CN115861996B - Data acquisition method and system based on Internet of things perception and AI neural network - Google Patents
Data acquisition method and system based on Internet of things perception and AI neural network Download PDFInfo
- Publication number
- CN115861996B CN115861996B CN202310119600.9A CN202310119600A CN115861996B CN 115861996 B CN115861996 B CN 115861996B CN 202310119600 A CN202310119600 A CN 202310119600A CN 115861996 B CN115861996 B CN 115861996B
- Authority
- CN
- China
- Prior art keywords
- license plate
- image
- value
- plate image
- length
- 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
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 27
- 238000000034 method Methods 0.000 title claims abstract description 27
- 230000008447 perception Effects 0.000 title claims abstract description 10
- 238000002834 transmittance Methods 0.000 claims description 29
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000004590 computer program Methods 0.000 claims description 3
- 230000000694 effects Effects 0.000 description 6
- 238000001514 detection method Methods 0.000 description 4
- 230000003044 adaptive effect Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Image Processing (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention relates to the technical field of image processing, in particular to a data acquisition method and system based on internet of things perception and an AI neural network. The method comprises the following steps: identifying a target area in the candidate area of the license plate image; according to the dark channel value distribution in the set window where each pixel point in the target area is located, determining a dark channel estimated value after the dark area of the license plate image is enhanced; acquiring the transmissivity of the license plate image, acquiring a first length between the position of the candidate region and the farthest end, a second length of the license plate image and the offset angle of the vehicle on the parking space, and acquiring the transmissivity correction factor of the candidate region according to the first length, the second length and the offset angle; performing image enhancement processing on the license plate image according to the atmospheric light value, the transmissivity and the transmissivity correction factor to obtain a target image; and acquiring license plate number data in the target image based on the AI neural network. The invention can accurately collect the license plate number data of the vehicle.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a data acquisition method and system based on internet of things perception and an AI neural network.
Background
In smart underground parking lots, it is necessary to collect license plate number data of vehicles on each parking space in order to help customers find vehicles. Because the light of the underground parking garage is usually weak, in order to improve the accuracy of data acquisition, image enhancement processing is required to be performed on the acquired license plate images.
In the prior art, image enhancement processing is performed on license plate images through a dark channel prior method, and as vehicles on parking spaces may incline, the brightness of different areas of the license plate images is different.
Disclosure of Invention
In order to solve the problem that license plate number data in license plate images are difficult to accurately extract in the prior art, the invention provides a data acquisition method and system based on internet of things perception and an AI neural network, and the adopted technical scheme is as follows:
the invention provides a data acquisition method based on internet of things perception and an AI neural network, which comprises the following steps:
acquiring a license plate image of a vehicle, taking the area of each row of pixel points in the license plate image as a candidate area, and identifying a target area in the candidate area;
determining a dark channel estimated value of the enhanced dark region of the license plate image according to dark channel value distribution in a set window where each pixel point in the target region is located;
estimating an atmospheric light value, and acquiring the transmissivity of the license plate image according to the dark channel estimated value and the atmospheric light value;
determining the farthest deviation end of the license plate image, acquiring a first length between the position of the candidate region and the farthest deviation end, a second length of the license plate image and an offset angle of the vehicle on a parking space, and acquiring a transmissivity correction factor of the candidate region according to the first length, the second length and the offset angle;
performing image enhancement processing on the license plate image according to the atmospheric light value, the transmissivity and the transmissivity correction factor to obtain an enhanced target image;
and acquiring license plate number data in the target image based on an AI neural network.
In some embodiments, the identifying a target region in the candidate region includes:
and acquiring a gray average value of each candidate region, and taking the candidate region with the gray average value larger than a set threshold value as the target region.
In some embodiments, the determining, according to the dark channel value distribution in the set window where each pixel point in the target area is located, a dark channel estimated value after the dark area of the license plate image is enhanced includes:
determining a minimum dark channel value within the set window;
calculating the sum of the minimum dark channel values of all pixel points in all the target areas;
and obtaining the total number of the pixel points of all the target area, and calculating the ratio between the sum value and the total number of the pixel points to serve as the dark channel estimated value.
In some embodiments, the estimating the atmospheric light value comprises:
identifying a pixel with the maximum brightness in the license plate image as a target pixel;
and determining the dark channel value of the target pixel point as the atmospheric light value.
In some embodiments, the obtaining the transmittance of the license plate image according to the dark channel estimation value and the atmospheric light value includes:
and calculating a difference between the atmospheric light value and the dark channel estimation value as the transmittance.
In some embodiments, the determining the furthest offset of the license plate image comprises:
and measuring the distance between the two ends of the license plate of the vehicle and a set reference horizontal line, and determining the end with the largest distance as the farthest deviation end.
In some embodiments, the obtaining the offset angle includes:
calculating the ratio between the second length and the standard license plate length, and performing inverse cosine operation on the ratio to obtain the offset angle.
In some embodiments, the obtaining the transmittance correction factor of the candidate region according to the first length, the second length, and the offset angle, and the corresponding calculation formula includes:
wherein ,
as a factor of the correction of the transmittance,
is the first
A first length between the location of the candidate region and the furthest offset of the license plate image,
for a second length of the license plate image,
in order for the angle of the offset to be,
representing taking the sine value of the offset angle,
is an index of candidate regions.
In some embodiments, the image enhancement processing is performed on the license plate image according to the atmospheric light value, the transmittance and the transmittance correction factor to obtain an enhanced target image, and the corresponding calculation formula includes:
wherein ,
for pixel points in the target image
The region in which the candidate region is enhanced,
is the pixel point in the license plate image
The candidate region in which the current region is located,
for the transmittance of the license plate image,
as a factor of the correction of the transmittance,
in the case of an atmospheric light value,
is the pixel point coordinates.
The invention provides a data acquisition system based on an Internet of things sensing and AI neural network, which comprises a memory and a processor, wherein the processor executes a computer program stored in the memory to realize the data acquisition method based on the Internet of things sensing and AI neural network.
The invention has the following beneficial effects: when the license plate is shifted due to oblique parking of the vehicle, the brightness degree of each row of pixel points in the license plate image is different under the influence of light, so that the invention takes the area where each row of pixel points in the license plate image are as a candidate area, and then carries out image enhancement processing of different degrees on each candidate area. Because the dark channel estimated value after the dark area of the license plate image is enhanced is determined according to the dark channel distribution in the set window where each pixel point in the target area is located, noise interference can be eliminated, and the accuracy of the dark channel estimated value is improved. The offset condition of the license plate can be mastered by determining the offset far end of the license plate image and the offset angle of the vehicle on the parking space, so that the influence condition of each candidate area in the license plate due to shadow can be determined according to the condition of the license plate. The transmissivity correction factors of the candidate areas are obtained according to the first length between the positions of the candidate areas and the farthest positions, the second length of the license plate images and the offset angle of the vehicle on the parking space, the position and the offset angle of each candidate area in the license plate images are considered, namely, the influence condition of shadows of each candidate area is considered, the accuracy of the transmissivity correction factors is improved, and accurate data are improved for self-adaptive enhancement of each candidate area of the license plate images. And carrying out image enhancement processing on the license plate image according to the atmospheric light value, the transmissivity and the transmissivity correction factor to obtain an enhanced target image, realizing self-adaptive enhancement processing of different areas of the license plate image, and improving the enhancement effect of the license plate image. Because the target image is an image after license plate image self-adaption enhancement, license plate number data of the target image are acquired through the AI neural network, and the accuracy of license plate number data acquisition is improved.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions and advantages of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are only some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic flow chart of a data acquisition method and system based on the Internet of things sensing and AI neural network according to an embodiment of the invention;
fig. 2 is a schematic diagram of a vehicle offset according to an embodiment of the present invention.
Detailed Description
In order to further describe the technical means and effects adopted by the invention to achieve the preset aim, the following is a detailed description of specific implementation, structure, characteristics and effects thereof based on the data acquisition method and system based on the internet of things sensing and AI neural network, which are provided by the invention, with reference to the accompanying drawings and the preferred embodiment. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The invention provides a data acquisition method and a data acquisition system based on the internet of things perception and an AI neural network, which are specifically described below with reference to the accompanying drawings.
Referring to fig. 1, a flow chart of a data acquisition method based on internet of things sensing and AI neural network according to an embodiment of the invention is shown, and the method includes the following steps:
s101, acquiring a license plate image of a vehicle, taking the area of each row of pixel points in the license plate image as a candidate area, and identifying a bright area in the candidate area as a target area.
The data acquisition method based on the internet of things sensing and the AI neural network can be applied to intelligent parking lots, and each parking lot in the intelligent parking lots is provided with an ultrasonic parking lot detection sensor for detecting the parking condition of the vehicle so as to acquire license plate data of the vehicle through an image acquisition device such as a camera after the vehicle is parked.
Specifically, the ultrasonic parking space detection sensor can detect the vehicle in the parking space, after the completion of parking of the vehicle in the parking space is detected, the license plate image of the vehicle can be acquired through the image acquisition device, and after the license plate image of the vehicle is acquired, a bright area in the license plate image can be identified as a target area. Wherein, license plate image is RGB image.
In the practical application scene, due to the consideration of cost, license plate images of vehicles on two parking spaces can be acquired through one image acquisition device, wherein the image acquisition device can be installed at the rear of a vehicle tail and used for acquiring license plate images of the vehicle tail, or installed at the front of a vehicle head and used for acquiring license plate images of the vehicle head.
In the embodiment of the invention, each column in the license plate image is used as a candidate area, and the identification of the target area in the candidate area comprises the following steps: and acquiring a gray average value of each candidate region, and taking the candidate region with the gray average value larger than a set threshold value as a target region.
It should be noted that the setting threshold may be set according to actual requirements, and is not limited in this disclosure.
The gray value of the pixel point in each candidate region can be obtained, the gray average value of each candidate region is calculated according to the gray value, whether the gray average value is larger than a set threshold value or not is judged, if yes, the candidate region is indicated to be a bright region, the candidate region is determined to be a target region, if not, the candidate region is indicated to be a dark region, and the candidate region is determined to be a non-target region.
S102, determining a dark channel estimated value after the dark area of the license plate image is enhanced according to the dark channel value distribution in the set window where each pixel point in the target area is located.
In the embodiment of the invention, according to the dark channel value distribution in the set window where each pixel point in the target area is located, the dark channel estimated value after the dark area of the license plate image is enhanced is determined, which comprises the following steps: and determining the minimum dark channel value in the set window, calculating the sum value of the minimum dark channel values of all the pixel points in all the target areas, obtaining the total number of the pixel points in all the target areas, and calculating the ratio between the sum value and the total number of the pixel points as a dark channel estimated value.
The setting window is a window built by taking each pixel point in the target area as a center, the size of the setting window can be set according to actual requirements, no limitation is made here, and optionally, the size of the setting window can be 3×3.
The dark channel value is the smallest channel value among the three channel values of RGB of the pixel point.
Specifically, the dark channel estimation value can be calculated by the following formula:
wherein ,
is a pixel point
Is used for the dark channel estimation value of (1),
for the total number of pixels of the total target area,
is a pixel point
Is the first of (2)
The channel value of the individual channels,
is a pixel point
Is provided with a setting window for the setting window,
is the first
Pixel points in each target area
The minimum dark channel value for all pixels within the window,
for the number of target areas,
for the index of the target area,
is the abscissa of the pixel point,
index for three channels of RGB.
From the following components
It can be seen that, in the embodiment of the present invention, three channel values of RGB of a pixel point in a set window of each pixel point in each target area are obtained and compared to determine a dark channel value of each pixel point in the set window, that is, a minimum channel value of the three channel values of RGB of each pixel point is taken as the dark channel value of the pixel point, then the size of the dark channel value of each pixel point in the set window is compared, the minimum dark channel value in the set window is determined, the minimum dark channel value in the set window is taken as the minimum dark channel value of the corresponding pixel point, and each of all target areas is takenThe average value of the minimum dark channel values of the pixel points is used as the dark channel estimation value, namely
. The minimum dark channel value of the corresponding pixel point is determined by setting the dark channel values of a plurality of pixel points in the window, so that noise interference can be avoided, the accuracy of the dark channel estimated value is improved, and a reliable basis is provided for subsequent image enhancement processing.
S103, estimating an atmospheric light value, and acquiring the transmissivity of the license plate image according to the dark channel estimated value and the atmospheric light value.
In an embodiment of the present invention, estimating an atmospheric light value includes: and identifying a pixel with the maximum brightness in the license plate image as a target pixel, and determining a dark channel value of the target pixel as an atmospheric light value.
In some embodiments, the gray value of each pixel in the license plate image is obtained and compared, the pixel with the largest gray value is determined to be the target pixel, then the three RGB channel values of the target pixel are obtained and compared, and the smallest channel value is determined to be the atmospheric light value.
Further, in the embodiment of the present invention, obtaining the transmittance of the license plate image according to the estimated value of the dark channel and the atmospheric light value includes: the difference between the atmospheric light value and the dark channel estimation value is calculated as the transmittance.
Specifically, the transmittance may be calculated by the following formula:
wherein ,
in order for the transmittance to be high,
in the case of an atmospheric light value,
is a dark channel estimate.
S104, determining the farthest offset end of the license plate image, acquiring a first length between the position of the candidate region and the farthest offset end, a second length of the license plate image and an offset angle of the vehicle on the parking space, and acquiring a transmissivity correction factor of the candidate region according to the first length, the second length and the offset angle.
The offset angle is an angle between a horizontal line where a license plate of a vehicle is located after the vehicle is parked and a corresponding parking line, for example, an angle between the horizontal line where the license plate of a parking space is located and the parking line behind a vehicle tail, and it can be understood that the angle of the vehicle on the parking space is also an offset angle of the license plate.
When a vehicle is parked on a parking space in an inclined way, illumination on a license plate changes, shadow shielding can occur in partial areas, so that the collected license plate images are different in brightness of pixel points in different columns, namely, the offset angles of the vehicle on the parking space are different, namely, the brightness of each candidate area in the license plate images is different, at the moment, if the license plate images are subjected to image enhancement through single transmittance, the enhancement degree of the dark areas of the license plate images is insufficient, or the enhancement degree of the bright areas excessively causes overexposure of the bright areas, and therefore, the transmittance of each candidate area needs to be adjusted according to the offset angles, so that adaptive image enhancement processing is carried out on each candidate area of the license plate images.
After the vehicle is obliquely parked on a parking space, one side of the license plate, which is far away from the image acquisition device, presents darker conditions, and in order to ensure the image enhancement effect, stronger image enhancement processing is performed on one side, which is far away from the license plate, and weaker image enhancement processing is performed on one side, which is far away from the license plate, so that the offset condition of the license plate needs to be determined before the image enhancement processing is performed on the license plate image.
In the embodiment of the invention, determining the farthest deviation end of the license plate image comprises the following steps: and measuring the distance between the two ends of the license plate of the vehicle and a set reference horizontal line, and determining one end with the largest distance as the farthest offset end.
The preset reference horizontal line can be a horizontal line where the pre-installed ultrasonic parking space detection sensor is located.
In some embodiments, the distance between the two ends of the license plate of the vehicle and the set reference horizontal line can be measured by the pre-installed ultrasonic parking space detection sensor and compared, so that the end farthest from the set reference horizontal line is determined to be the farthest end of the license plate image.
Further, in an embodiment of the present invention, the process for obtaining the offset angle includes: and calculating the ratio between the second length and the standard license plate length, and performing inverse cosine operation on the comparison value to obtain an offset angle.
Specifically, the offset angle of the vehicle on the parking space can be calculated by the following formula:
wherein ,
in order for the angle of the offset to be,
for a second length of the license plate image,
for the length of a standard license plate,
as an inverse cosine function.
The second length is the transverse length of the license plate image. In some embodiments, the abscissas of the left and right edges of the license plate image are obtained, and the absolute value of the difference between the two abscissas is calculated as the second length.
In some embodiments, the license plate of the vehicle which is not offset can be shot through the image acquisition device to obtain a standard license plate image, then the abscissa of the left and right side edges of the standard license plate image is acquired, and the absolute value of the difference between the two abscissas is calculated as the standard license plate length.
The image acquisition device is positioned at the same position when acquiring the license plate image and the standard license plate image, and the shooting parameters are unchanged.
Fig. 2 is a schematic diagram of vehicle offset according to an embodiment of the present invention, as shown in fig. 2, an offset angle of a vehicle on a parking space is obtained through a second length of a license plate image and a standard license plate degree, accuracy of the offset angle is guaranteed, and a reliable basis is provided for subsequent image enhancement processing.
Further, in the embodiment of the present invention, according to the first length, the second length and the offset angle, the transmittance correction factor of the candidate region is obtained, and the corresponding calculation formula includes:
wherein ,
as a factor of the correction of the transmittance,
is the first
A first length between the location of the candidate region and the furthest offset of the license plate image,
for a second length of the license plate image,
in order for the angle of the offset to be,
representing taking the sine value of the offset angle,
is an index of candidate regions.
The offset angle is as follows
Between 0 degrees and 45 degrees.
In some embodiments, for each candidate region, a first abscissa of the candidate region and a second abscissa of the license plate image that deviates from the farthest end may be acquired, and an absolute value of a difference between the first abscissa and the second abscissa is calculated as a first length between a location of the candidate region and the farthest end of the license plate image that deviates from the candidate region.
In the embodiment of the invention, the larger the offset angle of the license plate is, the larger the degree of influence of shadow is, so the offset angle is
The larger the license plate image is, the larger the enhanced intensity is, and correspondingly, the larger the transmissivity correction factor is, and conversely, the offset angle is
The smaller the enhanced intensity of the license plate image is, the smaller the transmittance correction factor is, and therefore, the transmittance correction factor is
Angle with offset angle
And has positive correlation.
Further, a first length between the candidate region and the far-most offset end of the license plate image
The larger the area is, the less the area is affected by shading, the less the area is enhanced, and accordingly, the transmittance correction factor
Smaller, conversely, the first length between the candidate region distance and the far-off end of the license plate image
The smaller the area is, the greater the extent to which the shadow affects the area, the greater the extent to which the area is enhanced, and accordingly the transmittance correction factor
The larger the transmittance correction factor is, therefore
And a first length of
And has a negative correlation.
In the embodiment of the invention, the first length between the position of each candidate area and the far-end deviated from the license plate image can be obtained, so that the transmissivity correction factor of each candidate area can be obtained according to the first length and the deviation angle, namely, the transmissivity correction factor of each candidate area can be determined according to the position of each candidate area and the deviation condition of the vehicle, and therefore, the transmissivity of each candidate area can be corrected according to the transmissivity correction factor. The transmissivity correction factors of each candidate region are determined by considering the influence of the offset of the license plate and the position of each candidate region on the shadow degree of the license plate, and the accuracy of the transmissivity correction factors is improved, so that different regions of the license plate image can be subjected to image enhancement processing of different degrees.
S105, performing image enhancement processing on the license plate image according to the atmospheric light value, the transmissivity and the transmissivity correction factor to obtain an enhanced target image.
In the embodiment of the invention, the license plate image is subjected to image enhancement processing according to the atmospheric light value, the transmissivity and the transmissivity correction factor, the enhanced target image is obtained, and the corresponding calculation formula comprises:
wherein ,
for pixel points in the target image
The region in which the candidate region is enhanced,
is the pixel point in the license plate image
The candidate region in which the current region is located,
for the transmittance of the license plate image,
as a factor of the correction of the transmittance,
in the case of an atmospheric light value,
is the coordinates of the pixel point.
As can be seen from the above formula for performing image enhancement processing on the license plate image, in the embodiment of the present invention, the transmittance correction factor is introduced to correct the transmittance of each candidate region in the license plate image, so as to implement adaptive image enhancement of different regions of the license plate image. The transmissivity correction factors are determined according to the offset angle and the area position of the license plate, when the transmissivity is corrected, the transmissivity of different areas can be corrected according to the offset angle and the area position of the license plate, so that the image enhancement processing of different areas of the license plate image is realized, the image enhancement processing of the dark area of the license plate image with higher degree can be performed, the image light-receiving processing of the bright area with weaker degree can be performed, the bright area of the license plate image can be prevented from being overexposed while the dark area of the license plate image is enhanced, the image enhancement effect of the license plate image is improved, and the definition of the license plate image is ensured.
S106, acquiring license plate number data in the target image based on the AI neural network.
In some embodiments, the target image may be input into a trained AI neural network, and license plate number data in the target image may be extracted by the AI neural network.
In the embodiment of the invention, the target image is the image obtained by carrying out self-adaptive enhancement processing on the license plate image, so that the image is clearer, and the license plate number in the target image can be accurately identified through the AI neural network, thereby improving the accuracy of license plate number data acquisition.
In summary, in the embodiment of the present invention, when the license plate is offset due to oblique parking of the vehicle, the brightness of each row of pixels in the license plate image is different under the influence of light, so that the present invention uses the area where each row of pixels in the license plate image is located as a candidate area, and then performs image enhancement processing to different degrees on each candidate area. Because the dark channel estimated value after the dark area of the license plate image is enhanced is determined according to the dark channel distribution in the set window where each pixel point in the target area is located, noise interference can be eliminated, and the accuracy of the dark channel estimated value is improved. The offset condition of the license plate can be mastered by determining the offset far end of the license plate image and the offset angle of the vehicle on the parking space, so that the influence condition of each candidate area in the license plate due to shadow can be determined according to the condition of the license plate. The transmissivity correction factors of the candidate areas are obtained according to the first length between the positions of the candidate areas and the farthest positions, the second length of the license plate images and the offset angle of the vehicle on the parking space, the position and the offset angle of each candidate area in the license plate images are considered, namely, the influence condition of shadows of each candidate area is considered, the accuracy of the transmissivity correction factors is improved, and accurate data are improved for self-adaptive enhancement of each candidate area of the license plate images. And carrying out image enhancement processing on the license plate image according to the atmospheric light value, the transmissivity and the transmissivity correction factor to obtain an enhanced target image, realizing self-adaptive enhancement processing of different areas of the license plate image, and improving the enhancement effect of the license plate image. Because the target image is an image after license plate image self-adaption enhancement, license plate number data of the target image are acquired through the AI neural network, and the accuracy of license plate number data acquisition is improved.
The invention also provides a data acquisition system based on the Internet of things sensing and AI neural network, which comprises a memory and a processor, wherein the processor executes a computer program stored in the memory to realize the data acquisition method based on the Internet of things sensing and AI neural network.
It should be noted that: the sequence of the embodiments of the present invention is only for description, and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments.
Claims (9)
1. The data acquisition method based on the internet of things perception and the AI neural network is characterized by comprising the following steps:
acquiring a license plate image of a vehicle, taking the area of each row of pixel points in the license plate image as a candidate area, and identifying a target area in the candidate area;
determining a dark channel estimated value of the enhanced dark region of the license plate image according to dark channel value distribution in a set window where each pixel point in the target region is located;
estimating an atmospheric light value, and acquiring the transmissivity of the license plate image according to the dark channel estimated value and the atmospheric light value;
determining the farthest deviation end of the license plate image, acquiring a first length between the position of the candidate region and the farthest deviation end, a second length of the license plate image and an offset angle of the vehicle on a parking space, and acquiring a transmissivity correction factor of the candidate region according to the first length, the second length and the offset angle;
performing image enhancement processing on the license plate image according to the atmospheric light value, the transmissivity and the transmissivity correction factor to obtain an enhanced target image;
the determination mode of the deviation from the farthest end is as follows: measuring and comparing the distances between the two ends of the license plate of the vehicle and the set reference horizontal line, and taking the end farthest from the set reference horizontal line as the farthest deviated end of the license plate image;
the second length is obtained by the following steps: acquiring the abscissa of the edges of the left side and the right side of the license plate image, and calculating the absolute value of the difference between the two abscissas as a second length;
acquiring license plate number data in the target image based on an AI neural network;
and acquiring a transmissivity correction factor of the candidate region according to the first length, the second length and the offset angle, wherein a corresponding calculation formula comprises:
wherein ,
for the transmissivity correction factor, +.>
Is->
A first length between the location of the candidate region and the far-end offset from the license plate image, < ->
For the second length of the license plate image, +.>
For the purpose of offset angle +>
Representing the sine value taking the offset angle, +.>
Is an index of candidate regions.
2. The method of claim 1, wherein the identifying the target region of the candidate regions comprises:
and acquiring a gray average value of each candidate region, and taking the candidate region with the gray average value larger than a set threshold value as the target region.
3. The method according to claim 1, wherein the determining the dark channel estimation value after the enhancement processing of the dark area of the license plate image according to the dark channel value distribution in the set window where each pixel point in the target area is located includes:
determining a minimum dark channel value within the set window;
calculating the sum of the minimum dark channel values of all pixel points in all the target areas;
and obtaining the total number of the pixel points of all the target area, and calculating the ratio between the sum value and the total number of the pixel points to serve as the dark channel estimated value.
4. The method of claim 1, wherein the estimating an atmospheric light value comprises:
identifying a pixel with the maximum brightness in the license plate image as a target pixel;
and determining the dark channel value of the target pixel point as the atmospheric light value.
5. The method of claim 1, wherein the obtaining the transmittance of the license plate image from the dark channel estimate and the atmospheric light value comprises:
and calculating a difference between the atmospheric light value and the dark channel estimation value as the transmittance.
6. The method of claim 1, wherein the determining the furthest offset of the license plate image comprises:
and measuring the distance between the two ends of the license plate of the vehicle and a set reference horizontal line, and determining the end with the largest distance as the farthest deviation end.
7. The method of claim 1, wherein the obtaining the offset angle comprises:
calculating the ratio between the second length and the standard license plate length, and performing inverse cosine operation on the ratio to obtain the offset angle.
8. The method according to claim 1, wherein the performing image enhancement processing on the license plate image according to the atmospheric light value, the transmittance, and the transmittance correction factor to obtain an enhanced target image, and the corresponding calculation formula includes:
wherein ,
for pixel point in target image +.>
Region in which the candidate region is enhanced, +.>
Is the pixel point in the license plate image>
Candidate region where ∈exists->
For the transmissivity of license plate image +.>
For the transmissivity correction factor, +.>
Is atmospheric light value, +.>
Is the coordinates of the pixel point.
9. A data acquisition system based on internet of things perception and AI neural network, comprising a memory and a processor, wherein the processor executes a computer program stored in the memory to implement a data acquisition method based on internet of things perception and AI neural network as claimed in any one of claims 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310119600.9A CN115861996B (en) | 2023-02-16 | 2023-02-16 | Data acquisition method and system based on Internet of things perception and AI neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310119600.9A CN115861996B (en) | 2023-02-16 | 2023-02-16 | Data acquisition method and system based on Internet of things perception and AI neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115861996A CN115861996A (en) | 2023-03-28 |
| CN115861996B true CN115861996B (en) | 2023-05-09 |
Family
ID=85658146
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310119600.9A Active CN115861996B (en) | 2023-02-16 | 2023-02-16 | Data acquisition method and system based on Internet of things perception and AI neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115861996B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116110053B (en) * | 2023-04-13 | 2023-07-21 | 济宁能源发展集团有限公司 | Container surface information detection method based on image recognition |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115331264A (en) * | 2022-10-13 | 2022-11-11 | 山东济矿鲁能煤电股份有限公司阳城煤矿 | Mine worker safety identification management method based on neural network |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH10185815A (en) * | 1996-12-20 | 1998-07-14 | Seitai Hikari Joho Kenkyusho:Kk | Method for measuring transmittance distribution, and method for forming ct image |
| CN114581914A (en) * | 2022-01-27 | 2022-06-03 | 中通服咨询设计研究院有限公司 | License plate recognition method in haze environment |
| CN114945061A (en) * | 2022-04-29 | 2022-08-26 | 江苏科技大学 | Mechanism for active compensation of underwater camera shooting and image enhancement method |
| CN115376117A (en) * | 2022-07-08 | 2022-11-22 | 南京航空航天大学 | License plate recognition method in rainy and foggy weather |
-
2023
- 2023-02-16 CN CN202310119600.9A patent/CN115861996B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115331264A (en) * | 2022-10-13 | 2022-11-11 | 山东济矿鲁能煤电股份有限公司阳城煤矿 | Mine worker safety identification management method based on neural network |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115861996A (en) | 2023-03-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107014294B (en) | Contact net geometric parameter detection method and system based on infrared image | |
| CN108470356B (en) | Target object rapid ranging method based on binocular vision | |
| CN115861996B (en) | Data acquisition method and system based on Internet of things perception and AI neural network | |
| EP2168079A1 (en) | Method and system for universal lane boundary detection | |
| CN109035287B (en) | Foreground image extraction method and device and moving vehicle identification method and device | |
| CN111444778A (en) | Lane line detection method | |
| CN112215794B (en) | Method and device for detecting dirt of binocular ADAS camera | |
| WO2021012647A1 (en) | Face verification method and apparatus, server and readable storage medium | |
| CN117115153B (en) | Intelligent printed circuit board quality detection method based on visual assistance | |
| CN107341793A (en) | A kind of target surface image processing method and device | |
| JP3550874B2 (en) | Monitoring device | |
| CN114620030A (en) | Parking performance testing method and device, server and product | |
| CN108563994B (en) | Parking lot parking space identification method based on image similarity | |
| CN116843584B (en) | Image data optimization enhancement method | |
| CN116188510B (en) | Enterprise emission data acquisition system based on multiple sensors | |
| CN111369806A (en) | Method and device for photographing, measuring speed and identifying license plate of freight train | |
| CN109661683B (en) | Structured light projection method, depth detection method and structured light projection device based on image content | |
| CN113450335B (en) | Road edge detection method, road edge detection device and road surface construction vehicle | |
| CN113834447B (en) | High-dynamic laser light bar self-adaptive imaging processing method in outdoor complex environment | |
| CN111611930B (en) | Parking space line detection method based on illumination consistency | |
| CN112183454B (en) | Image detection method and device, storage medium and terminal | |
| US10789688B2 (en) | Method, device, and system for enhancing changes in an image captured by a thermal camera | |
| CN116503481B (en) | Automatic parking position and orientation detecting system based on image visual guidance | |
| CN117495695B (en) | Low-light environment detection system based on millimeter wave and infrared image fusion | |
| CN116091909A (en) | Method for quantitatively measuring underwater light intensity distribution |
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 |