CN-121998576-A - Automatic food detection method, device and equipment

CN121998576ACN 121998576 ACN121998576 ACN 121998576ACN-121998576-A

Abstract

The application discloses a food automatic detection method, a device and equipment, and relates to the technical field of data processing, wherein the method comprises the steps of capturing multi-source task data of a current detection task to be checked from a multi-source data platform through a Robot Process Automation (RPA) technology, wherein the multi-source task data at least comprises one or more of a sample information text, a sample image and a detection scheme table; the method comprises the steps of automatically processing multi-source task data to generate structural audit information items, automatically comparing the structural audit information items from different information sources based on preset food detection business rules, and generating audit prompt information for the current detection task to be audited according to comparison results. The application realizes the fundamental transition from manual-based, repeated low-efficiency and error-prone to automatic-based, intelligent and efficient and error-free food detection data auditing, obviously improves auditing efficiency and reduces manual repeated labor.

Inventors

CHEN RENXI
GUO XIAOLEI
JI WANYI
QIU XIANHUI
XIA JIN
CHEN ZHENBIAO
HE QING
LIN ZHAOSHENG
JI WAN
DUAN YU
Meng Jiapeng

Assignees

华测检测认证集团股份有限公司

Dates

Publication Date: 20260508
Application Date: 20251226

Claims (10)

1. An automatic food detection method is characterized in that, the food automatic detection method comprises the following steps: Capturing multi-source task data of a current detection task to be checked from a multi-source data platform through a robot flow automation (RPA) technology, wherein the multi-source task data at least comprises one or more of a sample information text, a sample image and a detection scheme table; automatically processing the multi-source task data to generate a structured audit information item; based on a preset food detection business rule, automatically comparing the structured audit information items from different information sources; and generating audit prompt information aiming at the current detection task to be audited according to the comparison result.
2. The automated food product testing method of claim 1, wherein the multi-source data platform comprises at least a laboratory information management system and a food product sampling platform, and wherein capturing multi-source task data of a currently pending testing task from the multi-source data platform via robotic flow automation RPA technology comprises: Capturing sample information text from the laboratory information management system by an RPA technology; Capturing a sample image from the food-safe sampling platform by RPA techniques; And acquiring a detection scheme table of the current detection task to be checked and a related judgment standard file from a designated position associated with the task through an RPA technology.
3. The automated food product testing method of claim 2, wherein the step of automating the multi-source task data to generate structured audit information items includes: carrying out intelligent information extraction on the sample image in the multi-source task data to obtain a first structured information set; Performing condition analysis and standardization on a detection scheme table in the multi-source task data to obtain a second structured information set; Performing directional analysis on the judgment standard file to obtain a third structured information set; and integrating and aligning the data of the first structured information set, the second structured information set and the third structured information set to generate a unified structured audit information item.
4. The automated food product testing method of claim 3, wherein the step of intelligently extracting the sample images in the multi-source task data to obtain the first structured information set comprises: performing direction correction, graying, noise reduction and contrast enhancement on the sample image to generate a preprocessed image; Performing at least one of lateral, longitudinal, central diffusion and slide block cutting on the preprocessed image to generate a plurality of sub-area images; Performing optical character recognition OCR on each sub-area image and the original preprocessed image, and summarizing all recognition results to generate a primary text recognition result set; And inputting the content with the confidence coefficient lower than a preset threshold value in the initial text recognition result set and the corresponding image area into a natural language processing model, and selectively extracting and structuring to output a first structuring information set containing sample key information by adopting a named entity recognition technology.
5. The automated food product testing method of claim 3, wherein the step of conditional parsing and normalizing the test plan table in the multi-source task data to obtain the second structured information set comprises: analyzing the unit cell containing the production date condition in the detection scheme table, and converting the conditional statement into an associated judgment standard code number through a regular expression; carrying out standardized cleaning on the detection scheme table, splitting all merging cells, filling down contents, processing special characters and naming unified fields, and generating a cleaned detection scheme table; and generating a second structured information set based on the cleaned detection scheme table and the judgment standard code.
6. The automated food product testing method of claim 3, wherein the step of directionally parsing the decision criteria file to obtain a third structured information set comprises: judging the type of the judging standard file; If the type is editable PDF, locating and extracting text and table information related to the detection item through a PDF text extraction library in the python script; if the type is a scanned PDF, converting a PDF page into an image, and then extracting text information related to a detection item directionally by adopting the intelligent information extraction; And carrying out structuring processing on the extracted text, the table information and the text information, and establishing a mapping relation table of the detection items and the standard limit values to generate a third structuring information set.
7. The automated food product testing method of any one of claims 3-6, wherein automatically comparing the structured audit information items from different information sources based on preset food product testing business rules comprises: comparing the execution standard number in the structured audit information item with the execution standard number in the laboratory information management system; Comparing the production date in the structured audit information item with the production date in the laboratory information management system after format normalization; And matching and comparing the specification model in the structured audit information item with the specification model in the laboratory information management system and the platform in a similarity manner.
8. The automated food product testing method of claim 7, wherein the step of generating audit prompt information for the currently pending testing task based on the comparison results comprises: Analyzing the abnormal type and the abnormal grade of each abnormal record in the comparison result; Performing matching filling according to the abnormal type and a preset natural language prompt template library, and generating a preliminary prompt text of each abnormal record; And matching the abnormal grade with a preset processing strategy, generating an operation guide, and combining the operation guide with the corresponding preliminary prompt text to form the audit prompt information of the current detection task to be audited.
9. An automatic food detection device is characterized in that, the food automation detection device includes: The system comprises a data grabbing module, a detection scheme table and a detection scheme table, wherein the data grabbing module is used for grabbing multi-source task data of a current detection task to be checked from a multi-source data platform through a robot flow automation (RPA) technology, and the multi-source task data at least comprises one or more of a sample information text, a sample image and the detection scheme table; the data processing module is used for automatically processing the multi-source task data and generating a structured audit information item; The rule comparison module is used for automatically comparing the structured audit information items from different information sources based on a preset food detection business rule; And the prompt generation module is used for generating audit prompt information aiming at the current detection task to be audited according to the comparison result.
10. An automated food product inspection apparatus comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the computer program being configured to implement the steps of the automated food product inspection method according to any one of claims 1 to 8.

Description

Automatic food detection method, device and equipment Technical Field The present application relates to the field of data processing technologies, and in particular, to a method, an apparatus, and a device for automatically detecting food. Background Along with the rapid development of artificial intelligence and automation technology, the application of the method in the fields of natural language understanding, data processing, flow automation and the like is deepened continuously, and the method provides technical possibility for intelligent transformation of food detection data auditing. Food detection report auditing is used as a core link of quality control, and accurate consistency of multi-dimensional information such as sample information, inspection items, method applicability, qualification compliance, result reliability, judgment correctness and the like is required to be ensured, and the information is derived from various heterogeneous data sources such as national extraction platforms, sample photographs, laboratory Information Management Systems (LIMS), scheme tables, national and product standards and the like, so that auditing work has high complexity and cross-system integration requirements. Currently, some technologies are tried to be applied to local links of data auditing, such as an automatic auditing system of prepackaged food labels based on OCR and NLP, a LIMS abnormal result marking mechanism based on a rule engine and a statistical model, and research and patents for assisting report conclusion checking through semantic analysis. However, these prior art techniques focus on a specific link or single data type in the auditing process, and no integrated solution has been formed covering the whole chain from sample information entry, detection item matching, method qualification verification to result determination. Due to the lack of systematic grabbing, structuring and automatic comparison capabilities for multi-source heterogeneous data, the prior art is difficult to deal with the reality scenes of information dispersion, different formats and complex logic association in food detection and auditing. Therefore, the prior art still highly relies on manual work to check and compare the information of the cross-platform and cross-document in the food detection data auditing stage, so that the manual work is heavy in repeated work, the core leakage and the core error are easily caused by fatigue and negligence, the error rate is high, meanwhile, the error is often discovered in the auditing later stage and even in the report generating stage, the process reworking, the period delay and the process efficiency are low. The food detection industry is in need of a systematic technical scheme capable of penetrating through multi-source data and realizing full-flow automatic intelligent auditing. The foregoing is provided merely for the purpose of facilitating understanding of the technical solutions of the present application and is not intended to represent an admission that the foregoing is prior art. Disclosure of Invention The application mainly aims to provide a food automatic detection method, device and equipment, and aims to solve the technical problems of high repeatability labor capacity, high error rate and low flow efficiency caused by manually extracting information between multi-source heterogeneous data platforms in a food detection data auditing stage. In order to achieve the above object, the present application provides an automatic food detection method, which comprises: Capturing multi-source task data of a current detection task to be checked from a multi-source data platform through a robot flow automation (RPA) technology, wherein the multi-source task data at least comprises one or more of a sample information text, a sample image and a detection scheme table; automatically processing the multi-source task data to generate a structured audit information item; based on a preset food detection business rule, automatically comparing the structured audit information items from different information sources; and generating audit prompt information aiming at the current detection task to be audited according to the comparison result. In an embodiment, the multi-source data platform at least comprises a laboratory information management system and a food safety sampling platform, and the capturing multi-source task data of the currently pending detection task from the multi-source data platform by the robot flow automation RPA technology comprises: Capturing sample information text from the laboratory information management system by an RPA technology; Capturing a sample image from the food-safe sampling platform by RPA techniques; And acquiring a detection scheme table of the current detection task to be checked and a related judgment standard file from a designated position associated with the task through an RPA technology. In one embodiment, the step of automaticall