CN110443240B - Picture processing method, device and equipment and computer readable storage medium - Google Patents
Picture processing method, device and equipment and computer readable storage medium Download PDFInfo
- Publication number
- CN110443240B CN110443240B CN201910648054.1A CN201910648054A CN110443240B CN 110443240 B CN110443240 B CN 110443240B CN 201910648054 A CN201910648054 A CN 201910648054A CN 110443240 B CN110443240 B CN 110443240B
- Authority
- CN
- China
- Prior art keywords
- article
- picture
- pixel
- determining
- welt
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/24—Aligning, centring, orientation detection or correction of the image
- G06V10/245—Aligning, centring, orientation detection or correction of the image by locating a pattern; Special marks for positioning
-
- 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
- G06V10/255—Detecting or recognising potential candidate objects based on visual cues, e.g. shapes
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a picture processing method, a device, equipment and a computer readable storage medium, wherein the method comprises the following steps: obtaining a picture, and predicting the picture by adopting a deep learning algorithm to obtain pixel information of an article in the picture; determining a pixel area of an article according to pixel information of the article in the picture, and judging whether the article is a welt article; and if so, removing the pixel area of the welt article in the picture. The invention prevents the same article from being repeatedly grabbed due to the edge attaching condition to cause resource waste, and solves the problem of multi-grabbing (empty grabbing) caused by the change of the grabbing center when the article is attached.
Description
Technical Field
The present invention relates to the field of image processing, and in particular, to an image processing method, an image processing apparatus, a device and a computer-readable storage medium.
Background
With the development of computer technology, the technology of image processing is also more and more advanced, and the current article classification (such as garbage classification) also adopts the image processing technology to perform rapid classification, for example, on a garbage sorting production line, a conveyor belt is used for conveying garbage, a camera is used for taking images, the images taken by the camera are used for determining the garbage to be captured and then captured by a robot, but because the images taken by the camera are controlled by software/hardware, the view of the garbage entering the camera is random, and because the image taking speed of the camera is faster than that of the conveyor belt, a garbage object may appear in a plurality of images. If the same garbage target object is complete in a plurality of pictures, the recognized grabbing centers are also overlapped, and the robot can treat the object with the overlapped grabbing centers as an object, namely grab the object once. However, the garbage object may not be complete in the camera view field at different times, that is, the garbage edge (i.e., the garbage edge) occurs, the capture center of the object identified by the image is different from the image of the complete object, and the same garbage object is repeatedly captured by the robot. Obviously, the existing article edge attaching mode easily causes the article to be repeatedly grabbed.
Disclosure of Invention
The invention mainly aims to provide a picture processing method, a picture processing device, picture processing equipment and a computer readable storage medium, and aims to solve the technical problem that in the picture processing method in the prior art, articles are easily grabbed repeatedly due to article welting.
In order to achieve the above object, the present invention provides a picture processing method, including the steps of:
obtaining a picture, and predicting the picture by adopting a deep learning algorithm to obtain pixel information of an article in the picture;
determining a pixel area of an article according to pixel information of the article in the picture, and judging whether the article is a welted article;
and if so, removing the pixel area of the welt article in the picture.
Optionally, the step of determining a pixel area of the article according to pixel information of the article in the picture, and determining whether the article is a welt article includes:
determining a pixel area of an article according to pixel information of the article in the picture, and drawing a rectangular frame containing the article;
determining the arrangement direction of each article in the picture according to the rectangular frame corresponding to each article;
determining two opposite sides vertical to the arrangement direction in the picture;
calculating the distance from the rectangular frame corresponding to each article to two opposite sides, and determining the shortest distance;
and if the shortest distance is smaller than a preset constant, determining that the article is a welt article.
Optionally, the step of determining a pixel region of the article according to the pixel information of the article in the picture, and drawing a rectangular frame containing the article includes:
determining the maximum and minimum coordinates of the length/width of the pixel area of the article according to the pixel information of the article in the picture;
drawing a rectangular frame containing the pixel area of the article according to the maximum and minimum coordinates of the length/width of the pixel area of the article.
Optionally, before the step of determining a pixel region of the article according to pixel information of the article in the picture and determining whether the article is a welt article, the method includes:
extracting outline data of the article from pixel data of a front group of deep learning prediction results to push the outline data into a first set, wherein the first set is used for comparing with a second set;
the step of determining the pixel area of the article according to the pixel information of the article in the picture and judging whether the article is a welt article comprises the following steps:
traversing the pixel area of the article according to the pixel information of the article in the picture to obtain the pixel coordinate of the corresponding outline of the welting pixel area of the article;
and pressing the pixel coordinates of the corresponding outline of the article into the second set to determine whether the article is a welt article according to the pixel coordinates of the outline.
Optionally, the step of determining whether the article is a welt article according to the pixel coordinates of the outline includes:
traversing the first set with the second set as a primary set;
and if the contour similarity of the articles in the second set reaches a preset threshold after the articles in the first set are subjected to coordinate translation, determining that the articles are welt articles.
Further, to achieve the above object, the present invention further provides an image processing apparatus, including:
the acquisition module is used for acquiring pictures;
the prediction module is used for predicting the picture by adopting a deep learning algorithm so as to obtain pixel information of articles in the picture;
the judging module is used for determining a pixel area of an article according to pixel information of the article in the picture and judging whether the article is a welted article;
and if so, removing the pixel area of the welt article in the picture.
Optionally, the determining module includes:
the drawing unit is used for determining a pixel area of the article according to the pixel information of the article in the picture and drawing a rectangular frame containing the article;
the determining unit is used for determining the arrangement direction of each article in the picture according to the rectangular frame corresponding to each article;
the determining unit is further configured to determine two opposite sides perpendicular to the arrangement direction in the picture;
the calculation unit is used for calculating the distance from the rectangular frame corresponding to each article to the two opposite sides and determining the shortest distance;
the determining unit is further configured to determine that the article is a welt article if the shortest distance is smaller than a preset constant.
Optionally, the drawing unit is further configured to determine a maximum and minimum coordinate of a length/width of a pixel region of the article according to the pixel information of the article in the picture;
drawing a rectangular frame containing the pixel area of the article according to the maximum and minimum coordinates of the length/width of the pixel area of the article.
Further, to achieve the above object, the present invention also provides a picture processing device, which includes a memory, a processor, and a picture processing program stored on the memory and executable on the processor, and when executed by the processor, the picture processing program implements the steps of the picture processing method as described above.
Further, to achieve the above object, the present invention also provides a computer readable storage medium, on which a picture processing program is stored, which when executed by a processor implements the steps of the picture processing method as described above.
The technical scheme of the invention discloses a picture processing method, which comprises the steps of firstly obtaining a picture, predicting the picture by adopting a deep learning algorithm to obtain pixel information of an article in the picture, then determining a pixel region of the article according to the pixel information of the article in the picture, judging whether the article is a welted article, and if so, removing the pixel region of the welted article in the picture, so that in the process of classifying the article, the welted article in the picture is removed, the resource waste caused by repeated grabbing of the same article due to the welting condition is prevented, and the phenomenon of multi-grabbing (empty grabbing) caused by the change of a grabbing center during welting of the article is solved.
Drawings
Fig. 1 is a schematic structural diagram of a hardware operating environment of a device according to an embodiment of the present invention;
FIG. 2 is a flowchart illustrating a first embodiment of a method for processing pictures according to the present invention;
FIG. 3 is a flowchart illustrating a second embodiment of a method for processing pictures according to the present invention;
FIG. 4 is a schematic diagram of an application scenario of the present invention;
FIG. 5 is a flowchart illustrating a third exemplary embodiment of a method for processing pictures according to the present invention;
FIG. 6 is a functional block diagram of the image processing apparatus according to the present invention.
The implementation, functional features and advantages of the present invention will be further described with reference to the accompanying drawings.
Detailed Description
It should be understood that the specific embodiments described herein are merely illustrative of the invention and do not limit the invention.
The invention provides a picture processing device, and referring to fig. 1, fig. 1 is a schematic structural diagram of a device hardware operating environment according to an embodiment of the picture processing device of the invention.
As shown in fig. 1, the picture processing apparatus may include: a processor 1001, e.g. a CPU, a communication bus 1002, a user interface 1003, a network interface 1004, a memory 1005. Wherein a communication bus 1002 is used to enable connective communication between these components. The user interface 1003 may include a standard wired interface, a wireless interface. The network interface 1004 may optionally include a standard wired interface, a wireless interface (e.g., WI-FI interface). The memory 1005 may be a high-speed RAM memory or a non-volatile memory such as a disk memory. The memory 1005 may alternatively be a memory device separate from the processor 1001 described above.
Those skilled in the art will appreciate that the hardware configuration of the picture processing device shown in fig. 1 does not constitute a limitation of the picture processing device, and may include more or less components than those shown, or combine some components, or a different arrangement of components.
As shown in fig. 1, a memory 1005, which is a kind of computer-readable storage medium, may include therein an operating system, a network communication module, a user interface module, and a picture processing program. The operating system is a program for managing and controlling the picture processing equipment and software resources, and supports the operation of a network communication module, a user interface module, a picture processing program and other programs or software; the network communication module is used to manage and control the network interface 1004; the user interface module is used to manage and control the user interface 1003.
In the hardware configuration of the picture processing apparatus shown in fig. 1, the processor 1001 may call a picture processing program stored in the memory 1005, and execute the steps of the following picture processing method:
obtaining a picture, and predicting the picture by adopting a deep learning algorithm to obtain pixel information of an article in the picture;
determining a pixel area of an article according to pixel information of the article in the picture, and judging whether the article is a welted article;
and if so, removing the pixel area of the welt article in the picture.
Further, the step of determining the pixel area of the article according to the pixel information of the article in the picture and judging whether the article is a welt article comprises:
determining a pixel area of the article according to the pixel information of the article in the picture, and drawing a rectangular frame containing the article;
determining the arrangement direction of each article in the picture according to the rectangular frame corresponding to each article;
determining two opposite sides vertical to the arrangement direction in the picture;
calculating the distance from the rectangular frame corresponding to each article to two opposite sides, and determining the shortest distance;
and if the shortest distance is smaller than a preset constant, determining that the article is a welt article.
Further, the step of determining a pixel area of the article according to the pixel information of the article in the picture and drawing a rectangular frame containing the article includes:
determining the maximum and minimum coordinates of the length/width of the pixel area of the article according to the pixel information of the article in the picture;
and drawing a rectangular frame containing the pixel area of the article according to the maximum and minimum coordinates of the length/width of the pixel area of the article.
Further, before the step of determining a pixel area of the article according to pixel information of the article in the picture and determining whether the article is a welted article, the method includes:
extracting outline data of the article from the pixel data of the front group of deep learning prediction results to press the outline data into a first set, wherein the first set is used for comparing with a second set;
the step of determining the pixel area of the article according to the pixel information of the article in the picture and judging whether the article is a welt article comprises the following steps:
traversing the pixel area of the article according to the pixel information of the article in the picture to obtain the pixel coordinate of the corresponding outline of the welting pixel area of the article;
and pressing the pixel coordinates of the corresponding outline of the article into the second set to determine whether the article is a welt article according to the pixel coordinates of the outline.
Further, the step of determining whether the article is a welt article according to the pixel coordinates of the outline comprises:
traversing the first set with the second set as a primary set;
and if the contour similarity of the articles in the second set reaches a preset threshold value after the articles in the first set pass through coordinate translation, determining that the articles are welt articles.
The specific implementation of the image processing apparatus of the present invention is substantially the same as the embodiments of the image processing method described above, and is not described herein again.
The invention also provides a picture processing method.
While a logical order is shown in the flow chart, in some cases, the steps shown or described may be performed in an order different from that shown or described herein.
In the respective embodiments of the picture processing method, for convenience of description, the execution subject is omitted for explaining the respective embodiments.
Referring to fig. 2, fig. 2 is a schematic flowchart of a first embodiment of a picture processing method according to the present invention, where the picture processing method includes:
s10, acquiring a picture, and predicting the picture by adopting a deep learning algorithm to obtain pixel information of an article in the picture;
in the embodiment, the camera is used for shooting pictures, the camera can optionally collect one frame of picture at preset time intervals, then the collected one frame of picture is obtained for processing, and the preset time intervals are set according to actual requirements. In addition, one frame of picture can be collected in real time, and the collected frame of picture is obtained for processing.
After the picture is obtained, predicting the picture by adopting a deep learning algorithm to obtain pixel information of the article in the picture, wherein the mode of predicting the picture by adopting the deep learning algorithm to obtain the pixel information of the article in the picture is consistent with the prior art, and the details are not repeated here. Step S20, determining a pixel area of an article according to pixel information of the article in the picture, and judging whether the article is a welt article;
after the pixel information of the article in the picture is obtained, because the pixel information of different articles in the picture is different, or the pixels of the article in the picture and the background are different, the pixel area of the article can be determined according to the pixel information of the article in the picture, and then whether the article is a welted article or not is judged. Specifically, whether the article is welt useless or not is judged, a central point of the image can be set in the image as a coordinate origin, then, coordinates of all pixel points except the central point in the image are determined according to the coordinate origin, a rough welt area in the image is determined according to the coordinates of all the pixel points, then, whether the pixel area of the welt article is matched with the welt area or not is judged based on the pixel area of the welt article, and if the pixel area of the welt article is matched with the welt area, the article can be determined to be the welt article.
In addition, after the pixel information of the article in the picture is obtained, the pixel information is converted into pixel coordinates, and the result set is traversed, so that the pixel area of the article is determined according to the pixel information of the article in the picture, and whether the article is a welt article or not is judged.
And step S30, if yes, removing the pixel area of the welt article from the picture.
If the fact that the article is the welt article is determined in the picture, the pixel area of the welt article can be removed from the picture, the picture of the pixel area with the welt article removed is obtained, at the moment, the welt article is not contained in the picture, therefore, when the robot on the article sorting production line grabs the article, the situation that the same article is grabbed for multiple times due to the fact that the welt condition exists and multiple times of grabbing are achieved is avoided, and waste of resources is avoided.
In the embodiment, a picture is obtained first, the picture is predicted by adopting a deep learning algorithm to obtain pixel information of an article in the picture, then a pixel region of the article is determined according to the pixel information of the article in the picture, whether the article is a welt article is judged, if yes, the pixel region of the welt article is removed from the picture, so that in the process of classifying the article, the welt article in the picture is removed, resource waste caused by repeated grabbing of the same article due to the fact that a welt condition exists is prevented, and the phenomenon of multiple grabbing (empty grabbing) caused by the change of a grabbing center when the article is welt is solved.
Further, a second embodiment of the present invention is proposed based on the first embodiment of the picture processing method. In this embodiment, referring to fig. 3, the step S20 includes:
step S21, determining a pixel area of the article according to the pixel information of the article in the picture, and drawing a rectangular frame containing the article;
s22, determining the arrangement direction of each article in the picture according to the rectangular frame corresponding to each article;
step S23, determining two opposite sides vertical to the arrangement direction in the picture;
s24, calculating the distance from the rectangular frame corresponding to each article to two opposite sides, and determining the shortest distance;
and S25, if the shortest distance is smaller than a preset threshold value, determining that the article is a welted article.
In this embodiment, after obtaining pixel information of an article in a picture, a pixel region of the article is determined according to the pixel information of the article in the picture, and a rectangular frame containing the article is drawn, specifically, the step S21 includes:
determining the maximum and minimum coordinates of the length/width of the pixel area of the article according to the pixel information of the article in the picture;
drawing a rectangular frame containing the pixel area of the article according to the maximum and minimum coordinates of the length/width of the pixel area of the article.
The method comprises the steps of determining the maximum and minimum abscissa corresponding to the length (direction) of a pixel area of an article and the maximum and minimum ordinate corresponding to the width (direction) of the pixel area of the article according to pixel information of the article in a picture, calculating a difference value according to the maximum and minimum abscissa corresponding to the length to obtain the length of the pixel area of the article, calculating a difference value according to the maximum and minimum ordinate corresponding to the width to obtain the width of the pixel area of the article, and drawing a rectangular frame containing the pixel area of the article based on the length and the width of the pixel area.
After drawing the rectangular frame of the pixel area containing the articles, determining the arrangement direction of each article in the picture according to the rectangular frame corresponding to each article, specifically, looking at the position distribution of the rectangular frame of each article, if the arrangement direction is the transverse distribution, it is indicated that the arrangement direction of the articles in the picture is the transverse arrangement, and at this time, the moving direction of the conveyor belt on the article sorting line also runs transversely. If the articles are longitudinally distributed, the arrangement direction of the articles in the picture is also longitudinally arranged, and at the moment, the moving direction of the conveyor belt on the article sorting line also longitudinally runs.
After determining the arrangement direction of the articles in the picture, determining two opposite sides perpendicular to the arrangement direction in the picture. And then, calculating the distance from the rectangular frame corresponding to each article to the two opposite sides, determining the shortest distance in each calculated distance, and determining the article as a welted article if the shortest distance is smaller than a preset constant. For better understanding, referring to fig. 4, there are three articles on the picture, the arrangement direction of the articles is longitudinal arrangement, after determining the rectangular frame corresponding to the article, two opposite sides perpendicular to the arrangement direction are determined on the picture, at this time, the two opposite sides are two sides of the upper side in the picture, the distance from the rectangular frame corresponding to each article to the two opposite sides is calculated, and the distances from the rectangular frame corresponding to the article to the two opposite sides are compared to determine the shortest distance among the two distances, for example, the distance from the article at the top to the upper and lower sides is determined, and the shortest distance is determined among the two distances, and then, if the shortest distance is smaller than a preset constant, the article is determined to be a welted article. If the shortest distance is larger than the preset constant, the article is located in the middle position of the picture instead of the welt area. The specific value of the preset threshold is set according to actual needs, and is not limited herein.
In this embodiment, after determining that an article is a welt article, the pixel region of the welt article may be removed from the picture to obtain a picture of the pixel region from which the welt article is removed, specifically, the pixel region may be removed from the result set, and at this time, the picture does not include the welt article, so that when a robot on an article sorting production line grabs the article, the situation that the same article is grabbed many times and many times of grabbing is performed due to the welt condition does not occur, the waste of resources is avoided, and the phenomenon of multiple grabbing (empty grabbing) caused by the change of the grabbing center when the article is welt is solved.
It should be noted that the method can quickly judge the welting condition of the article, and is suitable for the condition that the length of the article is small relative to the visual field range in the moving direction of the conveyor belt, or the condition that the photographing speed (converted into the moving visual field direction of the conveyor belt) is high relative to the speed of the conveyor belt.
Further, a third embodiment of the present invention is proposed based on the first or second embodiment of the picture processing method.
In this embodiment, referring to fig. 5, before step S20, the method further includes:
step S40, extracting the outline data of the article from the pixel data of the previous group of deep learning prediction results to press the outline data into a first set, wherein the first set is used for comparing with a second set;
the step S20 further includes:
step S26, traversing the pixel area of the article according to the pixel information of the article in the picture to obtain the pixel coordinate of the corresponding outline of the welt pixel area of the article;
and S27, pressing the pixel coordinates of the corresponding outline of the article into a second set so as to determine whether the article is a welt article or not according to the pixel coordinates of the outline.
In this embodiment, before determining a pixel region of an article in a picture according to pixel information of the article, and determining whether the article is a welt article, profile data of the article may be extracted from pixel data of a previous group of deep learning prediction results, so as to be pushed into a first set, where the first set is used for comparison with a second set.
Then, according to the pixel information of the article in the picture, traversing the pixel area of the article to obtain the pixel coordinates of the corresponding contour of the welt pixel area of the article, in a specific manner consistent with the second embodiment, after obtaining the pixel coordinates of the corresponding contour of the welt pixel area of the article, pressing the pixel coordinates of the corresponding contour of the article into the second set to determine whether the article is a welt article according to the pixel coordinates of the contour. Specifically, the step of determining whether the article is a welt article according to the pixel coordinates of the outline includes:
traversing the first set with the second set as a primary set;
and if the contour similarity of the articles in the second set reaches a preset threshold value after the articles in the first set pass through coordinate translation, determining that the articles are welt articles.
That is, the second set is used as a main set, the first set is traversed, when the edge similarity between an article C in the first set and an article D in the second set after coordinate translation reaches the preset threshold (for example, 90%), that is, C and D are considered to be the same article, at this time, it is described that the article in the picture is a welt article, the pixel region of the welt article is removed from the picture, and a picture of the pixel region from which the welt article is removed is obtained, specifically, the pixel region can be removed from the result set, at this time, the picture does not contain the welt article, so that when a robot on an article sorting production line grabs the article, the situation that the same article is grabbed for many times due to the welt condition and multiple grabs empty times does not occur, the waste of resources is avoided, and the phenomenon of multiple grabs (empty grabbing) caused by the change of the grabbing center when the article is welt is solved.
In this embodiment, the recognition by the pixel-level edge matching is applied to a case where the length of the article is large relative to the field of view in the moving direction of the conveyor belt or a case where the photographing speed (converted into the moving field of view direction of the conveyor belt) is slow relative to the speed of the conveyor belt. According to the method, pixel-level data are matched, the consumed time is slightly increased, and the pixel area of the welt article is removed by comparing the matching result with the preset threshold value.
The invention also provides a picture processing device.
Referring to fig. 6, fig. 6 is a functional module schematic diagram of a first embodiment of a picture processing apparatus according to the present invention, where the picture processing apparatus is applied to a picture processing device.
The picture processing device includes:
an obtaining module 10, configured to obtain a picture;
the prediction module 20 is configured to predict the picture by using a deep learning algorithm to obtain pixel information of an article in the picture;
the judging module 30 is configured to determine a pixel area of an article according to pixel information of the article in the picture, and judge whether the article is a welt article;
a removing module 40, configured to remove a pixel area of the welt item from the picture if the pixel area is the same as the pixel area of the welt item.
Further, the judging module comprises:
the drawing unit is used for determining a pixel area of the article according to the pixel information of the article in the picture and drawing a rectangular frame containing the article;
the determining unit is used for determining the arrangement direction of each article in the picture according to the rectangular frame corresponding to each article;
the determining unit is further configured to determine two opposite sides perpendicular to the arrangement direction in the picture;
the calculation unit is used for calculating the distance from the rectangular frame corresponding to each article to the two opposite sides and determining the shortest distance;
the determining unit is further configured to determine that the article is a welt article if the shortest distance is smaller than a preset constant.
Further, the drawing unit is further configured to determine a maximum and minimum coordinate of a length/width of a pixel area of the article according to the pixel information of the article in the picture;
drawing a rectangular frame containing the pixel area of the article according to the maximum and minimum coordinates of the length/width of the pixel area of the article.
Further, the apparatus further comprises:
the extraction and pushing module is used for extracting the outline data of the article from the pixel data of the front group of deep learning prediction results so as to push the outline data into a first set, wherein the first set is used for comparing with a second set;
the judging module further comprises:
the traversing unit is used for traversing the pixel area of the article according to the pixel information of the article in the picture so as to obtain the pixel coordinate of the corresponding outline of the welt pixel area of the article;
and the pressing-in unit is used for pressing the pixel coordinates of the corresponding outline of the article into the second set so as to determine whether the article is a welt article or not according to the pixel coordinates of the outline.
Further, the pushing unit is further configured to traverse the first set with the second set as a main set;
and if the contour similarity of the articles in the second set reaches a preset threshold after the articles in the first set are subjected to coordinate translation, determining that the articles are welt articles.
The specific implementation of the image processing apparatus of the present invention is substantially the same as the embodiments of the image processing method described above, and will not be described herein again.
In addition, the embodiment of the invention also provides a computer readable storage medium.
The computer readable storage medium has stored thereon a picture processing program which, when executed by a processor, implements the steps of the picture processing method as described above.
The specific implementation of the computer-readable storage medium of the present invention is substantially the same as the embodiments of the image processing method described above, and is not described herein again.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one of 8230, and" comprising 8230does not exclude the presence of additional like elements in a process, method, article, or apparatus comprising the element.
The above-mentioned serial numbers of the embodiments of the present invention are merely for description and do not represent the merits of the embodiments.
Through the above description of the embodiments, those skilled in the art will clearly understand that the method of the above embodiments can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but in many cases, the former is a better implementation manner. Based on such understanding, the technical solutions of the present invention or portions thereof contributing to the prior art may be embodied in the form of a software product, which is stored in a storage medium (such as ROM/RAM, magnetic disk, optical disk) and includes instructions for enabling a terminal device (such as a mobile phone, a computer, a server, an air conditioner, or a network device) to execute the method according to the embodiments of the present invention.
The above description is only a preferred embodiment of the present invention, and not intended to limit the scope of the present invention, and all modifications of equivalent structures and equivalent processes, which are made by using the contents of the present specification and the accompanying drawings, or directly or indirectly applied to other related technical fields, are included in the scope of the present invention.
Claims (8)
1. An image processing method, characterized in that the image processing method comprises the following steps:
acquiring a picture, and predicting the picture by adopting a deep learning algorithm to obtain pixel information of an article in the picture;
determining a pixel area of an article according to pixel information of the article in the picture, and judging whether the article is a welted article;
if so, removing the pixel area of the welt article in the picture;
before the step of determining the pixel area of the article according to the pixel information of the article in the picture and judging whether the article is a welted article, the method comprises the following steps:
extracting outline data of the article from the pixel data of the front group of deep learning prediction results to press the outline data into a first set, wherein the first set is used for comparing with a second set;
the step of determining the pixel area of the article according to the pixel information of the article in the picture and judging whether the article is a welted article comprises the following steps:
traversing the pixel area of the article according to the pixel information of the article in the picture to obtain the pixel coordinate of the corresponding outline of the welting pixel area of the article;
pressing the pixel coordinates of the corresponding outline of the article into the second set to determine whether the article is a welt article according to the pixel coordinates of the outline;
the step of determining whether the article is a welt article according to the pixel coordinates of the outline comprises:
traversing the first set with the second set as a primary set;
and if the contour similarity of the articles in the second set reaches a preset threshold after the articles in the first set are subjected to coordinate translation, determining that the articles are welt articles.
2. The picture processing method according to claim 1, wherein the step of determining a pixel area of an article in the picture based on pixel information of the article and determining whether the article is a welt article comprises:
determining a pixel area of the article according to the pixel information of the article in the picture, and drawing a rectangular frame containing the article;
determining the arrangement direction of each article in the picture according to the rectangular frame corresponding to each article;
determining two opposite sides vertical to the arrangement direction in the picture;
calculating the distance from the rectangular frame corresponding to each article to two opposite sides, and determining the shortest distance;
and if the shortest distance is smaller than a preset constant, determining that the article is a welted article.
3. The picture processing method according to claim 2, wherein the step of determining a pixel region of the article based on pixel information of the article in the picture and drawing a rectangular frame containing the article comprises:
determining the maximum and minimum coordinates of the length/width of the pixel area of the article according to the pixel information of the article in the picture;
drawing a rectangular frame containing the pixel area of the article according to the maximum and minimum coordinates of the length/width of the pixel area of the article.
4. A picture processing apparatus, characterized in that the picture processing apparatus comprises:
the acquisition module is used for acquiring pictures;
the prediction module is used for predicting the picture by adopting a deep learning algorithm so as to obtain pixel information of articles in the picture;
the judging module is used for determining a pixel area of an article according to pixel information of the article in the picture and judging whether the article is a welted article;
a removing module, configured to remove a pixel region of the welt item from the picture if the pixel region is the same as the pixel region;
the device further comprises:
the extraction and pushing module is used for extracting the outline data of the article from the pixel data of the front group of deep learning prediction results so as to push the outline data into a first set, wherein the first set is used for comparing with a second set;
the judging module further comprises:
the traversing unit is used for traversing the pixel area of the article according to the pixel information of the article in the picture so as to obtain the pixel coordinate of the corresponding outline of the welt pixel area of the article;
a pressing-in unit, configured to press pixel coordinates of a corresponding outline of the article into the second set, so as to determine whether the article is a welt article according to the pixel coordinates of the outline;
the pushing unit is further used for traversing the first set by taking the second set as a main set;
and if the contour similarity of the articles in the second set reaches a preset threshold after the articles in the first set are subjected to coordinate translation, determining that the articles are welt articles.
5. The image processing apparatus of claim 4, wherein the determining module comprises:
the drawing unit is used for determining a pixel area of the article according to the pixel information of the article in the picture and drawing a rectangular frame containing the article;
the determining unit is used for determining the arrangement direction of each article in the picture according to the rectangular frame corresponding to each article;
the determining unit is further configured to determine two opposite sides perpendicular to the arrangement direction in the picture;
the calculation unit is used for calculating the distance from the rectangular frame corresponding to each article to the two opposite sides and determining the shortest distance;
the determining unit is further configured to determine that the article is a welt article if the shortest distance is smaller than a preset constant.
6. The picture processing apparatus according to claim 5, wherein said drawing unit is further configured to determine a maximum minimum coordinate of a length/width of a pixel area of an item based on pixel information of the item in the picture;
drawing a rectangular frame containing the pixel area of the article according to the maximum and minimum coordinates of the length/width of the pixel area of the article.
7. A picture processing device, characterized in that the picture processing device comprises a memory, a processor and a picture processing program stored on the memory and executable on the processor, which when executed by the processor implements the steps of the picture processing method according to any one of claims 1 to 3.
8. A computer-readable storage medium, on which a picture processing program is stored, which, when executed by a processor, implements the steps of the picture processing method according to any one of claims 1 to 3.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910648054.1A CN110443240B (en) | 2019-07-18 | 2019-07-18 | Picture processing method, device and equipment and computer readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910648054.1A CN110443240B (en) | 2019-07-18 | 2019-07-18 | Picture processing method, device and equipment and computer readable storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110443240A CN110443240A (en) | 2019-11-12 |
| CN110443240B true CN110443240B (en) | 2022-11-15 |
Family
ID=68430691
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910648054.1A Active CN110443240B (en) | 2019-07-18 | 2019-07-18 | Picture processing method, device and equipment and computer readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110443240B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2009200884A (en) * | 2008-02-22 | 2009-09-03 | Seiko Epson Corp | Image data processor and image data processing method |
| CN108230343A (en) * | 2018-01-05 | 2018-06-29 | 厦门华联电子股份有限公司 | A kind of image processing method and device |
| CN108830188A (en) * | 2018-05-30 | 2018-11-16 | 西安理工大学 | Vehicle checking method based on deep learning |
| CN108876795A (en) * | 2018-06-07 | 2018-11-23 | 四川斐讯信息技术有限公司 | A kind of dividing method and system of objects in images |
-
2019
- 2019-07-18 CN CN201910648054.1A patent/CN110443240B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2009200884A (en) * | 2008-02-22 | 2009-09-03 | Seiko Epson Corp | Image data processor and image data processing method |
| CN108230343A (en) * | 2018-01-05 | 2018-06-29 | 厦门华联电子股份有限公司 | A kind of image processing method and device |
| CN108830188A (en) * | 2018-05-30 | 2018-11-16 | 西安理工大学 | Vehicle checking method based on deep learning |
| CN108876795A (en) * | 2018-06-07 | 2018-11-23 | 四川斐讯信息技术有限公司 | A kind of dividing method and system of objects in images |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110443240A (en) | 2019-11-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111784685B (en) | Power transmission line defect image identification method based on cloud edge cooperative detection | |
| US10896349B2 (en) | Text detection method and apparatus, and storage medium | |
| KR102407664B1 (en) | Method, apparatus and apparatus for object detection | |
| Plastiras et al. | Efficient convnet-based object detection for unmanned aerial vehicles by selective tile processing | |
| US9449222B2 (en) | Image processing device and method | |
| CN112101317B (en) | Page direction identification method, device, equipment and computer readable storage medium | |
| CN106683073B (en) | License plate detection method, camera and server | |
| CN108427932B (en) | Fingerprint image identification method and device | |
| CN109376783B (en) | Robot adaptive terrain classification method, system and computer readable storage medium | |
| AU2018207032A1 (en) | Logo detection video analytics | |
| CN111767915A (en) | License plate detection method, device, equipment and storage medium | |
| CN110633712A (en) | Method, system and device for recognizing vehicle body color and computer readable medium | |
| CN110427845B (en) | Method, device and equipment for determining pixel center of article and readable storage medium | |
| CN111080665A (en) | Image frame identification method, device and equipment and computer storage medium | |
| CN110443240B (en) | Picture processing method, device and equipment and computer readable storage medium | |
| US9984294B2 (en) | Image classification method and apparatus for preset tour camera | |
| CN108647570B (en) | Zebra crossing detection method and device and computer readable storage medium | |
| CN114581419A (en) | Transformer insulating sleeve defect detection method, related equipment and readable storage medium | |
| CN108509876B (en) | Object detection method, device, apparatus, storage medium, and program for video | |
| CN110569921A (en) | Vehicle logo identification method, system, device and computer readable medium | |
| CN110458202B (en) | Picture processing method, device and equipment and computer readable storage medium | |
| CN111008987B (en) | Method and device for extracting edge image based on gray background and readable storage medium | |
| CN109308449B (en) | Foreign matter filtering video coding chip and method based on deep learning | |
| WO2023160061A1 (en) | Method and apparatus for determining moving object in image, electronic device, and storage medium | |
| CN118334320A (en) | Image object recognition processing method, device and equipment, storage medium and product |
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 | ||
| CB02 | Change of applicant information |
Address after: 523000 room 325, building 15, No. 1, Xuefu Road, Songshanhu Park, Dongguan City, Guangdong Province Applicant after: Guangdong Gongye Technology Co.,Ltd. Address before: Room 123-128, Building 1, No. 1, R&D Fifth Road, Songshanhu Park, Dongguan, Guangdong 523000 Applicant before: Dongguan Gongye Interconnection Technology Co.,Ltd. |
|
| CB02 | Change of applicant information | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |