A streaming text recognition method, device, electronic device and storage medium

Abstract

The present invention relates to a method, device, electronic device and storage medium for stream text recognition, the method comprising: obtaining a current stream graph; determining a blank candidate area of ​​the current stream graph; determining a cutting start point and a cutting end point of the current stream graph according to threshold parameters and strategy parameters of the blank candidate area; cutting and outputting the cutting result of the current stream graph according to the cutting start point and the cutting end point; identifying the cutting result of the current stream graph and the cutting result of the previous stream graph, and splicing them to obtain the final result. After determining the blank candidate area, the present application determines the cutting start point and the cutting end point of the current stream graph by the threshold parameters and strategy parameters of the blank candidate area, thereby improving the accuracy of determining the blank candidate area; obtaining the cutting result of the current stream graph by determining the cutting start point and the cutting end point solves the problem of text recognition errors in direct cutting; identifying and splicing to determine the final result improves the accuracy of the stream text recognition result.

Description

Technical Field

[0001] The present invention relates to the technical field of text recognition, and in particular to a streaming text recognition method, device, electronic equipment and storage medium. Background Art

[0002] At present, in the process of determining the boundary of the flow graph, it depends on the recognition results of the last or last few OCRs. Through the historical OCR recognition results, the length and width ratio of the boundary and the threshold parameters are determined to control and determine the boundary of the flow graph. Through the overlapping matching strategy of the flow graph, the final flow recognition results are spliced.

[0003] Problems: The boundary of the streaming graph is coupled with the result of OCR. In an offline environment, the boundary of the streaming graph cannot be determined locally to obtain the result of OCR, so the boundary of the streaming graph cannot be determined; the current boundary logic of the streaming graph has forced excision in some cases, resulting in the text being cut off, affecting the final text recognition effect. The current jigsaw strategy cannot obtain the correct splicing result for some incorrect inputs. Summary of the invention

[0004] The present invention provides a streaming text recognition method, device, electronic device and storage medium, which can solve the above technical problems.

[0005] The technical solution of the present invention to solve the above technical problems is as follows:

[0006] In a first aspect, an embodiment of the present invention provides a streaming text recognition method, comprising:

[0007] Get the current streaming graph;

[0008] Determine the blank candidate area of ​​the current flow graph;

[0009] Determine the starting point and the ending point of the current streaming graph according to the threshold parameters and strategy parameters of the blank candidate area;

[0010] Output the current stream graph cutting result according to the starting point and the ending point of the graph cutting;

[0011] Identify the current stream graph cutting result and the previous stream graph cutting result, and splice them to get the final result.

[0012] Furthermore, in the above-mentioned streaming text recognition method, determining the blank candidate area of ​​the current streaming graph includes:

[0013] Get the threshold parameters of the image area to be segmented and the blank candidate area of ​​the current streaming graph;

[0014] Pre-process the image area to be segmented and calculate the binary image;

[0015] The blank candidate area is determined according to the binarization image and the threshold parameter of the blank candidate area.

[0016] Furthermore, in the above-mentioned streaming text recognition method, pre-processing the image area and calculating the binarization image includes:

[0017] Compute the histogram from the vertical and horizontal projections.

[0018] Furthermore, in the above-mentioned streaming text recognition method, determining the cutting start point and the cutting end point of the current streaming graph according to the threshold parameters and strategy parameters of the blank candidate area includes:

[0019] Determine boundary information of the blank candidate area according to threshold parameters and strategy parameters of the blank candidate area;

[0020] The starting point and the ending point of the current streaming graph are determined according to the boundary information of the blank candidate area.

[0021] Furthermore, the above-mentioned streaming text recognition method also includes: updating the threshold parameters of the current blank candidate area.

[0022] Furthermore, in the above-mentioned streaming text recognition method, the threshold parameters of the blank candidate area include at least: a blank width threshold, a farthest threshold of the end position of the previous frame, and a nearest threshold of the end position of the previous frame;

[0023] The strategy parameters include the starting position of the last slice and the ending position of the last slice.

[0024] Furthermore, in the above-mentioned streaming text recognition method, identifying the current streaming graph cutting result and the previous streaming graph cutting result, and splicing them to obtain the final result, includes:

[0025] Get the current text information and text coordinate information, the previous text information and text coordinate information, and the overlap information between the current cut and the previous cut;

[0026] Determine the overlap between the current text and the previous text according to the current text information and text coordinate information, the previous text information and text coordinate information, and the overlap information between the current cut image and the previous cut image;

[0027] The characters are concatenated according to the degree of overlap to obtain a final result.

[0028] In a second aspect, an embodiment of the present invention further provides a streaming text recognition device, comprising:

[0029] Acquisition module: used to obtain the current streaming graph;

[0030] The first determination module is used to determine the blank candidate area of ​​the current flow graph;

[0031] The second determination module is used to determine the cutting start point and the cutting end point of the current streaming graph according to the threshold parameters and strategy parameters of the blank candidate area;

[0032] Graph cutting output module: used to cut and output the graph cutting result of the current streaming graph according to the graph cutting starting point and the graph cutting ending point;

[0033] Identification and splicing module: used to identify the slice result of the current streaming graph and the slice result of the previous streaming graph, and splice them to get the final result.

[0034] In a third aspect, an embodiment of the present invention further provides an electronic device, including: a processor and a memory;

[0035] The processor is used to execute any one of the streaming text recognition methods described above by calling the program or instruction stored in the memory.

[0036] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a program or instruction, wherein the program or instruction enables a computer to execute a streaming text recognition method as described in any one of the above items.

[0037] The beneficial effects of the present invention are as follows: the present invention relates to a method, device, electronic device and storage medium for streaming text recognition, the method comprising: obtaining the current streaming graph; determining a blank candidate area of ​​the current streaming graph; determining a cutting start point and a cutting end point of the current streaming graph according to threshold parameters and strategy parameters of the blank candidate area; cutting and outputting the cutting result of the current streaming graph according to the cutting start point and the cutting end point; identifying the cutting result of the current streaming graph and the cutting result of the previous streaming graph, and splicing them to obtain the final result. In the present application, after determining the blank candidate area, the cutting start point and the cutting end point of the current streaming graph are determined by the threshold parameters and strategy parameters of the blank candidate area, thereby improving the accuracy of determining the blank candidate area; obtaining the cutting result of the current streaming graph by determining the cutting start point and the cutting end point solves the problem of text recognition errors caused by direct cutting; determining the final result by identifying and splicing improves the accuracy of the streaming text recognition result. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 A streaming text recognition method provided by an embodiment of the present invention Figure 1 ;

[0039] Figure 2 A streaming text recognition method provided by an embodiment of the present invention Figure 2 ;

[0040] Figure 3 A streaming text recognition method provided by an embodiment of the present invention Figure 3 ;

[0041] Figure 4 A streaming text recognition method provided by an embodiment of the present invention Figure 4 ;

[0042] Figure 5 A diagram of a streaming text recognition device provided by an embodiment of the present invention;

[0043] Figure 6 A schematic block diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0044] The principles and features of the present invention are described below in conjunction with the accompanying drawings. The examples given are only used to explain the present invention and are not used to limit the scope of the present invention.

[0045] In order to more clearly understand the above-mentioned purpose, features and advantages of the present application, the present disclosure is further described in detail below in conjunction with the accompanying drawings and examples. It is to be understood that the described embodiments are part of the embodiments of the present disclosure, rather than all of the embodiments. The specific embodiments described herein are merely used to explain the present disclosure, rather than to limit the present application. All other embodiments obtained by those of ordinary skill in the art according to the described embodiments of the present application belong to the scope of protection of the present application.

[0046] It should be noted that, in this document, relational terms such as “first” and “second” are merely used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations.

[0047] The system architecture of this method consists of two parts, including the device side and the server side.

[0048] The device side includes: a streaming graph cutting strategy module, which receives the input of the streaming graph, obtains the cutting start point and end point of the current streaming graph through the streaming graph cutting strategy module, and uploads the data between the starting point and the end point to the server side for result display through compression and encryption, receives the splicing recognition result returned by the server, and displays it on the device side.

[0049] Figure 1 A streaming text recognition method provided by an embodiment of the present invention Figure 1 .

[0050] In a first aspect, an embodiment of the present invention provides a streaming text recognition method, combining Figure 1 , including five steps from S101 to S105:

[0051] In a first aspect, an embodiment of the present invention provides a streaming text recognition method, including five steps S101 to S105:

[0052] S101: Get the current flow graph.

[0053] Specifically, in the embodiment of the present application, the current flow diagram is obtained by scanning, and the current flow diagram can be obtained by scanning with a scanning device.

[0054] S102: Determine a blank candidate area of ​​the current flow graph.

[0055] Specifically, in an embodiment of the present application, after scanning and acquiring the current flow graph, the blank candidate area is determined by acquiring the image area to be segmented of the current flow graph and the threshold parameters of the blank candidate area. The blank candidate area in the present application can be a blank candidate area between text and text, or a blank area between letters of English words, etc., without specific limitation. The following text combines steps S201 to S203 to introduce in detail the process of determining the blank candidate area of ​​the current flow graph.

[0056] S103: Determine the starting point and the ending point of the current streaming graph according to the threshold parameters and the strategy parameters of the blank candidate area.

[0057] Specifically, in the embodiment of the present application, based on the threshold parameters and strategy parameters of the blank candidate area, the cutting start point and the cutting end point of the current streaming graph are determined, and the threshold parameters of the blank candidate area at least include: the blank width threshold, the farthest threshold of the end position of the previous frame, and the nearest threshold of the end position of the previous frame; the strategy parameters include the starting position of the previous cutting and the ending position of the previous cutting. It should be understood that the accuracy of determining the blank candidate area is improved by determining the cutting start point and the cutting end point of the current streaming graph through the threshold parameters and strategy parameters of the blank candidate area.

[0058] S104: Cut the graph according to the graph cutting starting point and the graph cutting ending point and output the graph cutting result of the current streaming graph.

[0059] Specifically, in the embodiment of the present application, the data cutting output between the cutting start point and the cutting end point obtains the cutting result of the current streaming graph, which solves the problem of text recognition errors caused by direct cutting.

[0060] S105: Identify the current stream graph cutting result and the previous stream graph cutting result, and stitch them together to obtain the final result.

[0061] Specifically, in the embodiment of the present application, the final result is determined by identifying the text coordinates, text information and image overlap information of the current streaming graph cutting result and the previous streaming graph cutting result, and then splicing them together to improve the accuracy of the final streaming text recognition result.

[0062] It should be understood that in the system architecture of the present application, the device receives the input of the streaming graph, obtains the current streaming graph cutting start point and end point through the streaming graph cutting strategy module, and can also upload the data between the starting point and the end point to the server after cutting through compression and encryption, or directly upload the data between the current streaming graph cutting start point and the end point to the server after cutting, and the text recognition module and spelling module of the server recognize the current streaming graph cutting result and the previous streaming graph cutting result, splice them to get the final result, and send it to the device for display. That is, a streaming scanning cutting logic that is completely independent of OCR is proposed, which realizes the decoupling of streaming cutting and OCR, and lays the foundation for realizing separate streaming puzzles on low-resource devices.

[0063] Figure 2 A streaming text recognition method provided by an embodiment of the present invention Figure 2 .

[0064] Furthermore, in the above-mentioned streaming text recognition method, the blank candidate area of ​​the current streaming graph is determined, combined with Figure 2 , including three steps from S201 to S203:

[0065] S201: Obtaining threshold parameters of the image area to be segmented and the blank candidate area of ​​the current streaming graph;

[0066] Specifically, in an embodiment of the present application, the image area to be segmented of the current streaming graph may be the image area to be segmented determined based on the current streaming graph, and the threshold parameter of the blank candidate area may be a parameter updated by the user based on the word count constraint of the current streaming graph.

[0067] S202: Pre-process the image area to be segmented and calculate a binary image;

[0068] Specifically, in the embodiment of the present application, the pre-processing may be text enhancement or the like, and the calculation of the binarized image is to calculate the histogram according to the vertical projection and the horizontal projection.

[0069] S203: Determine a blank candidate area according to the binarized image and the threshold parameter of the blank candidate area.

[0070] Specifically, in the embodiment of the present application, the blank candidate area is determined based on the histogram calculated by the vertical projection and the horizontal projection, the blank width threshold, the farthest threshold of the end position of the previous frame, and the closest threshold of the end position of the previous frame. It should be understood that the blank candidate area in the embodiment of the present application cannot be too large or too small, and the blank candidate area determined by the histogram and the threshold range is optimal.

[0071] Figure 3 A streaming text recognition method provided by an embodiment of the present invention Figure 3 .

[0072] Furthermore, in the above-mentioned streaming text recognition method, the starting point and the ending point of the current streaming graph are determined according to the threshold parameters and the strategy parameters of the blank candidate area, and the Figure 3 , including two steps S301 to S302:

[0073] S301: Determine boundary information of the blank candidate area according to the threshold parameters and strategy parameters of the blank candidate area.

[0074] S302: Determine the cutting start point and the cutting end point of the current stream graph according to the boundary information of the blank candidate area.

[0075] Specifically, in an embodiment of the present application, after the above steps determine the blank candidate area, the boundary information of the blank candidate area is determined by the threshold parameters and strategy parameters of the blank candidate area, and the threshold parameters of the blank candidate area include at least: a blank width threshold, a farthest threshold of the end position of the previous frame, and a closest threshold of the end position of the previous frame; the strategy parameters include the starting position of the previous cut and the end position of the previous cut. It should be understood that the present application combines the strategy parameters to improve the accuracy of the boundary information of the blank candidate area. After determining the boundary information, the starting point and the end point of the cut of the current streaming graph are determined according to the boundary information, thereby improving the accuracy of determining the starting point and the end point of the cut.

[0076] Furthermore, the above-mentioned streaming text recognition method also includes: updating the threshold parameters of the current blank candidate area.

[0077] Specifically, in the embodiment of the present application, the scanning device will have three parameters built-in during initial startup: the blank width threshold, the farthest threshold of the end position of the previous frame, and the nearest threshold of the end position of the previous frame. When the first stream graph is actually scanned, the above three parameters will be updated according to the word count constraint of the stream graph. Considering that the stream graph cannot be too long or too short, the word count is controlled to be 4 to 7 Chinese characters wide and 2 to 4 words wide in English.

[0078] Figure 4 A streaming text recognition method provided by an embodiment of the present invention Figure 4 .

[0079] Furthermore, in the above-mentioned streaming text recognition method, the sliced ​​image results of the current streaming graph are identified and spliced ​​to obtain the final result, and combined with Figure 4 ,include:

[0080] S401: Acquire current text information and text coordinate information, previous text information and text coordinate information, and overlap information between the current cut image and the previous cut image;

[0081] S402: determining the overlap between the current text and the previous text according to the current text information and text coordinate information, the previous text information and text coordinate information, and the overlap information between the current cut image and the previous cut image;

[0082] S403: splicing the characters according to the degree of overlap to obtain a final result.

[0083] Specifically and exemplarily, the result of the English word hello world when scanned for the first time is helloworld, and the result of the second scan is world, that is, world overlaps, and when spliced ​​together, world is covered.

[0084] It should be understood that in the present application, the degree of overlap between the current text and the previous text is determined based on the current text information and text coordinate information, the previous text information and text coordinate information, and the overlapping information of the current cut image and the previous cut image, and the text is spliced ​​according to the overlap degree to obtain the final result, thereby improving the accuracy of the final streaming text recognition.

[0085] Figure 5 A diagram of a streaming text recognition device provided in an embodiment of the present invention.

[0086] In a second aspect, an embodiment of the present invention further provides a streaming text recognition device, Figure 5 ,include:

[0087] Acquisition module 501: used to acquire the current flow graph.

[0088] Specifically, in the embodiment of the present application, the acquisition module 501 acquires the current flow diagram by scanning, and the current flow diagram can be obtained by scanning with a scanning device.

[0089] The first determination module 502 is used to determine the blank candidate area of ​​the current flow graph.

[0090] Specifically, in an embodiment of the present application, after scanning and acquiring the current flow graph, the first determination module 502 determines the blank candidate area through the image area to be segmented of the current flow graph and the threshold parameters of the blank candidate area. The blank candidate area in the present application can be a blank candidate area between text and text, or a blank area between letters of English words, etc., without specific limitation. The above text combines steps S201 to S203 to introduce in detail the process of determining the blank candidate area of ​​the current flow graph.

[0091] The second determination module 503 is used to determine the image cutting start point and the image cutting end point of the current stream graph according to the threshold parameters and strategy parameters of the blank candidate area.

[0092] Specifically, in the embodiment of the present application, the second determination module 503 determines the start and end points of the current stream graph based on the threshold parameters and strategy parameters of the blank candidate area. The threshold parameters of the blank candidate area include at least: the blank width threshold, the farthest threshold of the end position of the previous frame, and the nearest threshold of the end position of the previous frame; the strategy parameters include the start position of the previous cut and the end position of the previous cut. The start and end points of the current stream graph are determined by the threshold parameters and strategy parameters of the blank candidate area, which improves the accuracy of determining the blank candidate area.

[0093] The graph cutting output module 504 is used to cut and output the graph cutting result of the current stream graph according to the graph cutting starting point and the graph cutting ending point.

[0094] Specifically, in the embodiment of the present application, the graph cutting output module 504 obtains the graph cutting result of the current stream graph by outputting the data between the graph cutting start point and the graph cutting end point, thereby solving the problem of text recognition errors caused by direct graph cutting.

[0095] Identification and splicing module 505: used to identify the slice result of the current stream graph and the slice result of the previous stream graph, and splice them to obtain the final result.

[0096] Specifically, in the embodiment of the present application, the recognition and splicing module 505 recognizes the text coordinates, text information and image overlap information and then splices them to determine the final result, thereby improving the accuracy of the final streaming text recognition result.

[0097] In a third aspect, an embodiment of the present invention further provides an electronic device, including: a processor and a memory;

[0098] The processor is used to execute any one of the streaming text recognition methods described above by calling the program or instruction stored in the memory.

[0099] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a program or instruction, wherein the program or instruction enables a computer to execute a streaming text recognition method as described in any one of the above items.

[0100] Figure 6 It is a schematic block diagram of an electronic device provided by an embodiment of the present disclosure.

[0101] like Figure 6As shown, the electronic device includes: at least one processor 601, at least one memory 602 and at least one communication interface 603. The various components in the electronic device are coupled together through a bus system 604. The communication interface 603 is used to transmit information between external devices. It can be understood that the bus system 604 is used to achieve connection and communication between these components. In addition to the data bus, the bus system 604 also includes a power bus, a control bus and a status signal bus. However, for the sake of clarity, Figure 6 Various buses are labeled as bus system 604 .

[0102] It can be understood that the memory 602 in this embodiment can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories.

[0103] In some implementations, the memory 602 stores the following elements, executable units or data structures, or a subset thereof, or an extended set thereof: an operating system and application programs.

[0104] The operating system includes various system programs, such as a framework layer, a core library layer, a driver layer, etc., which are used to implement various basic services and process tasks based on hardware. The application includes various application programs, such as a media player (Media Player), a browser (Browser), etc., which are used to implement various application services. A program that implements any method of a streaming text recognition method provided in an embodiment of the present application can be included in the application.

[0105] In an embodiment of the present application, the processor 601 calls a program or instruction stored in the memory 602, specifically, a program or instruction stored in an application, and the processor 601 is used to execute the steps of each embodiment of a streaming text recognition method provided in an embodiment of the present application.

[0106] Get the current streaming graph;

[0107] Determine the blank candidate area of ​​the current flow graph;

[0108] Determine the starting point and the ending point of the current streaming graph according to the threshold parameters and strategy parameters of the blank candidate area;

[0109] Output the current stream graph cutting result according to the starting point and the ending point of the graph cutting;

[0110] Identify the current stream graph cutting result and the previous stream graph cutting result, and splice them to get the final result.

[0111] Any method in a streaming text recognition method provided in an embodiment of the present application can be applied to a processor 601, or implemented by a processor 601. The processor 601 can be an integrated circuit chip having signal processing capabilities. In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware in the processor 601 or an instruction in the form of software. The above-mentioned processor 601 can be a general-purpose processor, a digital signal processor (Digital Signal Processor, DSP), an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), a field programmable gate array (Field Programmable Gate Array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc.

[0112] The steps of any method in a streaming text recognition method provided in an embodiment of the present application can be directly embodied as being executed by a hardware decoding processor, or being executed by a combination of hardware and software units in a decoding processor. The software unit can be located in a mature storage medium in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory 602, and the processor 601 reads the information in the memory 602, and completes the steps of a streaming text recognition method in combination with its hardware.

[0113] Those skilled in the art will appreciate that although some embodiments described herein include certain features included in other embodiments but not other features, the combination of features from different embodiments is meant to be within the scope of the present application and to form different embodiments.

[0114] Those skilled in the art will appreciate that the description of each embodiment has its own emphasis, and for parts not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0115] Although the embodiments of the present application have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present application, and such modifications and variations are all within the scope defined by the appended claims.

[0116] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A streaming text recognition method, characterized in that: include: Get the current streaming graph; Determine a blank candidate area of ​​the current flow graph; Determine the starting point and the ending point of the current streaming graph according to the threshold parameters and strategy parameters of the blank candidate area; Outputting the cutting result of the current stream graph according to the cutting starting point and the cutting ending point; Identify the text of the slice result of the current stream graph, and concatenate the text to obtain a final result; The threshold parameters of the blank candidate area at least include: a blank width threshold, a farthest threshold of the end position of the previous frame, and a nearest threshold of the end position of the previous frame; The strategy parameters include the starting position of the last image cut and the ending position of the last image cut.

2. A streaming text recognition method according to claim 1, characterized in that: The determining of the blank candidate area of ​​the current stream graph includes: Get the threshold parameters of the image area to be segmented and the blank candidate area of ​​the current streaming graph; Pre-processing the image area to be segmented and calculating a binary image; The blank candidate area is determined according to the binarization image and a threshold parameter of the blank candidate area.

3. A streaming text recognition method according to claim 2, characterized in that: The pre-processing of the image region and calculating a binary image includes: Compute the histogram from the vertical and horizontal projections.

4. A streaming text recognition method according to claim 1, characterized in that: The determining of the cutting start point and the cutting end point of the current streaming graph according to the threshold parameter and the strategy parameter of the blank candidate area includes: Determine boundary information of the blank candidate area according to threshold parameters and strategy parameters of the blank candidate area; The starting point and the ending point of the current stream graph are determined according to the boundary information of the blank candidate area.

5. A streaming text recognition method according to claim 4, characterized in that: The method further includes: updating a threshold parameter of the current blank candidate area.

6. A streaming text recognition method according to claim 1, characterized in that: The identifying the text of the cut result of the current stream graph and splicing the text to obtain the final result includes: Get the current text information and text coordinate information, the previous text information and text coordinate information, and the overlap information between the current cut and the previous cut; Determine the overlap degree between the current text and the previous text according to the current text information and the text coordinate information, the previous text information and the text coordinate information, and the overlap information between the current cut image and the previous cut image; The characters are concatenated according to the degree of overlap to obtain a final result.

7. A streaming text recognition device, characterized in that: include: Acquisition module: used to obtain the current streaming graph; A first determination module: used to determine a blank candidate area of ​​the current flow graph; The second determination module is used to determine the cutting start point and the cutting end point of the current streaming graph according to the threshold parameters and strategy parameters of the blank candidate area; A graph cutting output module: used for cutting and outputting the graph cutting result of the current stream graph according to the graph cutting starting point and the graph cutting ending point; Identification and splicing module: used for identifying the text of the cut result of the current stream graph, and splicing the text to obtain the final result; The threshold parameters of the blank candidate area at least include: a blank width threshold, a farthest threshold of the end position of the previous frame, and a nearest threshold of the end position of the previous frame; The strategy parameters include the starting position of the last image cut and the ending position of the last image cut.

8. An electronic device, characterized in that: include: Processor and memory; The processor is used to execute a streaming text recognition method as described in any one of claims 1 to 6 by calling the program or instruction stored in the memory.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a program or instruction, and the program or instruction enables a computer to execute a streaming text recognition method as described in any one of claims 1 to 6.

Images