CN-121981124-A - Language landscape text grammar correction method and system based on image recognition

CN121981124ACN 121981124 ACN121981124 ACN 121981124ACN-121981124-A

Abstract

The invention relates to the technical field of computer vision and natural language processing, in particular to a language landscape text grammar correction method and system based on image recognition, which are characterized in that images containing language landscape texts are collected through a mobile terminal camera; the method comprises the steps of carrying out optical character recognition on an image, extracting a target language text, inputting the text into a grammar error diagnosis model to obtain a diagnosis result containing error positions, types and correction suggestions, and finally carrying out space alignment and superposition display on the diagnosis result and a text area in the image in an augmented reality mode on a mobile terminal screen. The invention provides a real-time, in-situ and visual quality checking and learning tool for city managers, scenic spot operators and foreign language learners directly aiming at grammar and translation quality problems of language landscape texts in public places, and effectively solves the problems that the traditional manual inspection efficiency is low and the online tool cannot be checked in linkage with a physical scene.

Inventors

LIU SHENG
LI YUAN
LIANG JIANGUO

Assignees

山东石油化工学院

Dates

Publication Date: 20260505
Application Date: 20260204

Claims (7)

1. The language landscape text grammar correction method based on image recognition is characterized by comprising the following steps of: S1, acquiring an image containing language landscape text through a camera of a mobile terminal; s2, performing optical character recognition on the image, and extracting text content of at least one target language; s3, inputting the identified target language text into a grammar error diagnosis model to obtain a grammar diagnosis result, wherein the grammar diagnosis result at least comprises an error position, an error type and correction suggestions; And S4, carrying out spatial position alignment and superposition display on the grammar diagnosis result and a text region in the image on a display screen of the mobile terminal in an augmented reality mode.
2. The method for grammar correction of language-landscape text based on image recognition of claim 1, wherein when the text extracted in the step S2 contains both original text and translated text, the step S3 further comprises: And carrying out semantic association analysis on the original text and the translation text to generate a translation quality comment, wherein the translation quality comment is used for evaluating the tradition and the accuracy of the translation and is taken as additional information to be included in a grammar diagnosis result.
3. The method for grammar correction of language landscape text based on image recognition as claimed in claim 2, wherein in step S3, the grammar error diagnosis model is a personalized grammar diagnosis model based on a native language background, and can perform depth attribution analysis according to the native language background information related to the text and output a diagnosis result.
4. The method for text grammar correction for language landscape based on image recognition as claimed in claim 3, wherein said step S4 comprises: s41, positioning a text region on a camera picture previewed in real time on a display screen or a shot still image; S42, highlighting words and sentences with grammar errors in the text region according to the error position by using a first visual style; And S43, in response to the triggering operation of the user on the highlight mark, a feedback window is popped up, wherein the error type, the correction proposal and the translation quality comment are integrally displayed.
5. A language landscape text grammar correction system based on image recognition, comprising: The image acquisition module is used for controlling the camera to acquire an image containing language landscape text; the OCR processing module is used for carrying out character recognition on the image and outputting the text of at least one target language; the grammar diagnosis module is used for receiving the text, calling a grammar error diagnosis model for processing and outputting a grammar diagnosis result, wherein the grammar diagnosis result at least comprises an error position, an error type and correction suggestions; and the AR rendering display module is used for aligning and superposing the diagnosis result with the text position in the image.
6. The language landscape text grammar correction system based on image recognition of claim 5, further comprising a translation quality analysis module for comparing and analyzing semantic consistency of the original text and the translation text to generate a translation quality comment, wherein the translation quality comment is used for evaluating the tradition and the accuracy of translation and is included as additional information into a grammar diagnosis result.
7. The language landscape text grammar correction system based on image recognition of claim 6, wherein the grammar diagnostic module is capable of dual mode configuration: The mode A is a cloud intelligent mode, and diagnostic services deployed in the cloud are called through a network API; Mode B is a terminal lightweight mode, and a lightweight syntax error diagnosis model is run locally at the terminal.

Description

Language landscape text grammar correction method and system based on image recognition Technical Field The invention relates to the technical field of computer vision and natural language processing, in particular to a language landscape text grammar correction method and system based on image recognition, which are particularly suitable for intelligent correction and specification of language landscape texts. Background With the acceleration of globalization process, language landscapes (Linguistic Landscape) of public places, such as bilingual marks, navigation information, propaganda posters and the like, and grammar accuracy and expression trails of foreign language translation have become important dimensions for measuring internationalization level. However, in reality, various grammar errors and Chinese English phenomena often occur in the translation text, and the effective transmission of international images and information is seriously affected. At present, the quality inspection means for such texts have significant bottlenecks, mainly depend on the following two types of technical schemes, and have fundamental defects: The first technical solution is based on Optical Character Recognition (OCR), whose functional scope is limited to converting text in an image into a sequence of characters, and the system architecture lacks functional modules for performing grammar and semantic analysis on the sequence, so that language errors of the text itself cannot be recognized and corrected. The second category of solutions is online grammatical analysis, which is capable of analyzing text, but if the user manually types in to convert the physical text into a digital format. The non-automatic interaction mode causes complicated flow and low efficiency, and cannot realize real-time and in-situ inspection of texts in a physical environment. Therefore, developing a language landscape text grammar correction method and system based on image recognition has become a problem to be solved urgently. Disclosure of Invention The invention aims to provide a language landscape text grammar correction method and system based on image recognition, which can seamlessly integrate visual perception, intelligent diagnosis and augmented reality feedback, so as to solve the problem that a first type of technical scheme (based on optical character recognition) in the background technology lacks depth analysis capability, and a second type of technical scheme (online grammar analysis) cannot interact with a physical scene directly. Still another object of the present invention is to enable a user to obtain accurate grammar error diagnosis, correction advice and immersive visual feedback in real time and automatically by using the method and the system of the present invention only to shoot the target text by using the mobile device. In order to achieve the purpose, the technical scheme provided by the invention is that the language landscape text grammar correction method based on image recognition comprises the following steps: S1, acquiring an image containing language landscape text through a camera of a mobile terminal; s2, performing optical character recognition on the image, and extracting text content of at least one target language; s3, inputting the identified target language text into a grammar error diagnosis model to obtain a grammar diagnosis result, wherein the grammar diagnosis result at least comprises an error position, an error type and correction suggestions; And S4, carrying out spatial position alignment and superposition display on the grammar diagnosis result and a text region in the image on a display screen of the mobile terminal in an augmented reality mode. Further, when the text extracted in step S2 includes an original text and a translated text, step S3 further includes: And carrying out semantic association analysis on the original text and the translation text to generate a translation quality comment, wherein the translation quality comment is used for evaluating the tradition and the accuracy of the translation and is taken as additional information to be included in a grammar diagnosis result. Further, in step S3, the grammar error diagnosis model is a personalized grammar diagnosis model based on a native language background, and is capable of performing depth attribution analysis according to the native language background information associated with the text, and outputting a diagnosis result. Further, the step S4 specifically includes: s41, positioning a text region on a camera picture previewed in real time on a display screen or a shot still image; S42, highlighting words and sentences with grammar errors in the text region according to the error position by using a first visual style; And S43, in response to the triggering operation of the user on the highlight mark, a feedback window is popped up, wherein the error type, the correction proposal and the translation quality comment are