WO2023246193A1 - 一种图像清晰度确定方法、装置、设备及存储介质 - Google Patents

一种图像清晰度确定方法、装置、设备及存储介质 Download PDF

Info

Publication number
WO2023246193A1
WO2023246193A1 PCT/CN2023/081657 CN2023081657W WO2023246193A1 WO 2023246193 A1 WO2023246193 A1 WO 2023246193A1 CN 2023081657 W CN2023081657 W CN 2023081657W WO 2023246193 A1 WO2023246193 A1 WO 2023246193A1
Authority
WO
WIPO (PCT)
Prior art keywords
character
target
image
area
sub
Prior art date
Application number
PCT/CN2023/081657
Other languages
English (en)
French (fr)
Inventor
谷爱国
Original Assignee
北京京东振世信息技术有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 北京京东振世信息技术有限公司 filed Critical 北京京东振世信息技术有限公司
Publication of WO2023246193A1 publication Critical patent/WO2023246193A1/zh

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
    • G06V30/10—Character recognition
    • G06V30/12—Detection or correction of errors, e.g. by rescanning the pattern
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
    • G06V30/10—Character recognition
    • G06V30/12—Detection or correction of errors, e.g. by rescanning the pattern
    • G06V30/133—Evaluation of quality of the acquired characters
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
    • G06V30/40—Document-oriented image-based pattern recognition
    • G06V30/42—Document-oriented image-based pattern recognition based on the type of document
    • G06V30/424—Postal images, e.g. labels or addresses on parcels or postal envelopes
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30168—Image quality inspection

Definitions

  • Embodiments of the present application relate to the field of image processing technology, for example, to an image sharpness determination method, device, equipment and storage medium.
  • the logistics system can first determine the image clarity of the signing image, and then perform subsequent operations on the signing image with higher image clarity.
  • image clarity is mainly determined by classifying the signature image or calculating the gradient of the pixels in the signature image.
  • Embodiments of the present application provide an image definition determination method, device, equipment and storage medium, which solves the problem of low accuracy in determining image definition.
  • an image sharpness determination method which may include:
  • Obtain a target image containing the target character extract the character area where the target character is located from the target image, and divide the character area into at least two character sub-areas;
  • For each character sub-region determine the recognition accuracy of the target character in the character sub-region, and determine the regional clarity of the character sub-region based on the recognition accuracy;
  • the image sharpness of the target image is determined based on the area sharpness of the at least two character sub-regions.
  • an image sharpness determining device which may include:
  • the character area division module is configured to obtain a target image containing the target character, extract the character area where the target character is located from the target image, and divide the character area into at least two character sub-areas;
  • a regional clarity determination module configured to determine the recognition accuracy of the target character within the character sub-region for each character sub-region, and determine the regional clarity of the character sub-region based on the recognition accuracy
  • the image definition determining module is configured to determine the image definition of the target image according to the regional definition of the at least two character sub-regions.
  • an electronic device which may include:
  • a memory communicatively connected to at least one processor; wherein,
  • the memory stores a computer program that can be executed by at least one processor, and the computer program is executed by at least one processor, so that when executed by at least one processor, the image definition determination method provided by any embodiment of the present application is implemented.
  • a computer-readable storage medium with computer instructions stored thereon.
  • the computer instructions are used to enable the processor to implement the image definition determining method provided by any embodiment of the present application when executed.
  • Figure 1 is a flow chart of an image sharpness determination method provided according to an embodiment of the present application.
  • Figure 2a is a schematic diagram of a signature image in an image sharpness determination method provided according to an embodiment of the present application
  • Figure 2b is a schematic diagram of the signing area extracted from the signing image in an image definition determination method provided according to an embodiment of the present application
  • Figure 2c is a schematic diagram of the character area extracted from the signature area in an image sharpness determination method provided according to an embodiment of the present application
  • Figure 2d is a diagram of Figure 2c in an image sharpness determination method provided according to an embodiment of the present application. A schematic diagram of the division result of the character area given on the basis;
  • Figure 3 is a flow chart of another image sharpness determination method provided according to an embodiment of the present application.
  • Figure 4 is a flow chart of yet another image sharpness determination method provided according to an embodiment of the present application.
  • Figure 5 is a flow chart of an optional example in yet another image sharpness determination method provided according to an embodiment of the present application.
  • Figure 6 is a structural block diagram of an image definition determining device provided according to an embodiment of the present application.
  • FIG. 7 is a schematic structural diagram of an electronic device that implements the image sharpness determination method according to the embodiment of the present application.
  • the image clarity obtained by the above two methods is a global definition. It cannot accurately capture whether each local area in the signature image is clear, and it cannot accurately capture the signature characters in this local area ( That is, whether the characters on the logistics signature are clear, and whether the signature characters are clear is the main focus of image clarity. Therefore, it is possible that the whole world is clear but some local areas are blurry and/or some signature characters are Blurred situations, resulting in lower accuracy in determining image sharpness.
  • FIG. 1 is a flow chart of an image sharpness determination method provided in an embodiment of the present application.
  • This embodiment can be applied to the situation of determining image sharpness, and is particularly suitable to the situation of jointly determining image sharpness through the regional sharpness of multiple character sub-regions on the target image.
  • This method can be performed by the image definition determining device provided by the embodiment of the present application.
  • the device can be implemented in the form of software and/or hardware.
  • the device can be integrated on an electronic device.
  • the electronic device can be various user terminals or server.
  • the method in the embodiment of this application includes the following steps:
  • the target image may be an image obtained by photographing the target object and include a target character on the target object.
  • the target character may be a target character on the target object, for example, it may be text, punctuation marks, image symbols, numbers, etc. , not limited here.
  • Image clarity can be determined by whether each target character in the target image is clear, because when each target character is clear, the target image will naturally be clear.
  • the character area where the target character is located can be extracted from the target image first, and the image sharpness can be determined by focusing on the area sharpness of the character area to avoid interference with the image sharpness by those non-character areas. It should be noted that when the number of target characters is at least two, the above character area can be understood as an area including all target characters. In practical applications, optionally, since the target image obtained after shooting the target object may contain objects other than the target object, in order to ensure the accuracy of character area extraction, the target object can be extracted from the target image first.
  • the object area where the character is located is, for example, the object area is detected through a certain target detection algorithm, and then the character area is extracted from the object area.
  • the signing image as an example, see Figures 2a to 2c, which respectively represent the signing image, the signing area detected from the signing image based on yolov3, and the character area extracted from the signing area.
  • x1,y1,x2,y2 Respectively represent the abscissa coordinate of the upper-left vertex, the ordinate of the upper-left vertex, the abscissa of the lower-right vertex, and the ordinate of the lower-right vertex of the signing area.
  • the above coordinates All are obtained by taking the upper left vertex of I as the origin, the horizontal axis as the horizontal axis pointing to the right, and the vertical axis as the vertical axis pointing downward. This explanation also applies to the following example about coordinates.
  • the character area accounts for a large proportion of the entire target image. This is especially obvious in the case of , because the blur of a certain local sub-area in the character area hardly affects the regional clarity of the entire character area. Therefore, in order to improve the accuracy of determining the image definition, the character area can be divided into at least two character sub-regions, so that in conjunction with subsequent steps, the image definition can be determined based on the regional definition of each character sub-region.
  • the character area extracted from the signing area I(x1,y1,x2,y2) is represented by I(xmin,ymin,xmax,ymax), where,xmin,ymin , xmax and ymax respectively represent the abscissa coordinate of the upper left vertex of the character area, the ordinate coordinate of the upper left vertex, the abscissa coordinate of the lower right vertex and the ordinate coordinate of the lower right vertex.
  • the following steps are performed for each character sub-region: identifying the target character in the character sub-region, that is, determining what character the target character is, thereby obtaining the recognition accuracy of the target character.
  • the recognition accuracy can be expressed or determined by recognition confidence/recognition probability, because the recognition confidence/recognition probability can directly reflect the accuracy of target character recognition, which is determined from the character
  • the granularity improves the determination accuracy of recognition accuracy. From a common sense, since the clearer the target character is, the easier it is to be recognized, and the recognition accuracy is correspondingly higher. Therefore, the regional clarity of the character sub-region can be determined by the recognition accuracy.
  • the number of target characters in a certain character sub-region may be one, two or more.
  • the recognition accuracy of each target character can be determined separately, and then The regional accuracy is determined based on each recognition accuracy.
  • CRNN Convolutional Recurrent Neural Network
  • S130 Determine the image sharpness of the target image according to the regional sharpness of the at least two character sub-regions.
  • the image sharpness of the target image is determined by the sharpness of each area.
  • the image sharpness can be quantitatively expressed by numerical values, or qualitatively expressed by the two categories of clear and blurred. No limitation is made here.
  • the definition of each region can be greater than the preset definition threshold, or the number of regions of the character sub-region corresponding to the region definition greater than the preset definition threshold exceeds the preset number. When the threshold is set, the target image is considered clear, otherwise it is considered blurry.
  • the technical solution of the embodiment of the present application obtains a target image containing the target character and extracts the character area where the target character is located from the target image to avoid non-character areas from interfering with the determination process of image clarity, and then the character area is Divide into at least two character sub-regions to ensure the accuracy of determining image clarity by processing at least two character sub-regions respectively; for each character sub-region, determine the recognition accuracy of the target character in the character sub-region, and The regional clarity of the character sub-region is determined based on the recognition accuracy.
  • the regional clarity can directly reflect whether the target character is clear or blurry; after obtaining the regional clarity of each character sub-region, it can be based on at least two character sub-regions.
  • the area sharpness of the area determines the image sharpness of the target image.
  • the above technical solution determines the image clarity by determining the regional clarity determined by the recognition accuracy of the target character in each character sub-region, which can accurately reflect whether each character sub-region (ie, local area) in the target image is clear, and more It can accurately reflect whether the target characters in each character sub-area are clear, thereby avoiding the situation where the whole world is clear but some character sub-areas and/or some target characters in some character sub-areas are blurred, thus ensuring that It proves the accuracy of determining the image sharpness.
  • the above-mentioned image sharpness determination method may also include: classifying the character sub-regions, where the classification category includes a first category and a second category, and The category clarity corresponding to the first category is greater than the category clarity corresponding to the second category; the classification probability of whether the character sub-region belongs to the first category or the second category is obtained; then the regional clarity of the character sub-region is determined based on the recognition accuracy, It may include: determining the regional clarity of the character sub-region based on recognition accuracy and classification probability.
  • the regional clarity determined according to the recognition accuracy can be understood as a kind of local clarity, and considering that many character processing algorithms such as character positioning algorithms and character recognition algorithms are more used to process text, punctuation marks and For characters such as numbers, the ability to process image symbols is relatively average, which means that the unclear image symbols in the character sub-region may not be reflected in the regional clarity determined based on the recognition accuracy. Therefore, in order to improve the determination accuracy of regional definition, it can be achieved by fusing the global determination scheme and the local determination scheme.
  • the character sub-region is classified into the first category or the second category, where the clarity of the category corresponding to the first category is greater than that of the second category, which means that the first category can be understood as a clear category, and the second category Can be understood as a fuzzy category.
  • the classification probability of whether the character sub-region belongs to the first category or the second category can be obtained, and then based on the classification probability, it is determined whether the character sub-region is clear as a whole, so that the clarity of the image symbol can be determined. Reflected. It should be emphasized that overall clarity does not mean that each target character within it is clear. Therefore, the classification probability and recognition accuracy can be combined to determine the regional clarity from multiple angles to ensure that the region is clear. degree of determination accuracy.
  • H i,j (x) P i,j +AVG i,j .
  • H i,j (x) is greater than the preset definition threshold thr, it is considered that the i,jth
  • Figure 3 is a flow chart of another image sharpness determination method provided in the embodiment of the present application. This embodiment is refined based on the above technical solutions.
  • extracting the character area where the target character is located from the target image may include: determining the character position information of the target character in the target image; and determining, based on the character position information, the target character in the target image corresponding to the target character.
  • the regional position information of the character area in the target image extract the character area from the target image based on the regional position information.
  • the method in this embodiment may include the following steps:
  • S220 Determine the character position information of the target character in the target image, and determine the region position information of the character area in the target image corresponding to the target character in the target image based on the character position information.
  • the character position information can be the position information of the target character in the target image. In practical applications, optionally, it can be represented by relative position information (for example, in the upper left 1/3 of the target image, etc.) and coordinate information. , not limited here. After obtaining the character position information, the regional position information of the character area in the target image can be directly determined based on it.
  • a character detection algorithm is used to locate each signature character in the signature image, thereby obtaining their respective character position information.
  • DBNet Differentiable Binarization Net
  • DBNet Differentiable Binarization Net
  • X1, Y1, X2,Y2 DBNet(I(x1,y1,x2,y2))
  • n means that there are n characters in the signature image. Signing character.
  • the area position information can indicate where the character area is located in the target image, the character area can be accurately extracted from the target image based on the area position information.
  • S250 Determine the image definition of the target image according to the regional definition of the at least two character sub-regions.
  • the technical solution in the embodiment of the present application can obtain the regional position information of the character area in the target image by determining the character position information of the target character in the target image, so that it can be extracted from the target image based on the regional position information. character area, thereby ensuring effective extraction of the character area.
  • the number of characters of the target character includes at least two.
  • the regional position information of the character area corresponding to the target character in the target image is determined, which may include : Based on the character position information of all target characters, determine the frame position information of the minimum circumscribed rectangular frame containing all target characters, and use the frame position information as the regional position information of the character area corresponding to the target character in the target image.
  • the position information determines the frame position information of the smallest circumscribed rectangular frame that contains all the target characters, and then uses the frame position information as the area position information, thereby ensuring that the extracted character area meets the above requirements.
  • the above-mentioned image sharpness determination method may also include: determining the sub-region positions of the at least two character sub-regions in the target image according to the region position information and the number of regions of the character sub-regions. Information; for each target character, according to the character position of the target character information and the sub-region position information of the at least two character sub-regions in the target image to determine the character sub-region where the target character is located.
  • the sub-region position information of each character sub-region in the target image can be determined respectively. That is, for any character sub-region, the regional position information of the entire character region and the character sub-region can be determined.
  • the relative position of the region in all character sub-regions and obtain its sub-region position information. Furthermore, for a certain target character, the character sub-region where the target character is located can be determined based on the relative relationship between its character position information and the position information of each sub-region. For example, the sub-region position information containing the character position information can be compared with The corresponding character sub-area is used as the character sub-area where the target character is located.
  • the sub-region position information of the i, j-th character sub-region can be expressed as as well as These four formulas respectively represent the abscissa coordinate of the upper left vertex of the character sub-region, the ordinate coordinate of the upper left vertex, the abscissa coordinate of the lower right vertex, and the ordinate coordinate of the lower right vertex.
  • the character position information of the k-th signature character When the following formula is satisfied, the signing character is considered to be located in the character sub-area:
  • Figure 4 is a flow chart of yet another image sharpness determination method provided in the embodiment of the present application. This embodiment is refined based on the above technical solutions.
  • the target image is an image obtained after shooting the target object, and the target object contains the target characters.
  • the above-mentioned sharpness determination method may also include: based on the regional position information of the character area in the target image. , determine whether the target characters captured in the target image cover all target characters; determine the character area based on the judgment results of integrity. The explanations of terms that are the same as or corresponding to the above embodiments will not be repeated here.
  • the method in this embodiment may include the following steps:
  • the target image is an image obtained by photographing a target object containing the target character, and the photographed target object contains the target symbol, this indicates that the target character belongs to the object symbol.
  • the target image contains all the object symbols, that is, whether the target symbols cover all the object symbols, because all the object symbols may not be captured due to deviations in the shooting angle.
  • the target object can be understood as a logistics signature, and the object symbols can be understood as all symbols in the logistics signature.
  • Figure 2c shows an example in which the signature symbols in the signature image include all object symbols. .
  • S320 Determine the regional position information of the character area corresponding to the target character in the target image in the target image, determine whether the target characters photographed in the target image cover all the target characters based on the area position information, and determine the character area based on the judgment result. of integrity.
  • the area position information can indicate where the character area covering all the target characters is located in the target image, it can be judged based on the area position information whether the target characters in the target image cover all the target characters. For example, according to the area position information, The determination is based on the relative positional relationship between the character area and the target image determined by the area position information, or based on the relative positional relationship combined with the relative positional relationship between the target object and all target characters, etc., which are not limited here.
  • the integrity of the character area can be determined based on the judgment result, that is, when the target character covers all the target characters, the photographed character area is considered complete, otherwise it is considered incomplete.
  • the character area is a key area in the target image, the integrity of the target image can also be obtained through the integrity of the character area. If the character area is complete, the target image is complete, and vice versa. .
  • S350 Determine the image definition of the target image according to the regional definition of the at least two character sub-regions.
  • the technical solution of the embodiment of the present application determines whether the target characters photographed in the target image cover all the target characters through the regional position information of the character area in the target image, so that the integrity of the character area can be determined based on the judgment result ( That is, determining the integrity of the target image), which cooperates with the determination process of image sharpness, thereby achieving the effect of accurately determining the shooting quality of the target image.
  • the regional position information is represented by the vertex coordinates of the regional vertices on the character area; based on the regional position information of the character area in the target image, the target captured in the target image is judged Whether the characters cover all target characters may include: obtaining a preset first critical value, determining the size value of the target image, determining the second critical value based on the first critical value and the size value; and determining the second critical value based on the sum of the vertex coordinates and the first critical value. /or the numerical relationship between the second critical values to determine whether the target characters captured in the target image cover all target characters.
  • the first critical value may be a preset value used to represent the distance from a certain boundary in the target image
  • the second critical value may be a value determined based on the first critical value and the size value used to represent the distance from a certain boundary in the target image.
  • the other boundary and the above-mentioned boundary may be two boundaries that are parallel to each other. For example, assuming that the size value includes width (W) and height (H), and the first critical value is T, then the second critical value may include WT and HT. Whether the target characters cover all target characters can be determined based on the numerical relationship between the vertex coordinates and the first critical value and/or the second critical value.
  • the vertex coordinates of the area vertices on the character area i.e., the abscissa coordinate of the upper left vertex of the character area, the ordinate coordinate of the upper left vertex, the abscissa coordinate of the lower right vertex, and the vertical coordinate of the lower right vertex
  • xmin, ymin, xmax and ymax when the vertex coordinates satisfy ), xmax>WT (this indicates that the character area is very close to the right boundary of the target image) and ymax>HT (this indicates that the character area is very close to the lower boundary of the target image) (the character area at this time is very close to the lower boundary of the target image) (near at least one boundary of the target image), it is considered that the target character does not cover all the target characters, otherwise it is considered that all the target characters are covered.
  • dividing the character area into at least two character sub-areas may include: if the character area is complete, then dividing the character area into at least two character sub-areas.
  • the character area is incomplete (that is, the target image is incomplete)
  • the significance of determining the image definition is limited, because even if the image definition is high, the incomplete target image is still unapplicable. Therefore, it is possible to first determine whether the character area is complete, and if it is complete, then perform the step of dividing the character area into at least two character sub-areas to determine the image clarity.
  • the above image sharpness determination method may also include: if the character area is incomplete, prompt information may be generated and displayed.
  • the prompt information may include information related to the fact that not all the object characters were captured, please re-photograph. Information.
  • the character area into N*N character sub-areas, classify the character sub-area for each character sub-area, and obtain the classification probability belonging to a clear category; and separately identify each signature in the character sub-area characters to obtain their respective recognition accuracy. Furthermore, the classification probability and each recognition accuracy are fused to obtain the regional clarity of the character sub-region. Therefore, the image definition of the signature image is determined based on the regional definition of the N*N character sub-regions.
  • Figure 6 is a structural block diagram of an image definition determination device provided by an embodiment of the present application.
  • the device is configured to execute the image definition determination method provided by any of the above embodiments.
  • This device belongs to the same inventive concept as the image definition determination method in the above embodiments.
  • the device may include: a character area dividing module 410, an area definition determining module 420, and an image definition determining module 430. in,
  • the character area dividing module 410 is configured to obtain a target image containing the target character, extract the character area where the target character is located from the target image, and divide the character area into at least two character sub-areas;
  • the area definition determination module 420 is configured to determine, for each character sub-region, the recognition accuracy of the target character in the character sub-region, and determine the regional clarity of the character sub-region based on the recognition accuracy;
  • the image definition determining module 430 is configured to determine the image definition of the target image based on the area definition of the at least two character sub-regions.
  • the above image definition determining device may also include:
  • the character sub-region classification module is configured to classify the character sub-region after targeting each character sub-region, wherein the classification category includes a first category and a second category, and the category definition corresponding to the first category is greater than that of the third category. Category clarity corresponding to the second category;
  • the classification probability obtaining module is configured to obtain the classification probability that the character sub-region belongs to the first category or the second category;
  • the area clarity determination module 420 may include:
  • the area clarity determination unit is configured to determine the area clarity of the character sub-region based on recognition accuracy and classification probability.
  • the character area dividing module 410 may include:
  • a character position information determination unit configured to determine the character position information of the target character in the target image
  • a region position information determination unit configured to determine, based on the character position information, the region position information of the character region in the target image corresponding to the target character in the target image;
  • the character area extraction unit is configured to extract the character area from the target image according to the area position information.
  • the number of characters of the target character includes at least two, and the regional position information determination unit can be set to:
  • the frame position information of the minimum circumscribed rectangular frame containing all the target characters is determined based on the character position information of all target characters, and the frame position information is used as the area position information of the character area corresponding to the target character in the target image in the target image.
  • the above image definition determining device may also include:
  • a sub-region position information determination module configured to determine the sub-region position information of the at least two character sub-regions in the target image according to the region position information and the number of regions of the character sub-regions;
  • the target character positioning module is configured to, for each target character, determine the character sub-region where the target character is located based on the character position information of the target character and the sub-region position information of the at least two character sub-regions in the target image. .
  • the target image includes an image obtained by photographing the target object, and the target object includes the object character.
  • the above-mentioned image definition determining device may also include:
  • the target character determination module is configured to determine whether the target characters photographed in the target image cover all the target characters based on the regional position information of the character area in the target image;
  • the integrity determination module is configured to determine the integrity of the character area based on the judgment result.
  • the regional position information is represented by the vertex coordinates of the regional vertices on the character area;
  • the target character judgment module may include:
  • the second critical value determination unit is configured to obtain the preset first critical value, determine the size value of the target image, and determine the second critical value based on the first critical value and the size value;
  • the target character determination unit is configured to determine whether the target characters captured in the target image cover all target characters based on the numerical relationship between the vertex coordinates and the first critical value and/or the second critical value.
  • the character area dividing module 410 may include:
  • Character area division unit set to divide the character area into at least two character sub-areas if the character area is complete.
  • the image definition determination device obtaineds the target image containing the target character through the character area dividing module, and extracts the character area where the target character is located from the target image to avoid non-character areas affecting image clarity.
  • the determination process causes interference and then divides the character area into It is at least two character sub-regions, so as to ensure the determination accuracy of image clarity by processing the at least two character sub-regions respectively; for each character sub-region, the target character in the character sub-region is determined through the region clarity determination module
  • the recognition accuracy is determined, and the regional clarity of the character sub-region is determined based on the recognition accuracy.
  • the regional clarity can directly reflect whether the target character is clear or blurry; the image clarity determination module is used to obtain the regional clarity of each character sub-region.
  • the image sharpness of the target image may be determined based on the area sharpness of the at least two character sub-regions.
  • the above device determines the image clarity by determining the area clarity determined by the recognition accuracy of the target characters in the at least two character sub-areas, and can accurately reflect whether each character sub-area (ie, local area) in the target image is clear. , and can accurately reflect whether the target characters in each character sub-area are clear, thereby avoiding the situation where the whole world is clear but some character sub-areas and/or some target characters in some character sub-areas are blurred. This ensures the accuracy of determining image clarity.
  • the image definition determination device provided by the embodiments of this application can execute the image definition determination method provided by any embodiment of this application, and has corresponding functional modules for executing the method.
  • FIG. 7 shows a schematic structural diagram of an electronic device 10 that can be used to implement embodiments of the present application.
  • Electronic devices are intended to refer to various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers.
  • Electronic devices may also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smart phones, wearable devices (eg, helmets, glasses, watches, etc.), and other similar computing devices.
  • the components shown herein, their connections and relationships, and their functions are examples only and are not intended to limit the implementation of the present application as described and/or claimed herein.
  • the electronic device 10 includes at least one processor 11, and a memory communicatively connected to the at least one processor 11, such as a read-only memory (Read-Only Memory, ROM) 12, a random access memory Random Access Memory (RAM) 13, etc., in which the memory stores a computer program that can be executed by at least one processor.
  • the processor 11 can operate according to the computer program stored in the read-only memory (ROM) 12 or from the storage unit. 18 A computer program loaded into random access memory (RAM) 13 to perform various appropriate actions and processes.
  • RAM 13 random access memory
  • various programs and data required for the operation of the electronic device 10 can also be stored.
  • the processor 11, the ROM 12 and the RAM 13 are connected to each other via the bus 14.
  • An input/output (I/O) interface 15 is also connected to the bus 14 .
  • the I/O interface 15 Multiple components in the electronic device 10 are connected to the I/O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a magnetic disk, an optical disk, etc. etc.; and communication unit 19, such as network card, modem, wireless communication transceiver, etc.
  • the communication unit 19 allows the electronic device 10 to exchange information/data with other devices through computer networks such as the Internet and/or various telecommunications networks.
  • Processor 11 may be a variety of general and/or special purpose processing components having processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (Central Processing Unit, CPU), a graphics processing unit (Graphics Processing Unit, GPU), various dedicated artificial intelligence (Artificial Intelligence, AI) computing chips, various running Processors for machine learning model algorithms, digital signal processors (Digital Signal Processing, DSP), and any appropriate processors, controllers, microcontrollers, etc.
  • the processor 11 performs various methods and processes described above, such as the image sharpness determination method.
  • the image sharpness determination method may be implemented as a computer program, which is tangibly included in a computer-readable storage medium, such as the storage unit 18 .
  • part or all of the computer program may be loaded and/or installed onto the electronic device 10 via the ROM 12 and/or the communication unit 19.
  • the processor 11 may be configured to perform the image sharpness determination method in any other suitable manner (eg, by means of firmware).
  • Various implementations of the systems and techniques described above may be implemented in digital electronic circuit systems, Integrated circuit systems, Field Programmable Gate Array (FPGA), Application Specific Integrated Circuit (ASIC), Application Specific Standard Parts (ASSP), System on Chip (System on Chip) Chip, SOC), load programmable logic device (Complex Programmable Logic Device, CPLD), computer hardware, firmware, software, and/or their combination.
  • FPGA Field Programmable Gate Array
  • ASIC Application Specific Integrated Circuit
  • ASSP Application Specific Standard Parts
  • System on Chip System on Chip
  • SOC System on Chip
  • load programmable logic device Complex Programmable Logic Device, CPLD
  • computer hardware firmware, software, and/or their combination.
  • Various embodiments may include implementation in at least one computer program executable and/or interpreted on a programmable system including at least one programmable processor, which may be a special purpose or A general-purpose programmable processor can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.
  • a programmable processor which may be a special purpose or
  • a general-purpose programmable processor can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.
  • Computer programs for implementing the methods of the present application may be written in any combination of at least one programming language. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when executed by the processor, the computer program causes the functions/operations specified in the flowcharts and/or block diagrams to be implemented.
  • a computer program may execute entirely on the machine, partly on the machine, as a stand-alone software package, partly on the machine and partly on a remote machine or entirely on the remote machine or server.
  • a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in connection with an instruction execution system, apparatus, or device.
  • Computer-readable storage media may include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or devices, or any suitable combination of the foregoing.
  • the computer-readable storage medium may be a machine-readable signal medium.
  • machine-readable storage media would include an electrical connection based on at least one wire, a portable computer disk, a hard disk, random access memory (RAM), read only memory (ROM), erasable programmable read only memory (( Erasable Programmable Read-Only Memory (EPROM) or flash memory), fiber optics, portable compact disk read-only memory (Compact Disc Read-Only Memory (CD-ROM)), optical storage devices, magnetic storage devices, or any of the above Suitable combination.
  • RAM random access memory
  • ROM read only memory
  • EPROM Erasable Programmable Read-Only Memory
  • CD-ROM Compact Disc Read-Only Memory
  • CD-ROM Compact Disc Read-Only Memory
  • the systems and techniques described herein may be implemented on an electronic device having a display device (e.g., a cathode ray tube (CRT) or liquid crystal) for displaying information to the user.
  • a display device e.g., a cathode ray tube (CRT) or liquid crystal
  • a display Liquid Crystal Display, LCD monitor
  • a keyboard and pointing device e.g., a mouse or a trackball
  • Other kinds of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and may be provided in any form, including Acoustic input, voice input or tactile input) to receive input from the user.
  • the systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., A user's computer having a graphical user interface or web browser through which the user can interact with implementations of the systems and technologies described herein), or including such backend components, middleware components, or any combination of front-end components in a computing system.
  • the components of the system may be interconnected by any form or medium of digital data communication (eg, a communications network). Examples of communication networks include: Local Area Network (LAN), Wide Area Network (WAN), blockchain network, and the Internet.
  • Computing systems may include clients and servers.
  • Clients and servers are generally remote from each other and typically interact over a communications network.
  • the relationship of client and server is created by computer programs running on corresponding computers and having a client-server relationship with each other.
  • the server can be a cloud server, also known as cloud computing server or cloud host. It is a host product in the cloud computing service system to solve the problems that exist in traditional physical host and virtual private server (VPS) services. It has the disadvantages of difficult management and weak business scalability.
  • VPN virtual private server

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Multimedia (AREA)
  • Artificial Intelligence (AREA)
  • Quality & Reliability (AREA)
  • Character Input (AREA)
  • Image Analysis (AREA)

Abstract

本申请实施例公开了一种图像清晰度确定方法、装置、设备及存储介质。该方法包括：获取包含目标字符的目标图像，并从目标图像中提取出目标字符所在的字符区域，将字符区域划分为至少两个字符子区域；针对于每个字符子区域，确定字符子区域内目标字符的识别准确度，并根据识别准确度确定字符子区域的区域清晰度；根据至少两个字符子区域的区域清晰度确定目标图像的图像清晰度。

Description

一种图像清晰度确定方法、装置、设备及存储介质
本申请要求在2022年6月24日提交中国专利局、申请号为202210729982.2的中国专利申请的优先权，该申请的全部内容通过引用结合在本申请中。
技术领域
本申请实施例涉及图像处理技术领域，例如涉及一种图像清晰度确定方法、装置、设备及存储介质。
背景技术
在物流领域中，快递员需要对物流签单进行拍摄，然后将拍摄得到的签单图像上传到物流系统中以进行后续的相关操作。
为了保证后续操作的有效性，物流系统可以先确定签单图像的图像清晰度，然后再对图像清晰度较高的签单图像执行后续操作。目前，主要是通过对签单图像进行二分类或计算签单图像中像素点的梯度来确定图像清晰度。
在实现本申请的过程中，发明人发现相关技术中存在以下技术问题：图像清晰度的确定精度不高。
发明内容
本申请实施例提供了一种图像清晰度确定方法、装置、设备及存储介质，解决了图像清晰度的确定精度不高的问题。
根据本申请的一方面，提供了一种图像清晰度确定方法，可以包括：
获取包含目标字符的目标图像，从目标图像中提取出目标字符所在的字符区域，并将字符区域划分为至少两个字符子区域；
针对于每个字符子区域，确定字符子区域内目标字符的识别准确度，根据识别准确度确定字符子区域的区域清晰度；
根据所述至少两个字符子区域的区域清晰度确定目标图像的图像清晰度。
根据本申请的另一方面，提供了一种图像清晰度确定装置，可以包括：
字符区域划分模块，设置为获取包含目标字符的目标图像，并从目标图像中提取出目标字符所在的字符区域，将字符区域划分为至少两个字符子区域；
区域清晰度确定模块，设置为针对于每个字符子区域，确定字符子区域内的目标字符的识别准确度，并根据识别准确度确定字符子区域的区域清晰度；
图像清晰度确定模块，设置为根据所述至少两个字符子区域的区域清晰度确定目标图像的图像清晰度。
根据本申请的另一方面，提供了一种电子设备，可以包括：
至少一个处理器；以及
与至少一个处理器通信连接的存储器；其中，
存储器存储有可被至少一个处理器执行的计算机程序，计算机程序被至少一个处理器执行，以使至少一个处理器执行时实现本申请任意实施例所提供的图像清晰度确定方法。
根据本申请的另一方面，提供了一种计算机可读存储介质，其上存储有计算机指令，该计算机指令用于使处理器执行时实现本申请任意实施例所提供的图像清晰度确定方法。
附图说明
图1是根据本申请实施例提供的一种图像清晰度确定方法的流程图；
图2a是根据本申请实施例提供的一种图像清晰度确定方法中的签单图像的示意图；
图2b是根据本申请实施例提供的一种图像清晰度确定方法中从签单图像中提取出的签单区域的示意图；
图2c是根据本申请实施例提供的一种图像清晰度确定方法中从签单区域中提取出的字符区域的示意图；
图2d是根据本申请实施例提供的一种图像清晰度确定方法中的在图2c的 基础上给出的字符区域的划分结果的示意图；
图3是根据本申请实施例提供的另一种图像清晰度确定方法的流程图；
图4是根据本申请实施例提供的再一种图像清晰度确定方法的流程图；
图5是根据本申请实施例提供的再一种图像清晰度确定方法中的可选示例的流程图；
图6是根据本申请实施例提供的一种图像清晰度确定装置的结构框图；
图7是实现本申请实施例的图像清晰度确定方法的电子设备的结构示意图。
具体实施方式
为了使本技术领域的人员更好地理解本申请方案，下面将结合本申请实施例中的附图，对本申请实施例中的技术方案进行清楚、完整地描述。
需要说明的是，本申请的说明书和权利要求书及上述附图中的术语“第一”、“第二”等是用于区别类似的对象，而不必用于描述特定的顺序或先后次序。应该理解这样使用的数据在适当情况下可以互换，以便这里描述的本申请的实施例能够以除了在这里图示或描述的那些以外的顺序实施。“目标”、“原始”等的情况类似，在此不再赘述。此外，术语“包括”和“具有”以及他们的任何变形，意图在于覆盖不排他的包含，例如，包含了一系列步骤或单元的过程、方法、系统、产品或设备不必限于清楚地列出的那些步骤或单元，而是可包括没有清楚地列出的或对于这些过程、方法、产品或设备固有的其它步骤或单元。
在介绍本申请实施例之前，先对背景技术中阐述的图像清晰度的确定精度较低的具体原因进行分析，以便结合后文可以更好地理解本申请实施例提出的图像清晰度确定方案具有更高的确定精度的原因所在。继续以签单图像为例，上述两种方案得到的图像清晰度是一种全局清晰度，其无法准确捕获签单图像内的各个局部区域是否清晰，更是无法准确捕获该局部区域内的签单字符(即物流签单上的字符)是否清晰，而签单字符是否清晰正是图像清晰度最主要的关注点，因此这就有可能出现全局清晰而某些局部区域模糊和/或某些签单字符 模糊的情况，从而导致图像清晰度的确定精度较低。
图1是本申请实施例中提供的一种图像清晰度确定方法的流程图。本实施例可适用于确定图像清晰度的情况，尤其适用于通过目标图像上多个字符子区域的区域清晰度共同确定图像清晰度的情况。该方法可以由本申请实施例提供的图像清晰度确定装置来执行，该装置可以由软件和/或硬件的方式实现，该装置可以集成在电子设备上，该电子设备可以是各种用户终端或是服务器。
参见图1，本申请实施例的方法包括如下步骤：
S110、获取包含目标字符的目标图像，从目标图像中提取出目标字符所在的字符区域，并将字符区域划分为至少两个字符子区域。
其中，目标图像可以是对目标对象进行拍摄后得到的包含有目标对象上的目标字符的图像，该目标字符可以是目标对象上的对象字符，例如可以是文字、标点符号、图像符号以及数字等，在此未做限定。
图像清晰度可以通过目标图像内的各目标字符是否清晰来确定，因为当各目标字符都较为清晰时，目标图像也自然较为清晰。为此，可以先从目标图像中提取出目标字符所在的字符区域，从而通过聚焦于字符区域的区域清晰度来确定图像清晰度，以避免那些非字符区域对于图像清晰度的干扰。需要说明的，当目标字符的字符数量是至少两个时，上述字符区域可以理解为包含全部目标字符的区域。在实际应用中，可选的，由于对目标对象进行拍摄后得到的目标图像可能包含除目标对象之外的对象，因此为了保证字符区域提取的精准度，可以先从目标图像中提取出目标对象所在的对象区域，如通过某目标检测算法检测出对象区域，然后再从该对象区域中提取出字符区域。示例性的，继续以签单图像为例，参见图2a-图2c，它们分别表示签单图像、基于yolov3从签单图像上检测出的签单区域、以及在签单区域上提取出的字符区域。当通过I表示签单图像时，签单区域可以通过I(x1,y1,x2,y2)进行表示，即I(x1,y1,x2,y2)＝yolov3(I)，其中x1,y1,x2,y2分别表示签单区域的左上顶点的横坐标、左上顶点的纵坐标、右下顶点的横坐标及右下顶点的纵坐标。需要说明的是，上述坐标 均是以I的左上顶点为原点、横轴是指向右方的水平轴及纵轴是指向下方的垂直轴为例得到的。这一说明也适用于下述有关于坐标的举例。
在此基础上，如果直接通过字符区域的区域清晰度来确定目标图像的图像清晰度，这依然存在图像清晰度的确定精度不高的问题，其在字符区域在整张目标图像中占比较大的情况下尤为明显，这是因为字符区域中某局部子区域的模糊很难影响到整个字符区域的区域清晰度。因此，为了提高图像清晰度的确定精度，可以将字符区域划分为至少两个字符子区域，以便与后续步骤相配合，通过各字符子区域的区域清晰度来确定图像清晰度。示例性的，继续以上述示例为例，假设从签单区域I(x1,y1,x2,y2)中提取出的字符区域通过I(xmin,ymin,xmax,ymax)进行表示，其中，xmin,ymin,xmax,ymax分别表示字符区域的左上顶点的横坐标、左上顶点的纵坐标、右下顶点的横坐标及右下顶点的纵坐标。现将I(xmin,ymin,xmax,ymax)划分为成N×N的字符子区域，以N＝3为例，参见图2d，其中的每个网格分别表示一个字符子区域。
S120、针对于每个字符子区域，确定字符子区域内目标字符的识别准确度，并根据识别准确度确定字符子区域的区域清晰度。
其中，分别对每个字符子区域执行下述步骤：对字符子区域内的目标字符进行识别，即确定该目标字符是什么字符，从而得到该目标字符的识别准确度。在实际应用中，可选的，该识别准确度可以通过识别置信度/识别概率进行表示或确定，这是因为识别置信度/识别概率可以直接反映出目标字符识别的准确性，这是从字符颗粒度上提高了识别准确度的确定精度。常理来看，由于目标字符越清晰则越易被识别，其的识别准确度也相应越高，因此可以通过识别准确度确定该字符子区域的区域清晰度。需要说明的是，某字符子区域内的目标字符的字符数量可能是一个、两个或多个，当其内存在至少两个目标字符时，可以分别确定每个目标字符的识别准确度，然后根据各识别准确度确定区域准确度。
示例性的，继续以上述示例为例，假设通过表示第i,j个字符子区域(即第i行第j列的字符子区域)，其中分别表示该 字符子区域内的多个签单字符(即目标字符)的左上顶点的横坐标的集合、左上顶点的纵坐标的集合、右下顶点的横坐标的集合以及右下顶点的纵坐标的集合。将卷积循环神经网络(Convolutional Recurrent Neural Network，CRNN)作为字符识别模型，作用于上，得到多个签单字符的识别准确度的集合Mi,j，即然后计算Mi,j的均值AVGi,j，即AVGi,j＝mean(Mi,j)，并将由此得到的AVGi,j作为该字符子区域的区域清晰度。
S130、根据所述至少两个字符子区域的区域清晰度确定目标图像的图像清晰度。
其中，通过每个区域清晰度确定目标图像的图像清晰度，在实际应用中，可选的，该图像清晰度可以通过数值进行定量表示，或是通过清晰和模糊这两个类别进行定性表示，在此未做限定。示例性的，以定性表示为例，可以在每个区域清晰度均大于预设清晰度阈值、或是与大于预设清晰度阈值的区域清晰度对应的字符子区域的区域数量超过预设数量阈值时，认为目标图像清晰，否则认为其模糊。
本申请实施例的技术方案，通过获取包含目标字符的目标图像，并从目标图像中提取出目标字符所在的字符区域，以避免非字符区域对于图像清晰度的确定过程造成干扰，进而将字符区域划分为至少两个字符子区域，以通过分别处理至少两个字符子区域来保证图像清晰度的确定精准度；针对于每个字符子区域，确定字符子区域内目标字符的识别准确度，并根据识别准确度确定字符子区域的区域清晰度，该区域清晰度可以直接反映出其中的目标字符是清晰还是模糊；在得到每个字符子区域的区域清晰度后，可以根据至少两个字符子区域的区域清晰度确定目标图像的图像清晰度。上述技术方案，通过每个字符子区域内目标字符的识别准确度确定出的区域清晰度确定图像清晰度，可以准确反映出目标图像内的每个字符子区域(即局部区域)是否清晰，更是可以准确反映出每个字符子区域内的目标字符是否清晰，由此可以避免出现全局清晰但是某些字符子区域和/或某些字符子区域内的某些目标字符模糊的情况，从而保 证了图像清晰度的确定精度。
一种可选的技术方案，在针对于每个字符子区域后，上述图像清晰度确定方法，还可以包括：对字符子区域进行分类，其中，分类类别包括第一类别和第二类别，与第一类别对应的类别清晰度大于与第二类别对应的类别清晰度；得到字符子区域属于第一类别或是第二类别的分类概率；则根据识别准确度确定字符子区域的区域清晰度，可以包括：根据识别准确度和分类概率确定字符子区域的区域清晰度。
其中，根据识别准确度确定出的区域清晰度可以理解为一种局部清晰度，而且考虑到很多字符处理算法如字符定位算法和字符识别算法等，它们更多是用于处理文字、标点符号和数字等字符，对于图像符号的处理能力较为一般，这意味着字符子区域内图像符号的不清晰可能难以体现在根据识别准确度确定出的区域清晰度上。因此，为了提高区域清晰度的确定精度，可以通过融合全局确定方案和局部确定方案的方式实现。示例性的，将字符子区域分类为第一类别或是第二类别，其中，与第一类别对应的类别清晰度大于第二类别，这意味着第一类别可以理解为清晰类别，第二类别可以理解为模糊类别。在对字符子区域进行分类后，可以得到其属于第一类别或是第二类别的分类概率，然后根据该分类概率确定该字符子区域在整体上是否清晰，从而可将图像符号的清晰与否体现出来。需要强调的是，整体上清晰并不意味着其内的每个目标字符均清晰，因此可以将分类概率和识别准确度相结合，由此从多个角度共同确定区域清晰度，保证了区域清晰度的确定精度。
示例性的，继续以上述示例为例，将resnet18作为分类网络，对至少两个字符子区域进行分类，即P(x)＝resnet18(x)，其中，x＝I(xmin,ymin,xmax,ymax)；P(x)是一张大小为N×N的分类概率图，表示经过分类网络后，输出为清晰类别的概率，概率越大则表示越清晰；通过Pi,j表示第i,j个字符子区域的分类概率。在此基础上，对于第i,j个字符子区域，通过如下式子表示其的区域清晰度Hi,j(x)：Hi,j(x)＝Pi,j+AVGi,j。假设Hi,j(x)大于预设清晰度阈值thr时，认为该第i,j个 字符子区域清晰，并且在所有字符子区域均清晰时，认为签单图像清晰，即h(x)＝1,if Hi,j(x)＞thr,i,j＝1,2,…,N，其中h(x)＝1表示签单图像清晰。
图3是本申请实施例中提供的另一种图像清晰度确定方法的流程图。本实施例以上述各技术方案为基础进行细化。本实施例中，可选的，从目标图像中提取出目标字符所在的字符区域，可包括：确定目标字符在目标图像中的字符位置信息；根据字符位置信息，确定目标图像中与目标字符对应的字符区域在目标图像中的区域位置信息；根据区域位置信息从目标图像中提取出字符区域。其中，与上述各实施例相同或相应的术语的解释在此不再赘述。
参见图3，本实施例的方法可以包括如下步骤：
S210、获取包含目标字符的目标图像。
S220、确定目标字符在目标图像中的字符位置信息，并根据字符位置信息，确定目标图像中与目标字符对应的字符区域在目标图像中的区域位置信息。
其中，字符位置信息可以是目标字符在目标图像中的位置信息，实际应用中，可选的，其可以通过相对位置信息(例如在目标图像的左上1/3处等)和坐标信息等进行表示，在此未做限定。在得到字符位置信息后，可以根据其直接确定字符区域在目标图像中的区域位置信息。
示例性的，继续以上述示例为例，采用字符检测算法定位出签单图像中的各签单字符，由此得到它们各自的字符位置信息。这里将可微分的二值化网络(Differentiable Binarization Net，DBNet)作为字符检测算法，在签单区域I(x1,y1,x2,y2)中检测出各签单字符的字符位置信息，即X1,Y1,X2,Y2＝DBNet(I(x1,y1,x2,y2))，其中 分别表示各签单字符的左上顶点的横坐标的集合、左上顶点的纵坐标的集合、右下顶点的横坐标的集合以及右下顶点的纵坐标的集合，n表示在该签单图像中存在n个签单字符。
S230、根据区域位置信息从目标图像中提取出字符区域，并将字符区域划分为至少两个字符子区域。
其中，由于区域位置信息可以表示出字符区域位于目标图像中的哪个位置，因此根据该区域位置信息可以从目标图像中准确提取出字符区域。
S240、针对于每个字符子区域，确定字符子区域内目标字符的识别准确度，并根据识别准确度确定字符子区域的区域清晰度。
S250、根据所述至少两个字符子区域的区域清晰度确定目标图像的图像清晰度。
本申请实施例中的技术方案，通过确定出的目标字符在目标图像中的字符位置信息，可以得到字符区域在目标图像中的区域位置信息，从而可以根据该区域位置信息从目标图像中提取出字符区域，由此保证了字符区域的有效提取。
在此基础上，一种可选的技术方案，目标字符的字符数量包括至少两个，根据字符位置信息，确定目标图像中与目标字符对应的字符区域在目标图像中的区域位置信息，可包括：根据所有目标字符的字符位置信息确定包含全部目标字符的最小外接矩形框的框位置信息，并将框位置信息作为目标图像中与目标字符对应的字符区域在目标图像中的区域位置信息。其中，为了保证提取得到的字符区域可以涵盖住全部目标字符，而且不会涵盖住除目标字符之外的其余对象(这是为了避免其余对象对图像清晰度的确定过程造成干扰)，可以根据字符位置信息确定包含全部目标字符的最小外接矩形框的框位置信息，然后将该框位置信息作为区域位置信息，从而保证了提取出的字符区域满足上述要求。示例性的，继续以上述示例为例，上述最小外接矩形框可以通过I(xmin,ymin,xmax,ymax)进行表示，其中，xmin＝min(X1)，ymin＝min(Y1)，xmax＝max(X2)，ymax＝max(Y2)，即xmin,ymin,xmax,ymax分别表示最小外接矩形框的左上顶点的横坐标、左上顶点的纵坐标、右下顶点的横坐标以及右下顶点的纵坐标，将这些坐标作为框位置信息，即区域位置信息。
另一种可选的技术方案，上述图像清晰度确定方法，还可包括：根据区域位置信息以及字符子区域的区域数量，分别确定所述至少两个字符子区域在目标图像中的子区域位置信息；针对于每个目标字符，根据目标字符的字符位置 信息以及所述至少两个字符子区域在所述目标图像中的子区域位置信息，确定目标字符所在的字符子区域。其中，根据区域位置信息和区域数量，可以分别确定每个字符子区域在目标图像中的子区域位置信息，即针对于任一字符子区域，可以根据整个字符区域的区域位置信息以及该字符子区域在所有字符子区域中的相对位置，得到其的子区域位置信息。进而，对于某个目标字符而言，可以根据其的字符位置信息和各子区域位置信息间的相对关系确定该目标字符所在的字符子区域，如将与包含该字符位置信息的子区域位置信息对应的字符子区域，作为该目标字符所在的字符子区域。
示例性的，继续以上述示例为例，假设字符子区域数量是N2(即N*N)，则第i,j个字符子区域的子区域位置信息可以表示为 以及这四个式子分别表示该字符子区域的左上顶点的横坐标、左上顶点的纵坐标、右下顶点的横坐标及右下顶点的纵坐标。在此基础上，当第k个签单字符的字符位置信息满足如下式子时，则认为该签单字符位于该字符子区域内：

其中，是S220的示例中给出的各签单字符的坐标集合，在此不再赘述。这里可以通过 表示该字符子区域内各签单字符的坐标集合。
图4是本申请实施例中提供的再一种图像清晰度确定方法的流程图。本实施例以上述各技术方案为基础进行细化。在本实施例中，可选的，目标图像是对目标对象进行拍摄后得到的图像，目标对象包含对象字符，上述清晰度确定方法，还可以包括：根据字符区域在目标图像中的区域位置信息，判断被拍摄到目标图像中的目标字符是否涵盖全部对象字符；根据判断结果确定字符区域 的完整性。其中，与上述各实施例相同或相应的术语的解释在此不再赘述。
参见图4，本实施例的方法可以包括如下步骤：
S310、获取包含目标字符的目标图像，其中，目标图像是对目标对象进行拍摄后得到的图像，目标对象包含对象字符。
其中，由于目标图像是对包含对象字符的目标对象进行拍摄后得到的图像，而且拍摄到的目标对象中包含目标符号，这说明目标字符属于对象符号。但是，目标图像中是否包含全部的对象符号，即目标符号是否涵盖住全部的对象符号，这是待确定的，因为有可能因拍摄角度的偏差而导致未拍摄到全部的对象符号。示例性的，继续以上述示例为例，目标对象可以理解为物流签单，而对象符号可以理解为物流签单中的全部符号，图2c给出了签单图像中的签单符号包含了全部对象符号的示例。
S320、确定目标图像中的与目标字符对应的字符区域在目标图像中的区域位置信息，并根据区域位置信息判断被拍摄到目标图像中的目标字符是否涵盖全部对象字符，根据判断结果确定字符区域的完整性。
其中，由于区域位置信息可以表示出涵盖住全部目标字符的字符区域位于目标图像中的哪个位置，因此可以根据该区域位置信息判断目标图像中的这些目标字符是否涵盖住全部对象字符，如根据该区域位置信息确定出的字符区域与目标图像间的相对位置关系来判断、根据该相对位置关系再结合目标对象与全部对象字符间的相对位置关系来判断等，在此未做限定。示例性的，可以根据判断结果确定字符区域的完整性，即当目标字符涵盖住全部对象字符时，则认为拍摄到的字符区域是完整的，否则认为不完整。在实际应用中，可选的，由于字符区域是目标图像中的关键区域，因此也可以通过字符区域的完整性来得到目标图像的完整性，如字符区域完整则目标图像完整，反之则不完整。
S330、根据区域位置信息从目标图像中提取出字符区域，并将字符区域划分为至少两个字符子区域。
S340、针对于每个字符子区域，确定字符子区域内目标字符的识别准确度， 并根据识别准确度确定字符子区域的区域清晰度。
S350、根据所述至少两个字符子区域的区域清晰度确定目标图像的图像清晰度。
本申请实施例的技术方案，通过字符区域在目标图像中的区域位置信息来判断被拍摄到目标图像中的这些目标字符是否涵盖住全部对象字符，从而可以根据判断结果确定字符区域的完整性(即确定目标图像的完整性)，这与图像清晰度的确定过程相配合，由此达到了准确确定目标图像的拍摄质量的效果。
在此基础上，一种可选的技术方案，区域位置信息通过字符区域上的区域顶点的顶点坐标进行表示；根据字符区域在目标图像中的区域位置信息，判断被拍摄到目标图像中的目标字符是否涵盖全部对象字符，可以包括：获取预先设置的第一临界值，且确定目标图像的尺寸值，根据第一临界值和尺寸值确定第二临界值；根据顶点坐标与第一临界值和/或第二临界值间的数值关系，判断被拍摄到目标图像中的目标字符是否涵盖全部对象字符。
其中，第一临界值可以是预先设置的用于表示与目标图像中的某条边界间的相距距离的数值，第二临界值可以是根据第一临界值和尺寸值确定出的用于表示与目标图像中的另一条边界间的相距距离的数值，该另一条边界与上述的该条边界可以是相互平行的两条边界。示例性的，假设尺寸值包括宽(W)和高(H)，第一临界值是T，那么第二临界值可以包括W-T和H-T。可以根据顶点坐标与第一临界值和/或第二临界值间的数值关系，判断目标字符是否涵盖全部对象字符。示例性的，继续以上述示例为例，假设字符区域上的区域顶点的顶点坐标(即字符区域的左上顶点的横坐标、左上顶点的纵坐标、右下顶点的横坐标及右下顶点的纵坐标)通过xmin、ymin、xmax以及ymax进行表示，那么当顶点坐标满足xmin＜T(这说明字符区域非常靠近目标图像的左边界)、ymin＜T(这说明字符区域非常靠近目标图像的上边界)、xmax＞W-T(这说明字符区域非常靠近目标图像的右边界)及ymax＞H-T(这说明字符区域非常靠近目标图像的下边界)中的至少一个不等式时(此时的字符区域非常靠 近目标图像的至少一条边界)，则认为目标字符未涵盖住全部对象字符，否则认为涵盖住全部对象字符。
另一种可选的技术方案，将字符区域划分为至少两个字符子区域，可包括：如果字符区域完整，则将字符区域划分为至少两个字符子区域。其中，由于在字符区域不完整(即目标图像不完整)时，确定图像清晰度的意义是有限的，因为即使图像清晰度较高，不完整的目标图像依然是无法应用的。因此，可以先确定字符区域是否完整，并在完整的情况下，再执行将字符区域划分为至少两个字符子区域的步骤来实现图像清晰度的确定。在此基础上，可选的，上述图像清晰度确定方法，还可包括：如果字符区域不完整，可以生成并展示提示信息，该提示信息可以包括与未拍摄到全部对象字符，请重新拍摄相关的信息。
为了从整体上更好地理解上述各技术方案的具体实现过程，下面结合具体示例，对其进行示例性说明。示例性的，继续以上述示例为例，如图5所示，这是一个基于区域划分和字符粒度识别来判断签单图像是否完整以及是否清晰的完整实现过程的具体示例。获取对物流签单进行拍摄后得到的签单图像，从签单图像中检测出物流签单所在的签单区域，并从签单区域中检测出包含物流签单上的签单字符的字符区域。根据字符区域上的签单字符是否覆盖住物流签单上的全部签单字符确定签单图像的完整性，如果不完整，则结束，否则执行下述步骤。将字符区域划分为N*N个字符子区域，针对于每个字符子区域，对该字符子区域进行分类，得到属于清晰类别的分类概率；并且，分别识别该字符子区域内的每个签单字符，得到各自的识别准确度。进而，将分类概率和各识别准确度进行融合，得到该字符子区域的区域清晰度。从而，根据N*N个字符子区域的区域清晰度确定签单图像的图像清晰度。
图6为本申请实施例提供的图像清晰度确定装置的结构框图，该装置设置为执行上述任意实施例所提供的图像清晰度确定方法。该装置与上述各实施例的图像清晰度确定方法属于同一个发明构思，在图像清晰度确定装置的实施例中未详尽描述的细节内容，可以参考上述图像清晰度确定方法的实施例。参见 图6，该装置可以包括：字符区域划分模块410、区域清晰度确定模块420和图像清晰度确定模块430。其中，
字符区域划分模块410，设置为获取包含目标字符的目标图像，从目标图像中提取出目标字符所在的字符区域，并将字符区域划分为至少两个字符子区域；
区域清晰度确定模块420，设置为针对于每个字符子区域，确定字符子区域内目标字符的识别准确度，并根据识别准确度确定字符子区域的区域清晰度；
图像清晰度确定模块430，设置为根据所述至少两个字符子区域的区域清晰度确定目标图像的图像清晰度。
可选的，上述图像清晰度确定装置，还可以包括：
字符子区域分类模块，设置为在针对于每个字符子区域之后，对字符子区域进行分类，其中，分类类别包括第一类别和第二类别，与第一类别对应的类别清晰度大于与第二类别对应的类别清晰度；
分类概率得到模块，设置为得到字符子区域属于第一类别或第二类别的分类概率；
区域清晰度确定模块420，可以包括：
区域清晰度确定单元，设置为根据识别准确度和分类概率确定字符子区域的区域清晰度。
可选的，字符区域划分模块410，可以包括：
字符位置信息确定单元，设置为确定目标字符在目标图像中的字符位置信息；
区域位置信息确定单元，设置为根据字符位置信息，确定目标图像中与目标字符对应的字符区域在目标图像中的区域位置信息；
字符区域提取单元，设置为根据区域位置信息从目标图像中提取出字符区域。
在此基础上，可选的，目标字符的字符数量包括至少两个，区域位置信息确定单元，可以设置为：
根据所有目标字符的字符位置信息确定包含全部的目标字符的最小外接矩形框的框位置信息，并将框位置信息作为目标图像中与目标字符对应的字符区域在目标图像中的区域位置信息。
再可选的，上述图像清晰度确定装置，还可以包括：
子区域位置信息确定模块，设置为根据区域位置信息以及字符子区域的区域数量，分别确定所述至少两个字符子区域在目标图像中的子区域位置信息；
目标字符定位模块，设置为针对于每个目标字符，根据目标字符的字符位置信息以及所述至少两个字符子区域在所述目标图像中的子区域位置信息，确定目标字符所在的字符子区域。
可选的，目标图像包括对目标对象进行拍摄后得到的图像，目标对象包含对象字符，上述图像清晰度确定装置，还可以包括：
目标字符判断模块，设置为根据字符区域在目标图像中的区域位置信息判断被拍摄到目标图像中的目标字符是否涵盖全部对象字符；
完整性确定模块，设置为根据判断结果确定字符区域的完整性。
在此基础上，可选的，区域位置信息是通过字符区域上的区域顶点的顶点坐标进行表示；目标字符判断模块，可以包括：
第二临界值确定单元，设置为获取预先设置的第一临界值，并确定目标图像的尺寸值，根据第一临界值和尺寸值确定第二临界值；
目标字符判断单元，设置为根据顶点坐标与第一临界值和/或第二临界值间的数值关系，判断被拍摄到目标图像中的目标字符是否涵盖全部对象字符。
再可选的，字符区域划分模块410，可以包括：
字符区域划分单元，设置为如果字符区域完整，则将字符区域划分为至少两个字符子区域。
本申请实施例中提供的图像清晰度确定装置，通过字符区域划分模块获取包含目标字符的目标图像，并从目标图像中提取出目标字符所在的字符区域，以避免非字符区域对于图像清晰度的确定过程造成干扰，进而将字符区域划分 为至少两个字符子区域，以通过分别处理所述至少两个字符子区域来保证图像清晰度的确定精准度；通过区域清晰度确定模块针对每个字符子区域，确定字符子区域内目标字符的识别准确度，并根据识别准确度确定字符子区域的区域清晰度，该区域清晰度可以直接反映出其中的目标字符是清晰还是模糊；通过图像清晰度确定模块在得到每个字符子区域的区域清晰度后，可以根据所述至少两个字符子区域的区域清晰度确定目标图像的图像清晰度。上述装置，通过所述至少两个字符子区域内目标字符的识别准确度确定出的区域清晰度确定图像清晰度，可以准确反映出目标图像内的每个字符子区域(即局部区域)是否清晰，更是可以准确反映出每个字符子区域内的目标字符是否清晰，由此可以避免出现全局清晰但是某些字符子区域和/或某些字符子区域内的某些目标字符模糊的情况，从而保证了图像清晰度的确定精度。
本申请实施例所提供的图像清晰度确定装置可执行本申请任意实施例所提供的图像清晰度确定方法，具备执行方法相应的功能模块。
值得注意的是，上述图像清晰度确定装置的实施例中，所包括的各个单元和模块只是按照功能逻辑进行划分的，但并不局限于上述的划分，只要能够实现相应的功能即可；另外，各功能单元的具体名称也只是为了便于相互区分，并不用于限制本申请的保护范围。
图7示出了可以用来实施本申请的实施例的电子设备10的结构示意图。电子设备旨在表示各种形式的数字计算机，诸如，膝上型计算机、台式计算机、工作台、个人数字助理、服务器、刀片式服务器、大型计算机、和其它适合的计算机。电子设备还可以表示各种形式的移动装置，诸如，个人数字处理、蜂窝电话、智能电话、可穿戴设备(如头盔、眼镜、手表等)和其它类似的计算装置。本文所示的部件、它们的连接和关系、以及它们的功能仅仅作为示例，并且不意在限制本文中描述的和/或者要求的本申请的实现。
如图7所示，电子设备10包括至少一个处理器11，以及与至少一个处理器11通信连接的存储器，如只读存储器(Read-Only Memory，ROM)12、随机访 问存储器(Random Access Memory，RAM)13等，其中，存储器存储有可被至少一个处理器执行的计算机程序，处理器11可以根据存储在只读存储器(ROM)12中的计算机程序或从存储单元18加载到随机访问存储器(RAM)13中的计算机程序，来执行各种适当的动作和处理。在RAM 13中，还可存储电子设备10操作所需的各种程序和数据。处理器11、ROM 12以及RAM 13通过总线14彼此相连。输入/输出(Input/Output，I/O)接口15也连接至总线14。
电子设备10中的多个部件连接至I/O接口15，包括：输入单元16，例如键盘、鼠标等；输出单元17，例如各种类型的显示器、扬声器等；存储单元18，如磁盘、光盘等；以及通信单元19，例如网卡、调制解调器、无线通信收发机等。通信单元19允许电子设备10通过诸如因特网的计算机网络和/或各种电信网络与其他设备交换信息/数据。
处理器11可以是各种具有处理和计算能力的通用和/或专用处理组件。处理器11的一些示例包括但不限于中央处理单元(Central Processing Unit，CPU)、图形处理单元(Graphics Processing Unit，GPU)、各种专用的人工智能(Artificial Intelligence，AI)计算芯片、各种运行机器学习模型算法的处理器、数字信号处理器(Digital Signal Processing，DSP)、以及任何适当的处理器、控制器、微控制器等。处理器11执行上文所描述的各个方法和处理，例如图像清晰度确定方法。
在一些实施例中，图像清晰度确定方法可被实现为计算机程序，其被有形地包含于计算机可读存储介质，如存储单元18。在一些实施例中，计算机程序的部分或者全部可以经由ROM 12和/或通信单元19而被载入和/或安装到电子设备10上。当计算机程序加载到RAM 13并由处理器11执行时，可以执行上文描述的图像清晰度确定方法的至少一个步骤。备选地，在其他实施例中，处理器11可以通过其他任何适当的方式(例如，借助于固件)而被配置为执行图像清晰度确定方法。
本文中以上描述的系统和技术的各种实施方式可以在数字电子电路系统、 集成电路系统、场可编程门阵列(Field Programmable Gate Array，FPGA)、专用集成电路(Application Specific Integrated Circuit，ASIC)、专用标准产品(Application Specific Standard Parts，ASSP)、芯片上系统的系统(System on Chip，SOC)、负载可编程逻辑设备(Complex Programmable Logic Device，CPLD)、计算机硬件、固件、软件、和/或它们的组合中实现。这些各种实施方式可以包括：实施在至少一个计算机程序中，该至少一个计算机程序可在包括至少一个可编程处理器的可编程系统上执行和/或解释，该可编程处理器可以是专用或者通用可编程处理器，可以从存储系统、至少一个输入装置、以及至少一个输出装置接收数据和指令，并且将数据和指令传输至该存储系统、该至少一个输入装置、以及该至少一个输出装置。
用于实施本申请的方法的计算机程序可以采用至少一个编程语言的任何组合来编写。这些计算机程序可以提供给通用计算机、专用计算机或是其他可编程数据处理装置的处理器，使得计算机程序当由处理器执行时使流程图和/或框图中所规定的功能/操作被实施。计算机程序可以完全在机器上执行、部分地在机器上执行，作为独立软件包部分地在机器上执行并且部分地在远程机器上执行或完全在远程机器或服务器上执行。
在本申请的上下文中，计算机可读存储介质可以是有形的介质，其可以包含或存储以供指令执行系统、装置或设备使用或与指令执行系统、装置或设备结合地使用的计算机程序。计算机可读存储介质可以包括但不限于电子的、磁性的、光学的、电磁的、红外的、或半导体系统、装置或设备，或者上述内容的任何合适组合。备选地，计算机可读存储介质可以是机器可读信号介质。机器可读存储介质的更具体示例会包括基于至少一个线的电气连接、便携式计算机盘、硬盘、随机存取存储器(RAM)、只读存储器(ROM)、可擦除可编程只读存储器((Erasable Programmable Read-Only Memory，EPROM)或快闪存储器)、光纤、便捷式紧凑盘只读存储器(Compact Disc Read-Only Memory，CD-ROM)、光学储存设备、磁储存设备、或上述内容的任何合适组合。
为了提供与用户的交互，可以在电子设备上实施此处描述的系统和技术，该电子设备具有：用于向用户显示信息的显示装置(例如，阴极射线管(Cathode Ray Tube，CRT)或者液晶显示器(Liquid Crystal Display，LCD)监视器)；以及键盘和指向装置(例如，鼠标或者轨迹球)，用户可以通过该键盘和该指向装置来将输入提供给电子设备。其它种类的装置还可以用于提供与用户的交互；例如，提供给用户的反馈可以是任何形式的传感反馈(例如，视觉反馈、听觉反馈、或者触觉反馈)；并且可以用任何形式(包括声输入、语音输入或者、触觉输入)来接收来自用户的输入。
可以将此处描述的系统和技术实施在包括后台部件的计算系统(例如，作为数据服务器)、或者包括中间件部件的计算系统(例如，应用服务器)、或者包括前端部件的计算系统(例如，具有图形用户界面或者网络浏览器的用户计算机，用户可以通过该图形用户界面或者该网络浏览器来与此处描述的系统和技术的实施方式交互)、或者包括这种后台部件、中间件部件、或者前端部件的任何组合的计算系统中。可以通过任何形式或者介质的数字数据通信(例如，通信网络)来将系统的部件相互连接。通信网络的示例包括：局域网(Local Area Network，LAN)、广域网(Wide Area Network，WAN)、区块链网络和互联网。
计算系统可以包括客户端和服务器。客户端和服务器一般远离彼此并且通常通过通信网络进行交互。通过在相应的计算机上运行并且彼此具有客户端-服务器关系的计算机程序来产生客户端和服务器的关系。服务器可以是云服务器，又称为云计算服务器或云主机，是云计算服务体系中的一项主机产品，以解决了传统物理主机与虚拟专用服务器(Virtual Private Server，VPS)服务中，存在的管理难度大，业务扩展性弱的缺陷。
应该理解，可以使用上面所示的各种形式的流程，重新排序、增加或删除步骤。例如，本申请中记载的各步骤可以并行地执行也可以顺序地执行也可以不同的次序执行，只要能够实现本申请的技术方案所期望的结果，本文在此不进行限制。

Claims (11)

  1. 一种图像清晰度确定方法，包括：
    获取包含目标字符的目标图像，并从所述目标图像中提取出所述目标字符所在的字符区域，将所述字符区域划分为至少两个字符子区域；
    针对于每个所述字符子区域，确定所述字符子区域内所述目标字符的识别准确度，并根据所述识别准确度确定所述字符子区域的区域清晰度；
    根据所述至少两个字符子区域的区域清晰度确定所述目标图像的图像清晰度。
  2. 根据权利要求1所述的方法，在所述针对于每个所述字符子区域之后，还包括：
    对所述字符子区域进行分类，其中，分类类别包括第一类别和第二类别，与所述第一类别对应的类别清晰度大于与所述第二类别对应的类别清晰度；
    得到所述字符子区域属于所述第一类别或是所述第二类别的分类概率；
    所述根据所述识别准确度确定所述字符子区域的区域清晰度，包括：
    根据所述识别准确度和所述分类概率确定所述字符子区域的区域清晰度。
  3. 根据权利要求1所述的方法，其中，所述从所述目标图像中提取出所述目标字符所在的字符区域，包括：
    确定所述目标字符在所述目标图像中的字符位置信息；
    根据所述字符位置信息，确定所述目标图像中与所述目标字符对应的字符区域在所述目标图像中的区域位置信息；
    根据所述区域位置信息从所述目标图像中提取出所述字符区域。
  4. 根据权利要求3所述的方法，其中，所述目标字符的字符数量是至少两个，所述根据所述字符位置信息，确定所述目标图像中与所述目标字符对应的字符区域在所述目标图像中的区域位置信息，包括：
    根据所有目标字符的字符位置信息确定包含全部所述目标字符的最小外接矩形框的框位置信息，并将所述框位置信息作为所述目标图像中与所述目标字符对应的字符区域在所述目标图像中的区域位置信息。
  5. 根据权利要求3所述的方法，还包括：
    根据所述区域位置信息以及所述字符子区域的区域数量，分别确定所述至少两个字符子区域在所述目标图像中的子区域位置信息；
    针对于每个所述目标字符，根据所述目标字符的所述字符位置信息以及所述至少两个字符子区域在所述目标图像中的子区域位置信息，确定所述目标字符所在的所述字符子区域。
  6. 根据权利要求1所述的方法，其中，所述目标图像是对目标对象进行拍摄后得到的图像，所述目标对象包含对象字符，所述方法还包括：
    根据所述字符区域在所述目标图像中的区域位置信息，判断被拍摄到的所述目标图像中的所述目标字符是否涵盖全部所述对象字符；
    根据判断结果确定所述字符区域的完整性。
  7. 根据权利要求6所述的方法，其中，所述区域位置信息通过所述字符区域上的区域顶点的顶点坐标进行表示；
    所述根据所述字符区域在所述目标图像中的区域位置信息，判断被拍摄到的所述目标图像中的所述目标字符是否涵盖全部所述对象字符，包括：
    获取预先设置的第一临界值，并且确定所述目标图像的尺寸值，根据所述第一临界值和所述尺寸值确定第二临界值；
    根据所述顶点坐标与所述第一临界值的数值关系和所述顶点坐标与所述第二临界值间的数值关系中的至少之一，判断被拍摄到的所述目标图像中的所述目标字符是否涵盖全部所述对象字符。
  8. 根据权利要求6所述的方法，其中，所述将所述字符区域划分为至少两个字符子区域，包括：
    响应于所述字符区域完整，将所述字符区域划分为至少两个字符子区域。
  9. 一种图像清晰度确定装置，包括：
    字符区域划分模块，设置为获取包含目标字符的目标图像，从所述目标图像中提取出所述目标字符所在的字符区域，将所述字符区域划分为至少两个字 符子区域；
    区域清晰度确定模块，设置为针对于每个所述字符子区域，确定所述字符子区域内所述目标字符的识别准确度，根据所述识别准确度确定所述字符子区域的区域清晰度；
    图像清晰度确定模块，设置为根据所述至少两个字符子区域的区域清晰度确定所述目标图像的图像清晰度。
  10. 一种电子设备，包括：
    至少一个处理器；以及
    与所述至少一个处理器通信连接的存储器；其中，
    所述存储器存储有可被所述至少一个处理器执行的计算机程序，所述计算机程序被所述至少一个处理器执行，以使所述至少一个处理器执行如权利要求1-8中任一项所述的图像清晰度确定方法。
  11. 一种计算机可读存储介质，所述计算机可读存储介质存储有计算机指令，所述计算机指令用于使处理器执行时实现如权利要求1-8中任一所述的图像清晰度确定方法。
PCT/CN2023/081657 2022-06-24 2023-03-15 一种图像清晰度确定方法、装置、设备及存储介质 WO2023246193A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202210729982.2A CN117333421A (zh) 2022-06-24 2022-06-24 一种图像清晰度确定方法、装置、设备及存储介质
CN202210729982.2 2022-06-24

Publications (1)

Publication Number Publication Date
WO2023246193A1 true WO2023246193A1 (zh) 2023-12-28

Family

ID=89281663

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/081657 WO2023246193A1 (zh) 2022-06-24 2023-03-15 一种图像清晰度确定方法、装置、设备及存储介质

Country Status (2)

Country Link
CN (1) CN117333421A (zh)
WO (1) WO2023246193A1 (zh)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109766890A (zh) * 2013-06-03 2019-05-17 支付宝（中国）网络技术有限公司 信息识别方法、设备和系统
CN111680688A (zh) * 2020-06-10 2020-09-18 创新奇智(成都)科技有限公司 字符识别方法及装置、电子设备、存储介质
CN111899243A (zh) * 2020-07-28 2020-11-06 阳光保险集团股份有限公司 一种图像清晰度评价方法、设备和计算机可读存储介质
CN113313113A (zh) * 2021-06-11 2021-08-27 北京百度网讯科技有限公司 证件信息获取方法、装置、设备以及存储介质

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109766890A (zh) * 2013-06-03 2019-05-17 支付宝（中国）网络技术有限公司 信息识别方法、设备和系统
CN111680688A (zh) * 2020-06-10 2020-09-18 创新奇智(成都)科技有限公司 字符识别方法及装置、电子设备、存储介质
CN111899243A (zh) * 2020-07-28 2020-11-06 阳光保险集团股份有限公司 一种图像清晰度评价方法、设备和计算机可读存储介质
CN113313113A (zh) * 2021-06-11 2021-08-27 北京百度网讯科技有限公司 证件信息获取方法、装置、设备以及存储介质

Also Published As

Publication number Publication date
CN117333421A (zh) 2024-01-02

Similar Documents

Publication Publication Date Title
CN108446698B (zh) 在图像中检测文本的方法、装置、介质及电子设备
US11810319B2 (en) Image detection method, device, storage medium and computer program product
CN112949767B (zh) 样本图像增量、图像检测模型训练及图像检测方法
CN114550177A (zh) 图像处理的方法、文本识别方法及装置
CN112101386B (zh) 文本检测方法、装置、计算机设备和存储介质
US11995905B2 (en) Object recognition method and apparatus, and electronic device and storage medium
WO2021151319A1 (zh) 卡片边框检测方法、装置、设备及可读存储介质
WO2023147717A1 (zh) 文字检测方法、装置、电子设备和存储介质
CN115861400B (zh) 目标对象检测方法、训练方法、装置以及电子设备
CN113627439A (zh) 文本结构化处理方法、处理装置、电子设备以及存储介质
US20230196805A1 (en) Character detection method and apparatus , model training method and apparatus, device and storage medium
CN112581344A (zh) 一种图像处理方法、装置、计算机设备及存储介质
CN113239807A (zh) 训练票据识别模型和票据识别的方法和装置
CN110390295B (zh) 一种图像信息识别方法、装置及存储介质
CN111950345A (zh) 摄像头的识别方法、装置、电子设备和存储介质
US10896339B2 (en) Detecting magnetic ink character recognition codes
CN114494751A (zh) 证照信息识别方法、装置、设备及介质
US11881044B2 (en) Method and apparatus for processing image, device and storage medium
CN113837194A (zh) 图像处理方法、图像处理装置、电子设备以及存储介质
US20230048495A1 (en) Method and platform of generating document, electronic device and storage medium
CN112396060A (zh) 基于身份证分割模型的身份证识别方法及其相关设备
WO2023246193A1 (zh) 一种图像清晰度确定方法、装置、设备及存储介质
CN113435358B (zh) 用于训练模型的样本生成方法、装置、设备、程序产品
CN114663418A (zh) 图像处理的方法及装置、存储介质及电子设备
CN115004261A (zh) 文本行检测

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23825851

Country of ref document: EP

Kind code of ref document: A1