CN114596576A - An image processing method, device, electronic device and storage medium - Google Patents

Abstract

The present disclosure provides an image processing method, device, electronic device and storage medium, which relate to the technical field of artificial intelligence, and further relate to the technical field of computer vision and deep learning, so as to at least solve the technical problem of low target object recognition efficiency in the related art. The specific implementation scheme is: acquiring a target image, wherein the target image includes the object to be recognized; detecting the target image to obtain target pixel data, wherein the target pixel data is used to represent at least one pixel in the object to be recognized and the object to be recognized. The positional relationship between the vertex coordinates; the target image is corrected based on the target pixel data, and the correction result is obtained.

Description

An image processing method, device, electronic device and storage medium

technical field

The present disclosure relates to the technical field of artificial intelligence, and further relates to the technical field of computer vision and deep learning, and in particular, to an image processing method, apparatus, electronic device, and storage medium.

Background technique

Text recognition for express orders generally includes two parts: text detection and text recognition. However, in the process of actual transportation and screening, the express packages are placed at random, the shooting angle is not fixed, and the pictures taken may be forward, upside-down, inclined, distorted, etc. Direct text detection and recognition is more difficult, and recognition after manual correction will greatly increase labor and time costs. Therefore, using the prior art to detect and set the express order has low accuracy.

SUMMARY OF THE INVENTION

The present disclosure provides an image processing method, device, electronic device and storage medium, so as to at least solve the technical problem of low accuracy in detecting express objects in the related art.

According to an aspect of the present disclosure, an image processing method is provided, including: acquiring a target image, where the target image includes an object to be recognized; and detecting the target image to obtain target pixel data, wherein the target pixel data is used to represent the target image to be identified. Identify the positional relationship between at least one pixel in the object and the vertex coordinates of the object to be identified; correct the target image based on the target pixel data to obtain a correction result.

According to yet another aspect of the present disclosure, an image processing apparatus is provided, comprising: an acquisition module for acquiring a target image, wherein the target image includes an object to be recognized; and a detection module for detecting the target image to obtain target pixels The target pixel data is used to represent the positional relationship between at least one pixel in the object to be recognized and the vertex coordinates of the object to be recognized; the correction module is used to correct the target image based on the target pixel data to obtain a correction result.

According to yet another aspect of the present disclosure, there is provided an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor. The at least one processor executes, so that the at least one processor can execute the image processing method proposed by the present disclosure.

According to yet another aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to cause a computer to execute the image processing method proposed by the present disclosure.

According to another aspect of the present disclosure, a computer program product is provided, including a computer program, and the computer program is executed by a processor to execute the image processing method proposed by the present disclosure.

In the present disclosure, a target image of a target scene is first acquired, wherein the target image includes an object to be recognized; then the target image is detected to obtain target pixel data, wherein the target pixel data is used to represent at least one of the objects to be recognized The positional relationship between the pixel and the vertex coordinates of the object to be recognized; finally, the target image is corrected based on the target pixel data, and the correction result is obtained. The recognition efficiency of the target image is improved. It is easy to notice that the target pixel data can be used to represent the positional relationship between at least one pixel in the object to be recognized and the vertex coordinates of the object to be recognized, and then the target image is corrected based on the target pixel data, which can further improve recognition. The accuracy of the method can reduce the false detection, and then solve the technical problem of the low accuracy of the express object detection in the related art.

It should be understood that what is described in this section is not intended to identify key or critical features of embodiments of the disclosure, nor is it intended to limit the scope of the disclosure. Other features of the present disclosure will become readily understood from the following description.

Description of drawings

The accompanying drawings are used for better understanding of the present solution, and do not constitute a limitation to the present disclosure. in:

1 is a partial view of a courier slip according to an embodiment of the present disclosure;

2 is a block diagram of a hardware structure of a computer terminal (or mobile device) for implementing a data processing method according to an embodiment of the present disclosure;

3 is a flowchart of a data processing method according to the first embodiment of the present disclosure;

4a is a diagram of an irregularly placed express delivery note according to an embodiment of the present disclosure;

4b is a diagram of an irregularly placed express delivery note according to an embodiment of the present disclosure;

Fig. 5a is a frame detection diagram of an express delivery slip according to an embodiment of the present disclosure;

5b is a correction diagram of an outer frame of a courier slip according to an embodiment of the present disclosure;

5c is a flowchart of another data processing method according to the second embodiment of the present disclosure;

Figure 6a is a sample image of an embodiment of the present disclosure;

6b is a central Gaussian distribution area diagram of a sample image according to an embodiment of the present disclosure;

6c is a flowchart of another data processing method according to the third embodiment of the present disclosure;

FIG. 7 is a structural block diagram of a data processing apparatus according to an embodiment of the present disclosure.

Detailed ways

Exemplary embodiments of the present disclosure are described below with reference to the accompanying drawings, which include various details of the embodiments of the present disclosure to facilitate understanding and should be considered as exemplary only. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the present disclosure. Also, descriptions of well-known functions and constructions are omitted from the following description for clarity and conciseness.

It should be noted that the terms "first", "second" and the like in the description and claims of the present disclosure and the above drawings are used to distinguish similar objects, and are not necessarily used to describe a specific sequence or sequence. It is to be understood that the data so used may be interchanged under appropriate circumstances such that the embodiments of the disclosure described herein can be practiced in sequences other than those illustrated or described herein. Furthermore, the terms "comprising" and "having" and any variations thereof, are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those expressly listed Rather, those steps or units may include other steps or units not expressly listed or inherent to these processes, methods, products or devices.

With the development of e-commerce and transportation, the express delivery industry is booming. In 2021, the number of express delivery in my country will exceed 10 billion, which will play a very important role in promoting consumption and accelerating the economic cycle. There are dozens of common express companies on the market, and each company has various styles of express orders, and the information is complicated. There are hundreds of couriers that each person comes into contact with in the process of distributing and delivering couriers at courier sites every day. If you manually enter the order number and recipient information, update the logistics status of the logistics company's internal information management system, or update the logistics status of the logistics company's internal information management system. It takes a lot of manpower and time to distribute orders based on information such as addresses. In an environment that requires express delivery speed, there is a high chance of errors in manual operations, which may lead to complaints.

At present, the methods for improving target detection and recognition mainly include the following:

Method 1. Four-direction classification method. The four-direction classification model can output four orientations of the picture, up and down, left and right, and then rotate 90°, 180°, 270° according to the direction to correct the express order picture.

Method 2. Regression method, using the regression model to directly detect the four vertices of the main body of the express order.

Method 3. The segmentation method, based on the segmentation algorithm, outputs the position of the main area and the forward 1/2 of the text, and the position of the upper left corner of the upper left corner of the text. Combine the position of the main body area and the position of the 1/4 area to determine the coordinates of the vertex of the main body 4 and the vertex of the starting point.

There are some problems in the related art, which are as follows: Method 1, the four-direction classification method, it is difficult to classify pictures with affine transformation angles or pictures rotated about 45 degrees taken by the camera obliquely. When the classification is correct, the rotated text still has a certain angle of inclination, which affects the subsequent text detection and recognition accuracy; method 2, regression method, when there are many layouts and complex styles of express orders, the vertex position will be inaccurate; method 3. Segmentation method, Figure 1 is a small local image of the express order. In the scenario shown in Figure 1, the feature distribution is scattered, and the segmentation accuracy is easily affected by large-area image features such as barcodes. 1 The /4 area map is error-prone, which in turn leads to errors in the judgment of the vertex starting point.

According to an embodiment of the present disclosure, an image processing method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings may be executed in a computer system such as a set of computer-executable instructions, and although the flowchart is shown in the flowchart A logical order is shown, but in some cases steps shown or described may be performed in a different order than shown.

The method embodiments provided by the embodiments of the present disclosure may be executed in a mobile terminal, a computer terminal, or a similar electronic device. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, 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 processors, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are by way of example only, and are not intended to limit implementations of the disclosure described and/or claimed herein. FIG. 2 shows a block diagram of the hardware structure of a computer terminal (or mobile device) for implementing an image processing method.

As shown in FIG. 2 , the computer terminal 200 includes a computing unit 201, which can be generated according to a computer program stored in a read only memory (ROM) 202 or a computer program loaded from a storage unit 208 into a random access memory (RAM) 203 Various appropriate actions and processes are performed. In the RAM 203, various programs and data necessary for the operation of the computer terminal 200 can also be stored. The computing unit 201 , the ROM 202 , and the RAM 203 are connected to each other through a bus 204 . An input/output (I/O) interface 205 is also connected to the bus 204 .

Various components in the computer terminal 200 are connected to the I/O interface 205, including: an input unit 206, such as a keyboard, a mouse, etc.; an output unit 207, such as various types of displays, speakers, etc.; a storage unit 208, such as a magnetic disk, an optical disk, etc. etc.; and a communication unit 209, such as a network card, modem, wireless communication transceiver, and the like. The communication unit 209 allows the computer terminal 200 to exchange information/data with other devices through a computer network such as the Internet and/or various telecommunication networks.

Computing unit 201 may be various general-purpose and/or special-purpose processing components with processing and computing capabilities. Some examples of computing units 201 include, but are not limited to, central processing units (CPUs), graphics processing units (GPUs), various specialized artificial intelligence (AI) computing chips, various computing units that run machine learning model algorithms, digital signal processing processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 201 performs the image processing method described herein. For example, in some embodiments, the image processing method may be implemented as a computer software program tangibly embodied on a machine-readable medium, such as storage unit 208 . In some embodiments, part or all of the computer program may be loaded and/or installed on the computer terminal 200 via the ROM 202 and/or the communication unit 209 . When a computer program is loaded into RAM 203 and executed by computing unit 201, one or more steps of the image processing method described herein may be performed. Alternatively, in other embodiments, the computing unit 201 may be configured to perform the image processing method by any other suitable means (eg, by means of firmware).

Various implementations of the systems and techniques described herein may be implemented in digital electronic circuitry, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), system-on-chips System (SOC), Load Programmable Logic Device (CPLD), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include being implemented in one or more computer programs executable and/or interpretable on a programmable system including at least one programmable processor that The processor, which may be a special purpose or general-purpose programmable processor, may 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 an output device.

It should be noted here that, in some optional embodiments, the electronic device shown in FIG. 2 may include hardware elements (including circuits), software elements (including computer code stored on a computer-readable medium), or hardware elements A combination of both components and software components. It should be noted that Figure 2 is only one example of a specific embodiment, and is intended to illustrate the types of components that may be present in the electronic device described above.

Under the above operating environment, the present disclosure provides the image processing method shown in FIG. 3 , and the method can be executed by the computer terminal shown in FIG. 2 or a similar electronic device. FIG. 3 is a flowchart of an image processing method provided according to the first embodiment of the present disclosure. As shown in Figure 3, the method may include the following steps:

Step S301, acquiring a target image, wherein the target image includes an object to be recognized.

The above-mentioned target image may be an image of an express package containing an object to be identified, wherein the object to be identified may be an express document on the express package.

The above-mentioned object to be identified may be the express document on the express package. The express document may include the following information: express number, recipient information, and the like.

The above-mentioned objects to be identified may also be invoices, electronic cards, posters, documents, etc. in the image.

In an optional embodiment, the target image may be acquired by a photographing device, where the photographing device may be a mobile phone, a camera, or the like.

In another optional embodiment, in the process of actual transportation and screening, since the angle of the express package is randomly placed and the camera shooting angle is not fixed, the express package in the photographed express package image will appear positive. , inverted, tilted, or even distorted, etc. Figures 4a and 4b are images of an irregularly placed express parcel in the present disclosure. Text detection and recognition is difficult, and recognition after manual correction will greatly increase labor costs and time costs. In the present disclosure, after the express package image is acquired, the express document in the express package image may be corrected to obtain a positive express document, thereby improving the detection accuracy of the express document.

Step S302: Detect the target image to obtain target pixel data, wherein the target pixel data is used to represent the positional relationship between at least one pixel in the object to be recognized and the vertex coordinates of the object to be recognized.

The above-mentioned target pixel data may be the positional relationship between all pixels in the object to be recognized and the vertex coordinates of the object to be recognized. The target pixel data can also be the positional relationship between one of the pixels in the object to be recognized and the vertex coordinates of the object to be recognized; the target pixel data can also be the relationship between multiple pixels in the object to be recognized and the vertex coordinates of the object to be recognized. Positional relationship. Optionally, the target pixel data may be the difference between the horizontal and vertical coordinates of the pixels in the main frame of the express document and the four vertices of the main frame.

In an optional embodiment, the above-mentioned at least one pixel may be a pixel in the central area of the express document. Since the pixels in the central area contain more textual information, the pixels in the central area can be used to more accurately represent the pixel and the pixel in the central area. Positional relationship between vertex coordinates.

In another optional embodiment, a detection model may be used to detect the target image to obtain target pixel data. Optionally, the detection model can use multi-channel segmentation to output the difference between the horizontal and vertical coordinates of the pixels in the main frame and the four vertices, so as to calculate the coordinates of the four vertices of the main body and determine the starting point coordinates of the four vertices. The coordinates of the starting point can be determined according to the orientation of the text in the pixel. Optionally, the coordinates of the starting point in the 4 vertices can be the upper left corner under the premise that the text is positive.

The above detection model may be a convolutional neural network (Convolutional Neural Networks, also referred to as CNN), wherein the convolutional neural network may be a type of feedforward neural network that includes convolutional computation and has a deep structure.

The above-mentioned starting point may be determined by the direction of the text. In an optional embodiment, the vertex of the upper left corner when the text is in the forward direction may be regarded as the starting point.

The above-mentioned vertex coordinates are the coordinates of the points on the four corners of the express document.

In another optional embodiment, the target object may be detected in all directions from different angles based on pixel information, so as to obtain target pixel data corresponding to the target image. Further, since the target pixel data is the positional relationship between at least one pixel in the object to be recognized and the vertex coordinates of the object to be recognized, therefore, 4 vertex coordinates and the starting point of 4 vertex coordinates can be determined by the target pixel data, Further, the object to be recognized in the target image can be corrected according to the coordinates of the four vertices and the starting point of the coordinates of the four vertices, so that the object to be recognized is in the forward direction.

Step S303, correcting the target image based on the target pixel data to obtain a correction result.

In an optional embodiment, four vertices and starting points of the object to be recognized in the target image may be obtained by logical calculation according to the target pixel data.

In another optional embodiment, the object to be recognized in the target image can be corrected according to the four vertices and the starting point, so that the object to be recognized can be displayed in the forward direction, because the text information in the object to be recognized displayed in the forward direction Therefore, when detecting the text information in the object to be recognized, the detection accuracy of the text information can be improved.

In another optional embodiment, the object to be recognized in the target image may be corrected through affine transformation to obtain a correction result, wherein the affine transformation refers to performing a linear transformation in a vector space in geometry and Following a translation, transform into another vector space. The above steps can greatly reduce the difficulty of character detection and recognition of positive characters, and can significantly improve the accuracy of character recognition, and manual correction is not required.

In another optional embodiment, after the corrected target image is obtained, a software development kit (Software Development Kit, also referred to as SDK) for express order recognition can be integrated into a mobile phone, a gun, a high-speed camera, etc. In terms of hardware, the software development kit can automatically extract the tracking number and recipient information on the express order in real time, which can not only ensure high recognition accuracy, but also greatly reduce the workload of manual accounting.

Fig. 5a is a detection diagram of an outer frame of a courier slip in the present disclosure, and Fig. 5b is a correction diagram of an outer frame of a courier slip in the present disclosure. As shown in Figure 5a and Figure 5b, in the present disclosure, the target image is detected by the detection model to obtain target pixel data, which can quickly detect the four vertices of the express order area in various directions, such as forward, inverted, inclined, and twisted. , and the starting point and sequence of the vertex can be determined according to the direction of the text, wherein the starting point of the vertex can be the upper left corner under the premise that the text is positive. In order to correct the express order area to the text to be positive, and improve the accuracy of subsequent text detection and recognition.

According to the above-mentioned steps S301 to S303 of the present disclosure, a target image of the target scene is first obtained, wherein the target image includes the object to be recognized; then the target image is detected to obtain target pixel data, wherein the target pixel data is used to represent the to-be-recognized object. The positional relationship between at least one pixel in the object and the vertex coordinates of the object to be recognized; finally, the target image is corrected based on the target pixel data to obtain a correction result. The recognition efficiency of the target image is improved. It is easy to notice that the target pixel data can be used to represent the positional relationship between at least one pixel in the object to be recognized and the vertex coordinates of the object to be recognized, and then the target image is corrected based on the target pixel data, which can further improve recognition. The accuracy of the method can reduce the false detection, and then solve the technical problem of the low accuracy of the express object detection in the related art.

Fig. 5c is a flowchart of an image processing method according to the second embodiment of the present disclosure. As shown in Fig. 5c, the method includes the following steps:

Step S501, acquiring a target image, wherein the target image includes an object to be recognized.

Step S502: Detect the target image to obtain target pixel data, wherein the target pixel data is used to represent the positional relationship between at least one pixel in the object to be recognized and the vertex coordinates of the object to be recognized.

Step S503, correcting the target image based on the target pixel data to obtain a correction result.

Step S504: Identify the target image based on the correction result to obtain a recognition result, wherein the recognition result is used to represent the text information of the object to be recognized in the target image.

Optionally, the target image is recognized based on the correction result to obtain a recognition result, wherein the recognition result is used to represent text information of the object to be recognized in the target image.

In an optional embodiment, the forward image of the object to be recognized in the target image can be determined according to the correction result, and the text information recorded in the object to be recognized can be obtained by recognizing the forward image of the object to be recognized, for example , recipient information, express order number, etc., so that the target image can be accurately identified, and then the text information of the object to be identified with high accuracy can be obtained.

When correcting the target image, it can be detected according to the coordinates of the four vertices of the detection subject and the coordinate information of the starting point. Through affine transformation, the main area of the express order picture in each direction can be corrected to the text positive picture. After this operation, the difficulty of character detection and recognition of forward characters can be greatly reduced, the accuracy of character recognition can be significantly improved, and manual correction is not required. After the correction is obtained, target image recognition can be performed based on the correction result, and the obtained recognition result can be used to represent the text information of the object to be recognized. In an optional embodiment, the recognition may also be performed according to an object image with higher accuracy, thereby improving the accuracy of the information in the recognized object.

Optionally, correcting the target image based on the target pixel data to obtain a correction result, including: determining the vertex coordinates of the object to be recognized and the sorting order of the vertex coordinates based on the target pixel data; Correction is performed to obtain the correction result.

The above sorting order can be clockwise or counterclockwise.

In an optional embodiment, the coordinates x1, y1, x2, y2, x3, y3, x4, and y4 of the 4 vertices of the main area of the express note picture can be obtained through logical calculation, wherein the coordinates can be clockwise or counterclockwise Clock order. Among them, x1, y1 is the upper left corner vertex when the text is forward, which can be regarded as the starting point.

In another optional embodiment, according to the obtained vertex information and the corresponding starting point information, the main area of the express order picture in all directions can be corrected to the forward direction of the text through affine transformation, that is, the corrected image can be obtained. result. Among them, affine transformation is also called affine mapping, which means that in geometry, a vector space is transformed into another vector space by performing a linear transformation followed by a translation. Forward text can greatly reduce the difficulty of text detection and recognition, and can significantly improve the accuracy of text recognition without manual correction.

Optionally, correcting the target image based on the vertex coordinates and the sorting order of the vertex coordinates to obtain a correction result, further comprising: determining the target coordinates in the vertex coordinates according to the sorting order of the vertex coordinates, wherein the target coordinates are the starting point in the vertex coordinates Coordinates: Correct the object to be recognized based on the target coordinates and vertex coordinates, and obtain the correction result.

The target coordinates in the above-mentioned vertices may be the origin coordinates. Among them, the starting point is the upper left corner under the premise that the text is positive.

By affine transformation of the obtained four vertex coordinates x1, y1, x2, y2, x3, y3, x4, y4, and the starting point x1, y1, the main area of the express order image in all directions can be corrected to the correct text. To pictures, to further improve the recognition accuracy of express objects.

Optionally, detecting the target image to obtain the target pixel data includes: using a detection model to detect the target image to obtain the target pixel data.

In an optional embodiment, a detection model may be used to detect the pixels of the object to be recognized in the target image to obtain target pixel data.

In an optional embodiment, the detection model can be used to detect multiple target images at the same time, which can greatly improve the detection efficiency.

In another optional example, the detection model can be used to simultaneously detect multiple packages in a target image, so as to further improve the detection efficiency of the target image.

Optionally, obtain an original sample, wherein the original sample includes: a sample image, sample coordinates corresponding to the sample image, and the sample coordinates are vertex coordinates of the object to be identified in the sample image; based on the sample image and the sample coordinates, the sample pixel data is determined, The sample pixel data is used to represent the positional relationship between the pixels in the sample image and the object to be recognized; the training data is determined based on the sample pixel data and the sample image; the initial model is trained based on the training data to obtain the detection model.

The above-mentioned sample image may be an express package containing an object to be recognized, wherein the object to be recognized in the sample image may be forward, upside-down, oblique, or even distorted.

The above-mentioned sample coordinates may be vertex coordinates of the object to be recognized in the sample image.

The above-mentioned sample pixel data may be the positional relationship between at least one pixel in the object to be recognized and the vertex coordinates of the object to be recognized.

The above training data is the data for training the initial model.

In an optional embodiment, an original sample may be obtained, wherein the original sample may include multiple sample images and sample coordinates corresponding to the multiple sample images. It should be noted that the sample coordinates corresponding to the multiple sample images It can be the sample coordinates that are manually annotated. Optionally, after obtaining multiple sample images, the coordinates of the sample images can be manually annotated to obtain the sample coordinates corresponding to each sample image. The sample coordinates of the original sample are generated.

Further, at least one pixel in the sample image and the sample coordinates can be logically calculated to determine the sample pixel data. In order to reduce the amount of calculation, the pixel of the target area in the sample image and the sample coordinates can be logically calculated to determine the sample pixel data. . The training data may be determined according to multiple sample images and sample pixel data corresponding to the multiple sample images. Optionally, the sample pixel data and the sample image may be constructed into a sample pair, multiple sample pairs may be obtained, training data may be generated according to the multiple sample pairs, and an initial model may be trained according to the training data to obtain a detection model.

Further, the detection model can be used to detect the target image, and the target pixel data of the object to be identified in the target image can be obtained. Since the target pixel data contains the positional relationship between the pixel and the coordinates, it is possible to detect the target image in the target image. identify the pixels of the object to be identified, and perform logical calculations on the pixels according to the target pixel data to obtain the sorting order of the vertex coordinates corresponding to the pixels in the object to be identified and the starting point coordinates in the vertex coordinates, according to the sorting order of the vertex coordinates and the vertex coordinates The coordinates of the starting point of the coordinates can be corrected for the object to be recognized, so that the object to be recognized can be displayed in the forward direction, so as to facilitate the recognition of the object to be recognized, thereby greatly improving the detection efficiency of the target image.

Optionally, determining sample pixel data based on the sample image and sample coordinates includes: acquiring a target area of an object to be recognized in the sample image; acquiring a difference between pixels in the target area and sample coordinates to determine sample pixel data.

In an optional embodiment, the above-mentioned target area may be the center area of the object to be recognized, the center area of the object to be recognized in the sample image may be obtained, and the difference between the pixels of the center area and the coordinates of the sample may be obtained, Optionally, the difference between the coordinates where the pixel is located and the sample coordinates can be determined, and the positional relationship between the pixel and the sample coordinates can be determined according to the difference, and then the sample pixel data can be determined.

The target area may be an area with larger characters in the object to be recognized; the target area may also be an area with relatively clear characters in the object to be recognized, and the target area may also be a central area of the object to be recognized and a place with a lot of information. In an optional embodiment, the above-mentioned target area may be regarded as a central Gaussian distribution area, where the Gaussian distribution is a normal distribution.

Figure 6a is a sample image in the present disclosure, and Figure 6b is a map of the central Gaussian distribution area in the sample image. In the present disclosure, by extracting only the pixels of the Gaussian distribution area of the object to be recognized in the sample image, to calculate The difference between the pixels of the Gaussian distribution area of the object to be recognized in the sample image and the horizontal and vertical coordinates of the four vertices in the sample coordinates is used to determine the sample data, which greatly reduces the number of output candidate positive samples, thereby reducing the amount of calculation and improving detection. performance.

In the present disclosure, a target image is acquired, wherein the target image includes an object to be recognized; the target image is detected to obtain target pixel data, wherein the target pixel data is used to represent at least one pixel in the object to be recognized and a vertex of the object to be recognized The positional relationship between the coordinates; the target image is corrected based on the target pixel data, and the correction result is obtained. When detecting the object to be recognized in the target image, pixel-level direction supervision can be introduced, and the target image can be detected through the detection model, and the positional relationship between the pixels in the object to be recognized and the coordinates of the vertices in the object to be recognized can be obtained. , by processing the target image through the position information in the target pixel data, a more accurate object image can be obtained, thereby improving the accuracy of the obtained object image.

Fig. 6c is a flowchart of an image processing method according to the third embodiment of the present disclosure. As shown in Fig. 6c, the method includes the following steps:

Step S601 , obtaining an original sample, wherein the original sample includes: a sample image and sample coordinates corresponding to the sample image, where the sample coordinates are vertex coordinates of an object to be recognized in the sample image.

Step S602: Determine sample pixel data based on the sample image and the sample coordinates, where the sample pixel data is used to represent the positional relationship between the pixels in the sample image and the object to be recognized.

Step S603: Determine training data based on the sample pixel data and the sample image.

In step S604, the initial model is trained based on the training data to obtain a detection model.

Step S605, acquiring a target image, wherein the target image includes the object to be recognized.

Step S606, the target image is detected to obtain target pixel data, wherein the target pixel data is used to represent the positional relationship between at least one pixel in the object to be recognized and the vertex coordinates of the object to be recognized.

Step S607, correcting the target image based on the target pixel data to obtain a correction result.

In step S608, the target image is recognized based on the correction result to obtain a recognition result, wherein the recognition result is used to represent the text information of the object to be recognized in the target image.

In the technical solutions of the present disclosure, the collection, storage, use, processing, transmission, provision, and disclosure of the user's personal information involved are all in compliance with relevant laws and regulations, and do not violate public order and good customs.

From the description of the above embodiments, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is better implementation. Based on this understanding, the technical solutions of the present disclosure essentially or the parts that make contributions to the prior art can be embodied in the form of software products, and the computer software products are stored in a storage medium and include several instructions for making a A terminal device (which may be a mobile phone, a computer, a server, or a network device, etc.) executes the methods of the various embodiments of the present disclosure.

The present disclosure also provides an image processing apparatus, which is used to implement the above embodiments and preferred implementations, and what has been described will not be repeated. As used below, the term "module" may be a combination of software and/or hardware that implements a predetermined function. Although the apparatus described in the following embodiments is preferably implemented in software, implementations in hardware, or a combination of software and hardware, are also possible and contemplated.

FIG. 7 is a structural block diagram of an image processing apparatus according to an embodiment of the present disclosure. As shown in FIG. 7 , a data processing apparatus 700 includes: an acquisition module 701 , a detection module 702 , and a correction module 703 .

an acquisition module 701, configured to acquire a target image, wherein the target image includes an object to be recognized;

The detection module 702 is used to detect the target image to obtain target pixel data, wherein the target pixel data is used to represent the positional relationship between at least one pixel in the object to be recognized and the vertex coordinates of the object to be recognized;

The correction module 703 is configured to correct the target image based on the target pixel data to obtain a correction result.

Optionally, the correction module 703 includes: a first determination unit for determining the vertex coordinates of the object to be recognized and the sorting order of the vertex coordinates based on the target pixel data; a correction unit for determining the vertex coordinates based on the sorting order of the vertex coordinates. The target image is corrected to obtain the correction result.

Optionally, the correction unit includes: a determination subunit for determining target coordinates in the vertex coordinates according to the sorting order of the vertex coordinates, where the target coordinates are the starting point coordinates in the vertex coordinates; a correction subunit for determining the target coordinates based on the target coordinates and the vertex coordinates to correct the object to be recognized to obtain the correction result.

Optionally, the detection module includes: a detection unit, configured to detect the target image by using the detection model to obtain target pixel data.

Optionally, the detection module includes: an obtaining unit for obtaining an original sample, wherein the original sample includes: a sample image, sample coordinates corresponding to the sample image, and the sample coordinates are the vertex coordinates of the object to be identified in the sample image; the second a determining unit for determining sample pixel data based on the sample image and the sample coordinates, wherein the sample pixel data is used to represent the positional relationship between the pixels in the sample image and the object to be recognized; the second determining unit is also used for determining the sample pixel data based on the sample pixel data The data and sample images are used to determine training data; the second determining unit is further configured to train an initial model based on the training data to obtain a detection model.

Optionally, the second determination unit includes: an acquisition subunit, used to acquire the target area of the object to be identified in the sample image; the acquisition subunit is also used to acquire the difference between the pixels in the target area and the sample coordinates, and determine Sample pixel data.

Optionally, the device further includes: an identification module, configured to identify the target image based on the correction result to obtain a recognition result, wherein the recognition result is used to represent text information of the object to be recognized in the target image.

It should be noted that the above modules can be implemented by software or hardware, and the latter can be implemented in the following ways, but not limited to this: the above modules are all located in the same processor; or, the above modules can be combined in any combination The forms are located in different processors.

According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, comprising a memory and at least one processor, where computer instructions are stored in the memory, and the processor is configured to execute the computer instructions to execute any one of the above method implementations steps in the example.

Optionally, the electronic device may further include a transmission device and an input/output device, wherein the transmission device is connected to the processor, and the input/output device is connected to the processor.

Optionally, in the present disclosure, the above-mentioned processor may be configured to perform the following steps through a computer program:

S1, acquire a target image, wherein the target image includes an object to be recognized;

S2, the target image is detected to obtain target pixel data, wherein the target pixel data is used to represent the positional relationship between at least one pixel in the object to be recognized and the vertex coordinates of the object to be recognized;

S3, correcting the target image based on the target pixel data to obtain a correction result.

Optionally, for specific examples in this embodiment, reference may be made to the examples described in the foregoing embodiments and optional implementation manners, and details are not described herein again in this embodiment.

According to an embodiment of the present disclosure, the present disclosure also provides a non-transitory computer-readable storage medium storing computer instructions, the non-transitory computer-readable storage medium having computer instructions stored therein, wherein the computer instructions are configured to run Steps in any one of the above method embodiments are executed.

Optionally, in this embodiment, the above-mentioned non-volatile storage medium may be configured to store a computer program for executing the following steps:

S1, acquire a target image, wherein the target image includes an object to be recognized;

S2, the target image is detected to obtain target pixel data, wherein the target pixel data is used to represent the positional relationship between at least one pixel in the object to be recognized and the vertex coordinates of the object to be recognized;

S3, correcting the target image based on the target pixel data to obtain a correction result.

Optionally, in this embodiment, the above-mentioned non-transitory computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination. More specific examples of readable storage media would include one or more wire-based electrical connections, portable computer disks, hard disks, random access memory (RAM), read only memory (ROM), erasable programmable read only memory ( EPROM or flash memory), fiber optics, compact disk read only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

According to an embodiment of the present disclosure, the present disclosure also provides a computer program product. Program code for implementing the audio processing method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general purpose computer, special purpose computer or other programmable data processing apparatus, such that the program code, when executed by the processor or controller, performs the functions/functions specified in the flowcharts and/or block diagrams. Action is implemented. The program code may execute entirely on the machine, partly on the machine, partly on the machine and partly on a remote machine as a stand-alone software package or entirely on the remote machine or server.

In the above-mentioned embodiments of the present disclosure, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference may be made to related descriptions of other embodiments.

In the several embodiments provided in the present disclosure, it should be understood that the disclosed technical content may be implemented in other manners. The device embodiments described above are only illustrative, for example, the division of the units may be a logical function division, and there may be other division methods in actual implementation, for example, multiple units or components may be combined or Integration into another system, or some features can be ignored, or not implemented. On the other hand, the shown or discussed mutual coupling or direct coupling or communication connection may be through some interfaces, indirect coupling or communication connection of units or modules, and may be in electrical or other forms.

The units described as separate components may or may not be physically separated, and components shown as units may or may not be physical units, that is, may be located in one place, or may be distributed to multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution in this embodiment.

In addition, each functional unit in each embodiment of the present disclosure may be integrated into one processing unit, or each unit may exist physically alone, or two or more units may be integrated into one unit. The above-mentioned integrated units may be implemented in the form of hardware, or may be implemented in the form of software functional units.

The integrated unit, if implemented in the form of a software functional unit and sold or used as an independent product, may be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the present disclosure can be embodied in the form of software products in essence, or the part that contributes to the prior art, or all or part of the technical solutions, and the computer software product is stored in a storage medium , including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present disclosure. The aforementioned storage medium includes: U disk, read only memory (ROM), random access memory (RAM), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.

The above are only the preferred embodiments of the present disclosure. It should be pointed out that for those skilled in the art, without departing from the principles of the present disclosure, several improvements and modifications can be made. It should be regarded as the protection scope of the present disclosure.

Claims (15)

1. An image processing method, comprising: acquiring a target image, wherein the target image includes an object to be recognized; Detecting the target image to obtain target pixel data, wherein the target pixel data is used to represent the positional relationship between at least one pixel in the object to be recognized and the vertex coordinates of the object to be recognized; The target image is corrected based on the target pixel data to obtain a correction result. 2. The method according to claim 1, wherein the performing correction on the target image based on the target pixel data to obtain a correction result, comprising: Determine the vertex coordinates of the object to be recognized and the sorting order of the vertex coordinates based on the target pixel data; The target image is corrected based on the vertex coordinates and the sorting order of the vertex coordinates to obtain the correction result. 3. The method according to claim 2, wherein the correction of the target image based on the vertex coordinates and the sorting order of the vertex coordinates to obtain a correction result, comprising: Determine the starting point coordinates in the vertex coordinates according to the sorting order of the vertex coordinates; The object to be recognized is corrected based on the coordinates of the starting point and the coordinates of the vertex to obtain the correction result. 4. The method of claim 1, wherein the method further comprises: acquiring an original sample, wherein the original sample includes: a sample image, sample coordinates corresponding to the sample image, and the sample coordinates are vertex coordinates of the object to be identified in the sample image; determining sample pixel data based on the sample image and the sample coordinates, wherein the sample pixel data is used to represent the positional relationship between the pixels in the sample image and the object to be recognized; determining training data based on the sample pixel data and the sample image; The initial model is trained based on the training data to obtain a detection model; Wherein, the target image is detected to obtain target pixel data, including: The target image is detected by using the detection model to obtain target pixel data. 5. The method of claim 4, wherein the determining sample pixel data based on the sample image and the sample coordinates comprises: obtaining the target area of the object to be identified in the sample image; The difference between the pixels in the target area and the sample coordinates is acquired, and the sample pixel data is determined. 6. The method of claim 1, wherein the method further comprises: The target image is recognized based on the correction result to obtain a recognition result, wherein the recognition result is used to represent text information of the object to be recognized in the target image. 7. An image processing device, comprising: an acquisition module for acquiring a target image, wherein the target image includes an object to be recognized; The detection module is configured to detect the target image to obtain target pixel data, wherein the target pixel data is used to represent the difference between at least one pixel in the object to be recognized and the vertex coordinates of the object to be recognized. Positional relationship; A correction module, configured to correct the target image based on the target pixel data to obtain a correction result. 8. The apparatus of claim 7, wherein the correction module comprises: a first determining unit, configured to determine the vertex coordinates of the object to be identified and the sorting order of the vertex coordinates based on the target pixel data; A correction unit, configured to correct the target image based on the vertex coordinates and the sorting order of the vertex coordinates to obtain the correction result. 9. The apparatus according to claim 8, wherein the correction unit comprises: A determination subunit, configured to determine the target coordinates in the vertex coordinates according to the sorting order of the vertex coordinates, wherein the target coordinates are the starting point coordinates in the vertex coordinates; A correction subunit, configured to correct the object to be recognized based on the target coordinates and the vertex coordinates to obtain the correction result. 10. The apparatus according to claim 7, wherein the detection module comprises: an obtaining unit, configured to obtain an original sample, wherein the original sample includes: a sample image, sample coordinates corresponding to the sample image, and the sample coordinates are vertex coordinates of the object to be recognized in the sample image; a second determining unit, configured to determine sample pixel data based on the sample image and the sample coordinates, wherein the sample pixel data is used to represent a position between a pixel in the sample image and the object to be recognized relation; The second determining unit is further configured to determine training data based on the sample pixel data and the sample image; The second determining unit is further configured to train an initial model based on the training data to obtain a detection model; Wherein, the detection module further includes: A detection unit, configured to detect the target image by using the detection model to obtain target pixel data. 11. The apparatus according to claim 10, wherein the second determining unit comprises: an acquisition subunit for acquiring the target area of the object to be identified in the sample image; The obtaining subunit is further configured to obtain the difference between the pixels in the target area and the sample coordinates, and determine the sample pixel data. 12. The apparatus of claim 7, wherein the apparatus further comprises: The recognition module is configured to recognize the target image based on the correction result to obtain a recognition result, wherein the recognition result is used to represent the text information of the object to be recognized in the target image. 13. An electronic device comprising: at least one processor; and a memory communicatively coupled to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the execution of any of claims 1-6 Methods. 14. A non-transitory computer-readable storage medium storing computer instructions for causing the computer to perform the method of any of claims 1-6. 15. A computer program product comprising a computer program which, when executed by a processor, implements the method of any of claims 1-6.

Images