foreground points belonging to the same class into the target frames of the same foreground moving target.
The image weak and small target detection apparatus of this embodiment is used to implement the foregoing image weak and small target detection method, and therefore specific implementations of the image weak and small target detection apparatus can be seen in the foregoing embodiments of the image weak and small target detection method, for example, the image acquisition module 100, the filtering processing module 200, the background modeling module 300, and the target classification module 400 are respectively used to implement steps S11, S12, S13, and S14 in the foregoing image weak and small target detection method, so that specific implementations of the apparatus may refer to descriptions of corresponding partial embodiments, and are not repeated herein.
The application also provides a weak little target detection equipment of image, includes:
a memory for storing a computer program;
a processor for executing the computer program to implement the operations of the steps of the image weak small object detection method as described in any one of the above.
The image weak and small target detection method specifically comprises the following steps:
obtaining a gray level image of a current frame;
carrying out top-hat morphological filtering processing on the gray level image;
performing background modeling according to the gray value of each pixel point of the gray image subjected to top hat morphological filtering, and detecting foreground points in each pixel point of the gray image;
and performing clustering analysis processing on each foreground point, and dividing the foreground points belonging to the same class into target frames of the same foreground moving target.
According to the method, the purpose of inhibiting the background is achieved by performing top-hat morphological filtering processing on the gray level image of the current frame, then background modeling is performed based on the gray level value of the gray level image, distinguishing of foreground points and background points is achieved, and a target frame is defined through cluster analysis, so that small and weak targets are accurately and effectively identified, accurate and quick detection and identification of the small and weak targets are achieved, and development and application of an image detection technology are facilitated.
The present application further provides a computer-readable storage medium, which stores a computer program, and the computer program, when executed by a processor, implements the steps of the image weak small object detection method according to any one of the above.
The computer readable storage medium may be Random Access Memory (RAM), memory, Read Only Memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.
The embodiments are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same or similar parts among the embodiments are referred to each other. The device disclosed by the embodiment corresponds to the method disclosed by the embodiment, so that the description is simple, and the relevant points can be referred to the method part for description.
Those of skill would further appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both, and that the various illustrative components and steps have been described above generally in terms of their functionality in order to clearly illustrate this interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. 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 invention.
Claims (10)
1. A method for detecting a weak and small image target is characterized by comprising the following steps:
obtaining a gray level image of a current frame;
carrying out top-hat morphological filtering processing on the gray level image;
performing background modeling according to the gray value of each pixel point of the gray image after the top-hat morphological filtering processing, and detecting foreground points in each pixel point of the gray image;
and performing clustering analysis processing on each foreground point, and dividing the foreground points belonging to the same class into target frames of the same foreground moving target.
2. The method for detecting the weak and small image target as claimed in claim 1, wherein after obtaining the gray image of the current frame, before performing the top-hat morphological filtering process on the gray image, the method further comprises:
acquiring gray maximum values and gray minimum values corresponding to the historical frame gray images of the current frame in a preset number;
and performing linear normalization processing on the gray value of each pixel point of the gray image by taking the maximum gray value and the minimum gray value as boundaries.
3. The method for detecting the small and weak image target as claimed in claim 2, wherein after the linear normalization processing is performed on the gray values of the pixels of the gray image, the method further comprises:
performing platform processing on a gray value histogram corresponding to the gray value of each pixel point of the gray image to obtain a platform histogram of the gray image;
stretching the gray level image based on the platform histogram; and then according to the gray value of each pixel point of the gray image after stretching processing, executing the step of top-hat morphological filtering processing on the gray image.
4. The method for detecting a small object in an image as claimed in claim 2, further comprising, after obtaining the gray image of the current frame:
determining a maximum gray boundary value and a minimum gray boundary value of a gray value according to a preset pixel gray value proportion;
changing the gray value larger than the maximum gray boundary value into the maximum gray boundary value and changing the gray value smaller than the minimum gray boundary value into the minimum gray boundary value in each pixel point of the gray image; and then, the operation of obtaining the maximum gray value and the minimum gray value corresponding to the preset number of historical frame gray images of the gray image of the current frame is executed.
5. The method for detecting weak and small image targets as claimed in any one of claims 1 to 4, wherein after dividing foreground points belonging to the same class into target frames of the same foreground moving target, the method further comprises:
and eliminating the target frames which do not meet the prior condition in each target frame.
6. An image weak small object detection device, comprising:
the image acquisition module is used for acquiring a gray level image of the current frame;
the filtering processing module is used for carrying out top-hat morphological filtering processing on the gray level image;
the background modeling module is used for carrying out background modeling according to the gray value of each pixel point of the gray image after the top cap morphological filtering processing, and detecting foreground points in each pixel point of the gray image;
and the target classification module is used for carrying out clustering analysis processing on each foreground point and dividing the foreground points belonging to the same class into target frames of the same foreground moving target.
7. The apparatus according to claim 6, further comprising a normalization processing module, configured to obtain a maximum gray value and a minimum gray value corresponding to a preset number of historical frame gray images of the gray image of the current frame after obtaining the gray image of the current frame and before performing top-hat morphological filtering processing on the gray image; and performing linear normalization processing on the gray value of each pixel point of the gray image by taking the maximum gray value and the minimum gray value as boundaries.
8. The image dim-small target detection device according to claim 6, further comprising a histogram processing module, configured to perform a flattening process on the gray value histogram corresponding to the gray value of each pixel of the gray image after performing a linear normalization process on the gray value of each pixel of the gray image, so as to obtain a flattened histogram of the gray image; stretching the gray level image based on the platform histogram; and then according to the gray value of each pixel point of the gray image after stretching processing, executing the step of top-hat morphological filtering processing on the gray image.
9. An image weak small object detection apparatus, comprising:
a memory for storing a computer program;
a processor for executing the computer program to implement the operations of the steps of the image weak small object detection method according to any one of claims 1 to 5.
10. A computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program which, when executed by a processor, implements the steps of the image weak small object detection method according to any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911382402.1A CN111047624A (en) | 2019-12-27 | 2019-12-27 | Image dim target detection method, device, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911382402.1A CN111047624A (en) | 2019-12-27 | 2019-12-27 | Image dim target detection method, device, equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111047624A true CN111047624A (en) | 2020-04-21 |
Family
ID=70240853
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911382402.1A Pending CN111047624A (en) | 2019-12-27 | 2019-12-27 | Image dim target detection method, device, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111047624A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112733589A (en) * | 2020-10-29 | 2021-04-30 | 广西科技大学 | Infrared image pedestrian detection method based on deep learning |
| CN113554632A (en) * | 2021-07-30 | 2021-10-26 | 沭阳天勤工具有限公司 | Artificial intelligence-based workpiece detection optimization method in turbid water body |
| WO2021217445A1 (en) * | 2020-04-28 | 2021-11-04 | 深圳市大疆创新科技有限公司 | Image processing method, device and system, and storage medium |
| CN114092777A (en) * | 2021-11-29 | 2022-02-25 | 广州辰创科技发展有限公司 | Efficient identification method and system based on small target |
| CN114332756A (en) * | 2021-12-15 | 2022-04-12 | 中国飞行试验研究院 | Target identification method realized by fast-Vibe and classification algorithm |
| CN117934417A (en) * | 2024-01-24 | 2024-04-26 | 北京首发公路养护工程有限公司 | Method, device, equipment and medium for identifying apparent defects of road based on neural network |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101408981A (en) * | 2008-11-25 | 2009-04-15 | 中国电子科技集团公司第二十八研究所 | Method for detecting infrared puniness target |
| WO2011022279A2 (en) * | 2009-08-18 | 2011-02-24 | Behavioral Recognition Systems, Inc. | Background model for complex and dynamic scenes |
| CN102184552A (en) * | 2011-05-11 | 2011-09-14 | 上海理工大学 | Moving target detecting method based on differential fusion and image edge information |
| CN103473549A (en) * | 2013-09-23 | 2013-12-25 | 北京智诺英特科技有限公司 | Image target detecting method and device |
| CN105469428A (en) * | 2015-11-26 | 2016-04-06 | 河海大学 | Morphological filtering and SVD (singular value decomposition)-based weak target detection method |
| CN106327526A (en) * | 2016-08-22 | 2017-01-11 | 湖南挚新科技发展有限公司 | Image object tracking method and image object tracking system |
| CN109033972A (en) * | 2018-06-27 | 2018-12-18 | 上海数迹智能科技有限公司 | A kind of object detection method, device, equipment and storage medium |
| CN109345472A (en) * | 2018-09-11 | 2019-02-15 | 重庆大学 | A kind of infrared moving small target detection method of complex scene |
| CN110415208A (en) * | 2019-06-10 | 2019-11-05 | 西安电子科技大学 | A kind of adaptive targets detection method and its device, equipment, storage medium |
-
2019
- 2019-12-27 CN CN201911382402.1A patent/CN111047624A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101408981A (en) * | 2008-11-25 | 2009-04-15 | 中国电子科技集团公司第二十八研究所 | Method for detecting infrared puniness target |
| WO2011022279A2 (en) * | 2009-08-18 | 2011-02-24 | Behavioral Recognition Systems, Inc. | Background model for complex and dynamic scenes |
| CN102184552A (en) * | 2011-05-11 | 2011-09-14 | 上海理工大学 | Moving target detecting method based on differential fusion and image edge information |
| CN103473549A (en) * | 2013-09-23 | 2013-12-25 | 北京智诺英特科技有限公司 | Image target detecting method and device |
| CN105469428A (en) * | 2015-11-26 | 2016-04-06 | 河海大学 | Morphological filtering and SVD (singular value decomposition)-based weak target detection method |
| CN106327526A (en) * | 2016-08-22 | 2017-01-11 | 湖南挚新科技发展有限公司 | Image object tracking method and image object tracking system |
| CN109033972A (en) * | 2018-06-27 | 2018-12-18 | 上海数迹智能科技有限公司 | A kind of object detection method, device, equipment and storage medium |
| CN109345472A (en) * | 2018-09-11 | 2019-02-15 | 重庆大学 | A kind of infrared moving small target detection method of complex scene |
| CN110415208A (en) * | 2019-06-10 | 2019-11-05 | 西安电子科技大学 | A kind of adaptive targets detection method and its device, equipment, storage medium |
Non-Patent Citations (4)
| Title |
|---|
| 杨杰: "基于数学形态学的Top-hat变换", 《视频目标检测和跟踪及其应用》 * |
| 牟新刚: "平台直方图均衡", 《红外探测器成像与信息处理》 * |
| 王保曾: "雷达/红外的可视化仿真与目标信息的获取", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
| 王文峰: "图像目标归一化处理", 《MATLAB计算机视觉与机器认知》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021217445A1 (en) * | 2020-04-28 | 2021-11-04 | 深圳市大疆创新科技有限公司 | Image processing method, device and system, and storage medium |
| CN112733589A (en) * | 2020-10-29 | 2021-04-30 | 广西科技大学 | Infrared image pedestrian detection method based on deep learning |
| CN113554632A (en) * | 2021-07-30 | 2021-10-26 | 沭阳天勤工具有限公司 | Artificial intelligence-based workpiece detection optimization method in turbid water body |
| CN114092777A (en) * | 2021-11-29 | 2022-02-25 | 广州辰创科技发展有限公司 | Efficient identification method and system based on small target |
| CN114092777B (en) * | 2021-11-29 | 2024-04-12 | 广州辰创科技发展有限公司 | Efficient recognition method and system based on small targets |
| CN114332756A (en) * | 2021-12-15 | 2022-04-12 | 中国飞行试验研究院 | Target identification method realized by fast-Vibe and classification algorithm |
| CN117934417A (en) * | 2024-01-24 | 2024-04-26 | 北京首发公路养护工程有限公司 | Method, device, equipment and medium for identifying apparent defects of road based on neural network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111047624A (en) | Image dim target detection method, device, equipment and storage medium | |
| CN107451999B (en) | Foreign matter detection method and device based on image recognition | |
| CN107256225B (en) | Method and device for generating heat map based on video analysis | |
| CN105404847B (en) | A kind of residue real-time detection method | |
| CN107085714B (en) | Forest fire detection method based on video | |
| CN109918971B (en) | Method and device for detecting number of people in monitoring video | |
| CN109086724B (en) | Accelerated human face detection method and storage medium | |
| CN106056079B (en) | A kind of occlusion detection method of image capture device and human face five-sense-organ | |
| US11093737B2 (en) | Gesture recognition method and apparatus, electronic device, and computer-readable storage medium | |
| CN111723644A (en) | Method and system for detecting occlusion of surveillance video | |
| CN108765491A (en) | A kind of SAR image Ship Target Detection method | |
| CN110728697A (en) | Infrared dim target detection tracking method based on convolutional neural network | |
| CN105469105A (en) | Cigarette smoke detection method based on video monitoring | |
| CN105279772B (en) | A kind of trackability method of discrimination of infrared sequence image | |
| CN110415208A (en) | A kind of adaptive targets detection method and its device, equipment, storage medium | |
| CN109712158A (en) | A kind of infrared small target catching method based on target background pixel statistical restraint | |
| CN105812618B (en) | A kind of method for testing motion and motion detection apparatus | |
| CN108765406A (en) | A kind of snow mountain detection method based on infrared remote sensing image | |
| CN111611907A (en) | Image-enhanced infrared target detection method | |
| CN111967345A (en) | Method for judging shielding state of camera in real time | |
| CN111461078A (en) | Anti-fishing monitoring method based on computer vision technology | |
| CN112614109A (en) | Image quality evaluation method, device and computer readable storage medium | |
| WO2024016632A1 (en) | Bright spot location method, bright spot location apparatus, electronic device and storage medium | |
| CN110728668A (en) | Airspace high-pass filter for maintaining small target form | |
| CN112927262A (en) | Camera lens shielding detection method and system based on video |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200421 |
|
| RJ01 | Rejection of invention patent application after publication |