CN115294606B - Millimeter wave image dark target enhancement method - Google Patents
Millimeter wave image dark target enhancement method Download PDFInfo
- Publication number
- CN115294606B CN115294606B CN202210938996.5A CN202210938996A CN115294606B CN 115294606 B CN115294606 B CN 115294606B CN 202210938996 A CN202210938996 A CN 202210938996A CN 115294606 B CN115294606 B CN 115294606B
- Authority
- CN
- China
- Prior art keywords
- human body
- image
- axis
- gray value
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 38
- 230000000295 complement effect Effects 0.000 claims abstract description 71
- 230000002708 enhancing effect Effects 0.000 claims abstract description 10
- 238000004364 calculation method Methods 0.000 claims description 16
- 230000004927 fusion Effects 0.000 claims description 12
- 210000000746 body region Anatomy 0.000 claims description 9
- 230000000877 morphologic effect Effects 0.000 claims description 9
- 238000007499 fusion processing Methods 0.000 claims description 3
- 230000009191 jumping Effects 0.000 claims description 2
- 238000001514 detection method Methods 0.000 description 8
- 238000010586 diagram Methods 0.000 description 6
- 238000004422 calculation algorithm Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000010521 absorption reaction Methods 0.000 description 2
- 238000002592 echocardiography Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 241000405965 Scomberomorus brasiliensis Species 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010191 image analysis Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000010187 selection method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/90—Dynamic range modification of images or parts thereof
- G06T5/94—Dynamic range modification of images or parts thereof based on local image properties, e.g. for local contrast enhancement
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/54—Extraction of image or video features relating to texture
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30196—Human being; Person
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/07—Target detection
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Human Computer Interaction (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a millimeter wave image dark target enhancement method, which comprises the steps of firstly calculating the axis position, the head top position and the shoulder position in a human body in a millimeter wave image of the human body, limiting a processing area, then calculating the average gray value of the human body, creating a complementary set image, and storing a complementary set of pixel points of which the gray value is lower than the average gray value in an original image in the complementary set image; extracting the area of the complement image which is the dark target, excluding the error enhancement part caused by the body structure, and only keeping the enhancement area of the dark target. And finally, weighting and fusing the complementary set image and the original image, reserving texture features of the dark target, enhancing the gray features of the dark target, and obtaining the millimeter wave image after the dark target is enhanced. The method provided by the invention combines the gray scale characteristics of the dark target and the position relation between the dark target and the human body area, effectively enhances the gray scale characteristics of the dark target, and effectively retains the texture characteristics of the dark target.
Description
Technical Field
The invention belongs to the field of millimeter wave image target detection of a human body, and particularly relates to a method for enhancing a dark target of a millimeter wave image.
Background
Millimeter wave image target detection is the key to realize the detection of contraband carried on the body surface of a human body, can be widely applied to security inspection work of airports, stations and the like, and is an effective substitute for the existing human body security inspection means. The imaging of the human body by using the millimeter waves is a precondition for realizing millimeter wave image target detection, and the active millimeter wave imaging technology irradiates the millimeter waves to the human body and receives millimeter wave echoes by using a millimeter wave radar to generate an image according to the strength difference of the echoes. According to the effect on incident millimeter waves, the nature of contraband objects carried by the human body can be divided into two categories: the reflection action of the first kind of millimeter waves is stronger than the absorption action, and the millimeter waves are represented as a region with a gray value higher than that of a human body in a millimeter wave image (a rectangular frame in fig. 1 (a), which is called as a bright target below); one type of millimeter wave has a stronger absorption effect than reflection effect on incident millimeter waves, and appears in a millimeter wave image as an area having a much lower gray value than the human body (a rectangular frame in fig. 1 (b), hereinafter referred to as a dark object).
At present, the millimeter wave image target detection method mostly uses a machine learning technology, and the training and testing principle can be briefly summarized as follows: after learning a large number of features of a desired target, a location similar to the learned features is found in the test image. The "feature of the target" refers to a feature such as texture and gray level unique to the target region (relative to the human body region). As can be seen from fig. 1, a bright target has richer texture than a dark target, and the target feature of the dark target is similar to the black background outside the human body area, and the difference between the bright target and the dark target in the target feature causes the detection rate of the dark target to be poor.
Aiming at the problem, the gray scale feature of the dark target and the position relation between the dark target and the human body area are combined, so that the texture feature of the dark target is effectively reserved while the gray scale feature of the dark target is effectively enhanced.
Disclosure of Invention
The method aims at the problem of poor detection rate of the dark target caused by unobvious target features of the dark target relative to the bright target in the millimeter wave image of the human body, combines the gray scale features of the dark target and the position relation between the dark target and the human body area, effectively enhances the gray scale features of the dark target, and effectively retains the texture features of the dark target.
A method for enhancing a dark target in a millimeter wave image comprises the following steps:
step 1, calculating the axis position, the vertex position and the shoulder position in the human body in the millimeter wave image of the human body, and limiting a processing area;
step 2, calculating the average gray value of the human body;
step 3, creating a complementary set image, and storing a complementary set of pixel points with gray values lower than the average gray value in the original image in the complementary set image;
and 4, extracting the area which is the dark target in the complementary image, excluding the error enhancement part caused by the body structure, and only reserving the enhancement area of the dark target.
And 5, weighting and fusing the complementary set image and the original image, reserving texture features of the dark target, enhancing the gray features of the dark target, and obtaining the millimeter wave image after the dark target is enhanced.
Further, the specific method of step 1 is as follows;
because the background gray value of the millimeter wave human body image is almost 0, the human body region can be segmented by using the maximum inter-class variance method to obtain a binary image of the human body region. And establishing a plane rectangular coordinate system by taking the upper left corner point of the image as an original point O, wherein the downward direction and the rightward direction of the original point are respectively the positive directions of an x axis and a y axis.
Calculating the position of the middle shaft:
and respectively taking points at the outermost edges of the left shank and the right shank of the human body under the same ordinate, and respectively recording the abscissa of the left shank and the abscissa of the right shank as legLeft and legRight. Let the abscissa of the medial axis of the human body be axis, then axis is expressed as:
calculating the position of the head top:
calculating the position of the head by means of axis, recording the vertical coordinate of the top of the head as headTop and the binary image of the human body as I OTSU The method for calculating the headTop with the total number of lines of the image being imgRows comprises the following specific steps:
1-1: in I OTSU At point (axis, 0), go in the positive y-axis direction.
1-2: and when traversing to a pixel point with the first gray value not being 0, recording the vertical coordinate of the pixel point as the headTop, stopping traversing, and finishing the calculation.
Calculating the shoulder position:
firstly, in a binary image I OTSU In (c), traverse from the vertex position (axis, headTop) in the negative x-axis direction to the first point that is not 0, with the abscissa labeled leftArm. Traversing to a first point which is not 0 sequentially along the positive direction of the y axis aiming at the axis-leftArm pixel points from (leftArm, headTop) to (axis, headTop), and recording the ordinate of the point as leftShoulder i ,i∈[0,axis-leftArm)。
Similarly, in the binary image I OTSU In (c), from the vertex position (coordinate) along the x-axisThe positive direction is traversed to the first point that is not 0 and the abscissa is denoted as rightArm. Sequentially traversing to a first point which is not 0 along the positive direction of the y axis for the rightArm-axis pixel points in the (axis, headTop) to (rightArm, headTop), and recording the ordinate of the point as rightShoulder j ,j∈[0,rightArm-axis)。
Finally, in the ordinate set leftSholder i ∩rightShoulder j The maximum value of the ordinate is found as the ordinate of the Shoulder position and is marked as Shoulder.
Further, the specific method in step 2 is as follows;
since the background gray value of the millimeter wave image of the human body is 0, the average gray value of the non-0 pixel point is the average gray value of the human body region. Recording the average gray value of the human body as avgGrey and the millimeter wave image of the human body as I src Respectively record I src The total row and column numbers of imgRows and imgCols. The calculation method of the avgGrey comprises the following specific steps:
2-1: traverse I from origin src All pixel points of (1), record I src The total gray value of all the pixels and the number of the pixels with the gray values not being 0.
2-2: after the traversal is finished, calculating I src The ratio of the total gray value of all the pixels to the number of pixels with gray values not 0 is marked as I src Avggley, the calculation ends.
Further, the specific method in step 3 is as follows;
the complementary set of images is I enhance The maximum gray value which can be stored by a single pixel point of the complementary set image file is maxGrey, I enhance The specific steps of the generation are as follows:
3-1: creating a sheet and I src Complementary set image I of the same size enhance ，I enhance The initial gray values of all the pixels in the pixel array are 0.
3-2: traverse I src All pixel points of the part below the shoulders of the middle human body take the avgGray as a complementary threshold value for I src Pixel points with internal gray value lower than avgGray are shown in I enhance In which the complement value of the gray value of the pixel point is taken as I enhance And assigning values to the pixel points at the corresponding positions.
3-3: and after traversing is finished, finishing the calculation.
For the interpretation of the complement in step 3-2: take an 8-bit unsigned single-channel grayscale map as an example: the maximum gray value that a single pixel can store is 255, and if the gray value of a pixel is 20, the complement value of the pixel is 255-20=235.
Further, the specific method in step 4 is as follows;
in a binary image I of a human body OTSU Traversing the inner edge of the human body central axis from the bottom of the image upwards, stopping when meeting the first point which is not 0, recording the vertical coordinate of the point as button Down, and the vertical coordinate of the point is the vertical coordinate of the human body crotch. Therefore, the human body leg vertical coordinate interval is [ button Down, imgRows ], the left leg of the human body is positioned on the left side of the human body central axis, the right leg of the human body is positioned on the right side of the human body central axis, the inner positions of the two legs of the human body are positioned by utilizing the information, the complementary region of the inner outer edge of the two legs is removed, and the complementary image I 'obtained after the complementary region of the outer edge of the human body is removed is obtained' enhance 。
Traversing image pixels in the transverse direction from the outer side of the human body to the inner side of the human body, and setting a traversing termination condition to ensure that the complementary sets of the dark objects cannot be removed together. Taking the removal of the outer edge complementary region on the left side of the human body as an example, a specific removal method of the outer edge complementary region on the left side of the human body is as follows:
4-1: in the case where x is equal to 0]In the region, starting from the position of x =0, traversing I in rows in the positive direction of the x axis enhance And setting the gray value of the non-zero pixel points to zero in the traversal process. When the current pixel point is nonzero and the gray value of the next pixel point at the traversal position is 0, the nonzero pixel point at the current position is set to zero, the traversal of the line is finished, and the initial position of the pixel point at the next line is jumped to continue traversal according to the rule.
4-2: finish traversing I enhance All rows of (2), the calculation is complete.
The morphological opening operation is used to eliminate the possible presence of clutter in the non-dark target complement region: maximum of shorter side length of imageBit is m, a pair of rectangular structural elements I 'with origin at the center of size m is used' enhance The morphological opening operation processing is carried out, and the complementary set image after the morphological opening operation processing is recorded as I ″) enhance 。
Further, the specific method of step 5 is as follows.
I' was calculated using the same method as in step 2 enhance Is recorded as avgGrey enhance . In order to enhance the gray scale feature of the dark target and simultaneously retain the original texture feature of the dark target, avgGrey is used enhance And I src Calculating a weighting coefficient by using the average gray value avgGrey, and performing weighted fusion I src And I enhance . The fused image is recorded as I fusion Then, I fusion The fusion process of (a) is represented as:
I fusion the method has the advantages that the texture features of the dark target are effectively reserved while the gray features of the dark target are effectively enhanced for the final enhancement result of the dark target in the human millimeter wave image.
The invention has the following beneficial effects:
aiming at the problem of poor detection rate of dark targets caused by unobvious target features of dark targets relative to bright targets in millimeter wave images of a human body, the gray scale features of the dark targets and the position relation between the dark targets and the human body area are combined, so that the texture features of the dark targets are effectively reserved while the gray scale features of the dark targets are effectively enhanced.
Drawings
FIG. 1 is a schematic diagram of bright and dark targets in a millimeter-wave image of a human body;
FIG. 2 is a schematic diagram of the segmentation result of the human millimeter wave image OTSU;
FIG. 3 is a schematic view of a spatial coordinate system of the present invention;
FIG. 4 is a schematic diagram of shoulder position calculation;
fig. 5 is a schematic diagram of an enhancement process of the present invention.
Detailed Description
The method of the present invention is further illustrated with reference to the following figures and examples.
A millimeter wave image dark target enhancement method specifically comprises the following steps:
step 1, calculating the position of a middle axis, the position of the vertex and the position of shoulders of a human body;
since the background Gray value of the millimeter wave human body image is almost 0, the human body region can be segmented by using the maximum inter-class variance Method (OTSU, OTSU n.a Threshold Selection Method from Gray-Level Histograms [ J ]. IEEE Transactions on Systems Man & Cybernetics,2007,9 (1): 62-66.) to obtain the binary image of the human body region as shown in fig. 2. As shown in fig. 3, a planar rectangular coordinate system is established with the upper left corner point of the image as an origin O, the origin is downward and rightward respectively the positive directions of the x axis and the y axis, and a schematic diagram of the coordinate system and the following key points are marked in fig. 3.
Calculating the position of the middle shaft:
and respectively taking points at the outermost edges of the left shank and the right shank of the human body under the same ordinate, and respectively recording the abscissa of the left shank and the abscissa of the right shank as legLeft and legRight. Let the abscissa of the medial axis of the human body be axis, then axis is expressed as:
calculating the head top position:
calculating the head position by means of axis, recording the head top ordinate as headTop and the human body binary image as I OTSU The method for calculating the headTop with the total number of lines of the image being imgRows comprises the following specific steps:
1-1: in I OTSU At point (axis, 0), go in the positive y-axis direction.
1-2: and when traversing to a pixel point with the first gray value not being 0, recording the vertical coordinate of the pixel point as the headTop, stopping traversing, and finishing the calculation.
The pseudo code for the headTop calculation algorithm is as follows:
calculating the shoulder position:
firstly, in a binary image I OTSU In (d), traverse is performed from the vertex position (coordinate) to the first point other than 0 in the negative x-axis direction, and the abscissa is labeled leftArm. Traversing to a first point which is not 0 sequentially along the positive direction of the y axis aiming at the axis-leftArm pixel points from (leftArm, headTop) to (axis, headTop), and recording the ordinate of the point as leftShoulder i ,i∈[0,axis-leftArm)。
Similarly, in the binary image I OTSU In (c), traverse is performed from the vertex position (coordinate (axis, headTop)) to the first point that is not 0 in the positive x-direction, and the abscissa is recorded as rightArm. Sequentially traversing to a first point which is not 0 along the positive direction of the y axis for the rightArm-axis pixel points in the (axis, headTop) to (rightArm, headTop), and recording the ordinate of the point as rightShoulder j ,j∈[0,rightArm-axis)。
Finally, in the ordinate set leftSholder i ∩rightShoulder j The maximum value of the ordinate is found as the ordinate of the Shoulder position and is marked as Shoulder.
The schematic diagram of calculating the shoulder position is shown in fig. 4, wherein the arrow direction represents the pixel point traversal direction.
Step 2, calculating the average gray value of the human body;
since the background gray value of the millimeter wave image of the human body is 0, the average gray value of the non-0 pixel point is the average gray value of the human body region. Recording the average gray value of the human body as avgGrey and the millimeter wave image of the human body as I src Respectively record I src The total row and column numbers of imgRows and imgCols. The calculation method of the avgGrey comprises the following specific steps:
2-1: traverse I from origin src All pixel points of (1), record I src The total gray value and the gray value of all the pixelsThe number of pixels not equal to 0.
2-2: after the traversal is finished, calculating I src The ratio of the total gray value of all the pixels to the number of pixels with gray values not 0 is marked as I src Avggley, the calculation ends.
The pseudo code for the avgGrey computing algorithm is as follows:
step 3, creating a complementary set image, and storing I in the complementary set image src A complementary set of pixel points having a medium gray value lower than the average gray value;
the complementary set of images is I enhance The maximum gray value which can be stored by a single pixel point of the complementary set image file is maxGrey, I enhance The specific steps of the generation are as follows:
3-1: creating a sheet and I src Complementary set image I of the same size enhance ，I enhance The initial gray values of all the pixels in the pixel array are 0.
3-2: traverse I src All pixel points of the part below the shoulder of the middle human body take avgGray as a complementary threshold value for I src Pixel points with internal gray value lower than avgGray are shown in I enhance In which the complement value of the gray value of the pixel point is taken as I enhance And assigning values to the pixel points at the corresponding positions.
3-3: and after traversing is finished, finishing the calculation.
For the interpretation of the complement in step 3-2: take an 8-bit unsigned single-channel grayscale map as an example: the maximum gray value that a single pixel can store is 255, and if the gray value of a pixel is 20, the complement value of the pixel is 255-20=235.
The pseudo code for the complement image generation algorithm is as follows:
an example of a complement image is shown in fig. 5 (a).
Step 4, extracting the area of the dark target in the complementary set image;
as shown in fig. 5 (a), the outer edge of the human body has similar gray features to the dark object, so there will be a complement of the outer edge of the human body and the dark object in the complement image at the same time, and therefore the complement area of the non-dark object needs to be removed.
In a binary image I of a human body OTSU Traversing the inner edge of the human body central axis from the bottom of the image upwards, stopping when meeting the first point which is not 0 (see fig. 3, and recording the ordinate of the point as button Down), and the ordinate of the point is the ordinate of the human body crotch. Therefore, the human body leg vertical coordinate interval is [ button Down, imgRows ], the left leg of the human body is positioned on the left side of the human body central axis, the right leg of the human body is positioned on the right side of the human body central axis, the inner positions of the two legs of the human body are positioned by utilizing the information, the complementary region of the inner outer edge of the two legs is removed, and the complementary image I 'obtained after the complementary region of the outer edge of the human body is removed is obtained' enhance 。
Traversing image pixels in the transverse direction from the outer side of the human body to the inner side of the human body, and setting a traversing termination condition to ensure that the complementary sets of the dark objects cannot be removed together. Taking the removal of the outer edge complementary region on the left side of the human body as an example, a specific removal method of the outer edge complementary region on the left side of the human body is as follows:
4-1: in the case where x is equal to 0]In the region, starting from the position of x =0, traversing I in rows in the positive direction of the x axis enhance And setting the gray value of the non-zero pixel points to zero in the traversal process. When the current pixel point is nonzero and the gray value of the next pixel point at the traversal position is 0, zeroing the nonzero pixel point at the current position, ending the traversal of the line and jumping to the initial position (the position where x = 0) of the pixel point at the next line to continue the traversal according to the rule.
4-2: finish traversing I enhance All rows of (2), the calculation is complete.
The pseudocode for the removal of the complement region of the outer edge of the human body is as follows:
mathematical Morphology (Serra j. Image Analysis and physical Morphology-Volume i.academic, 1982.) calculation can eliminate small connected domains and retain larger connected domains (connected domains: a set of interconnected valued pixels in an image). The morphological opening operation is used to eliminate the possible presence of clutter in the non-dark target complement region: the highest bit of the shorter side length of the image is m, and a rectangular structural element pair I 'with an origin of size m × m at the center is used' enhance The morphological opening operation processing is carried out, and the complementary set image after the morphological opening operation processing is recorded as I ″) enhance . For example, FIG. 5 (b) I' enhance The size is 400 × 768, so m =4. FIG. 5 (c) is an I ″, which is obtained by performing arithmetic processing on FIG. 5 (b) enhance 。
And 5, weighting and fusing the complementary set image and the original image.
I' was calculated using the same method as in step 2 enhance Is recorded as avgGrey enhance . In order to enhance the gray scale feature of the dark target and simultaneously retain the original texture feature of the dark target, avgGrey is used enhance And I src Calculating a weighting coefficient by using the average gray value avgGrey, and performing weighted fusion I src And I enhance . The fused image is recorded as I fusion Then, I fusion The fusion process of (a) is represented as:
I fusion for the final enhancement result of the dark target in the millimeter wave image of the human body (i.e. the millimeter wave image after the enhancement of the dark target, as shown in FIG. 5 (d)), there areEffectively enhancing the gray scale characteristics of the dark target and effectively retaining the texture characteristics of the dark target.
Claims (4)
1. A method for enhancing a dark target in a millimeter wave image is characterized by comprising the following steps:
step 1, calculating the axis position, the vertex position and the shoulder position in the human body in the millimeter wave image of the human body, and limiting a processing area;
step 2, calculating the average gray value of the human body;
step 3, creating a complementary set image, and storing a complementary set of pixel points with gray values lower than the average gray value in the original image in the complementary set image;
step 4, extracting the area which is the dark target in the complementary image, excluding the error enhancement part caused by the body structure, and only reserving the enhancement area of the dark target;
step 5, weighting and fusing the complementary set image and the original image, reserving texture features of the dark target, enhancing gray features of the dark target, and obtaining a millimeter wave image after the dark target is enhanced;
the specific method of the step 1 is as follows;
because the background gray value of the millimeter wave human body image is almost 0, the human body region can be segmented by using the maximum inter-class variance method to obtain a binary image of the human body region; establishing a plane rectangular coordinate system by taking the upper left corner point of the image as an original point O, wherein the downward direction and the rightward direction of the original point are respectively the positive directions of an x axis and a y axis;
calculating the position of the middle shaft:
respectively taking points at the outermost edges of the left shank and the right shank of the human body under the same ordinate, and respectively recording the abscissa of the left shank and the abscissa of the right shank as legLeft and legRight; let the abscissa of the medial axis of the human body be axis, then axis is expressed as:
calculating the head top position:
calculating the position of the head by means of axis, recording the vertical coordinate of the top of the head as headTop and the binary image of the human body as I OTSU Image lineThe method for calculating headTop with imgRows comprises the following specific steps:
1-1: in I OTSU At point (axis, 0), traversing in the positive y-axis direction;
1-2: when traversing to a pixel point with a first gray value not being 0, recording the vertical coordinate of the pixel point as a headTop and stopping traversing, and finishing the calculation;
calculating the shoulder position:
firstly, in a binary image I OTSU Traversing from the vertex position to the first point which is not 0 along the negative direction of the x-axis, and recording the abscissa as leftArm, wherein the coordinate of the vertex position is (axis, headTop); traversing to a first point which is not 0 sequentially along the positive direction of the y axis aiming at the axis-leftArm pixel points from (leftArm, headTop) to (axis, headTop), and recording the ordinate of the point as leftShoulder i ,i∈[0,axis-leftArm)；
Similarly, in the binary image I OTSU Traversing from the vertex position to the first point which is not 0 along the positive direction of the x-axis, and marking the abscissa as rightArm, wherein the coordinate of the vertex position is (axis, headTop); sequentially traversing to a first point which is not 0 along the positive direction of the y axis aiming at the rightArm-axis pixel points from (axis, headTop) to (rightArm, headTop), and recording the ordinate of the point as rightShoulder j ,j∈[0,rightArm-axis)；
Finally, in the ordinate set leftSholder i ∩rightShoulder j Finding the maximum value of the vertical coordinate as the vertical coordinate of the Shoulder position, and recording as Shoulder;
the specific method of the step 2 is as follows;
since the background gray value of the human millimeter wave image is 0, the average gray value of the non-0 pixel points is the average gray value of the human body area; recording the average gray value of the human body as avgGrey and the millimeter wave image of the human body as I src Respectively record I src The total row and column numbers of imgRows and imgCols; the calculation method of the avgGrey comprises the following specific steps:
2-1: traverse I from origin src All pixel points of (1), record I src The total gray value of all the pixel points and the number of the pixel points with the gray values not being 0;
2-2: after the traversal is finished, calculating I src The ratio of the total gray value of all the pixels to the number of pixels with gray values not 0 is marked as I src Avggley, the calculation ends.
2. The method for enhancing the dark objects in the millimeter wave images according to claim 1, wherein the step 3 is as follows;
the complementary set of images is I enhance The maximum gray value which can be stored by a single pixel point of the complementary set image file is maxGrey, I enhance The specific steps of the generation are as follows:
3-1: creating a sheet and I src Complementary set image I of the same size enhance ，I enhance The initial gray values of all the internal pixel points are 0;
3-2: traverse I src All pixel points of the part below the shoulders of the middle human body take the avgGray as a complementary threshold value for I src Pixel points with internal gray value lower than avgGray are shown in I enhance In which the complement value of the gray value of the pixel point is taken as I enhance Assigning the pixel points of the corresponding positions;
3-3: after traversing, finishing calculation;
for the interpretation of the complement in step 3-2: take an 8-bit unsigned single-channel grayscale map as an example: the maximum gray value that a single pixel can store is 255, and if the gray value of a pixel is 20, the complement value of the pixel is 255-20=235.
3. The method for enhancing the dark target in the millimeter wave image according to claim 2, wherein the step 4 is as follows;
in a binary image I of a human body OTSU Traversing the inner edge human body central axis from the bottom of the image upwards, stopping when meeting a first point which is not 0, recording the vertical coordinate of the point as button Down, and the vertical coordinate of the point is the vertical coordinate of the human body crotch; therefore, the human body leg vertical coordinate interval is [ button Down, imgRows ], the human body left leg is positioned at the left side of the human body central axis, the human body right leg is positioned at the right side of the human body central axis, and the human body legs are positioned by utilizing the informationSide position, and removing the complementary region at the inner outer edge of the two legs to obtain a complementary image I after the complementary region at the outer edge of the human body is removed e ′ nhance ；
Traversing image pixels in the transverse direction from the outer side of the human body to the inner side of the human body, and setting a traversing termination condition to ensure that the complementary sets of the dark targets cannot be removed together; taking the removal of the outer edge complementary region on the left side of the human body as an example, a specific removal method of the outer edge complementary region on the left side of the human body is as follows:
4-1: in the case where x is equal to 0]In the region, starting from the position of x =0, traversing I in rows in the positive direction of the x axis enhance Setting the gray value of the non-zero pixel point to zero in the traversal process; when the current pixel point is nonzero and the gray value of the next pixel point at the traversal position is 0, zeroing the nonzero pixel point at the current position, ending the traversal of the line and jumping to the initial position of the pixel point at the next line to continue the traversal according to the rule;
4-2: finish traversing I enhance All the rows of (2) are calculated;
the morphological opening operation is used to eliminate the possible presence of clutter in the non-dark target complement region: let m be the highest bit of the shorter edge length of the image, use the rectangular structural element pair I with the origin at the center and the size of m × m e ′ nhance Morphological opening operation processing, recording the complementary set image after the morphological opening operation processing as I e ″ nhance 。
4. The method for enhancing the dark objects in the millimeter wave images according to claim 3, wherein the step 5 is as follows;
i was calculated using the same method as in step 2 e ″ nhance Is recorded as avgGrey enhance (ii) a In order to enhance the gray scale characteristics of the dark target and simultaneously retain the original texture characteristics of the dark target, avgGrey is used enhance And I src Calculating a weighting coefficient by using the average gray value avgGrey, and performing weighted fusion I src And I enhance (ii) a The fused image is recorded as I fusion Then, I fusion The fusion process of (a) is represented as:
I fusion the method has the advantages that the texture features of the dark target are effectively reserved while the gray features of the dark target are effectively enhanced for the final enhancement result of the dark target in the human millimeter wave image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210938996.5A CN115294606B (en) | 2022-08-05 | 2022-08-05 | Millimeter wave image dark target enhancement method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210938996.5A CN115294606B (en) | 2022-08-05 | 2022-08-05 | Millimeter wave image dark target enhancement method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115294606A CN115294606A (en) | 2022-11-04 |
| CN115294606B true CN115294606B (en) | 2023-03-21 |
Family
ID=83827593
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210938996.5A Active CN115294606B (en) | 2022-08-05 | 2022-08-05 | Millimeter wave image dark target enhancement method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115294606B (en) |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9424499B2 (en) * | 2014-07-11 | 2016-08-23 | Oce-Technologies B.V. | Method for printing a grayscale raster image by grayscale value dispersion |
| CN110264523B (en) * | 2019-06-25 | 2021-06-18 | 亮风台（上海）信息科技有限公司 | Method and equipment for determining position information of target image in test image |
| CN113158899B (en) * | 2021-04-22 | 2022-07-29 | 中国科学院地理科学与资源研究所 | Village and town development state measurement method based on remote sensing luminous dark target enhancement technology |
| CN113487642A (en) * | 2021-07-08 | 2021-10-08 | 杭州电子科技大学 | Method for detecting in-vitro target by using millimeter waves for significance vision |
-
2022
- 2022-08-05 CN CN202210938996.5A patent/CN115294606B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN115294606A (en) | 2022-11-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107680054B (en) | Multi-source image fusion method in haze environment | |
| CN102156996B (en) | Image edge detection method | |
| CN103971102B (en) | Static Gesture Recognition Method Based on Finger Contour and Decision Tree | |
| CN101493891B (en) | Characteristic extracting and describing method with mirror plate overturning invariability based on SIFT | |
| CN104978715A (en) | Non-local mean image denoising method based on filtering window and parameter self-adaption | |
| CN106940791B (en) | A kind of pedestrian detection method based on low-dimensional histograms of oriented gradients | |
| CN102074017B (en) | Method and device for detecting and tracking barbell central point | |
| CN106485252B (en) | Dot matrix target image Feature point recognition method is tested in image registration | |
| CN107506795A (en) | A kind of local gray level histogram feature towards images match describes sub- method for building up and image matching method | |
| Fang et al. | Laser stripe image denoising using convolutional autoencoder | |
| Hou et al. | Combining prior knowledge with CNN for weak scratch inspection of optical components | |
| CN101661618A (en) | Method for extracting and describing image characteristics with turnover invariance | |
| CN110009670A (en) | The heterologous method for registering images described based on FAST feature extraction and PIIFD feature | |
| CN101630365B (en) | Method for extracting and describing DAISY-based feature with mirror face turning invariance | |
| CN117788472B (en) | Method for judging corrosion degree of rivet on surface of aircraft skin based on DBSCAN algorithm | |
| CN110348307B (en) | Path edge identification method and system for crane metal structure climbing robot | |
| CN117557565B (en) | Detection method and device for lithium battery pole piece | |
| Li et al. | Wafer crack detection based on yolov4 target detection method | |
| Gooda et al. | Automatic detection of road cracks using EfficientNet with residual U-net-based segmentation and YOLOv5-based detection | |
| Zhao et al. | Passenger flow monitoring of elevator video based on computer vision | |
| CN115294606B (en) | Millimeter wave image dark target enhancement method | |
| CN111178111A (en) | Two-dimensional code detection method, electronic device, storage medium and system | |
| CN106384103A (en) | Vehicle face recognition method and device | |
| CN115937205A (en) | Method, device and equipment for generating surface defect ceramic tile image and storage medium | |
| CN103606146A (en) | Corner point detection method based on circular target |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |