CN116563631A - AI-based instrument panel visual identification method, system and storage medium - Google Patents
AI-based instrument panel visual identification method, system and storage medium Download PDFInfo
- Publication number
- CN116563631A CN116563631A CN202310531877.2A CN202310531877A CN116563631A CN 116563631 A CN116563631 A CN 116563631A CN 202310531877 A CN202310531877 A CN 202310531877A CN 116563631 A CN116563631 A CN 116563631A
- Authority
- CN
- China
- Prior art keywords
- data
- information
- preset
- instrument panel
- image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 230000000007 visual effect Effects 0.000 title claims abstract description 41
- 238000000034 method Methods 0.000 title claims abstract description 30
- 230000006399 behavior Effects 0.000 claims abstract description 34
- 230000008859 change Effects 0.000 claims description 52
- 230000007613 environmental effect Effects 0.000 claims description 36
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Chemical compound O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 34
- 238000007781 pre-processing Methods 0.000 claims description 28
- 238000003062 neural network model Methods 0.000 claims description 14
- 238000011282 treatment Methods 0.000 claims description 7
- 230000007797 corrosion Effects 0.000 claims description 5
- 238000005260 corrosion Methods 0.000 claims description 5
- 238000001514 detection method Methods 0.000 abstract description 17
- 238000005516 engineering process Methods 0.000 abstract description 6
- 238000004458 analytical method Methods 0.000 description 18
- 238000013528 artificial neural network Methods 0.000 description 10
- 238000005286 illumination Methods 0.000 description 10
- 238000004519 manufacturing process Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 6
- 230000009471 action Effects 0.000 description 4
- 230000005856 abnormality Effects 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000013499 data model Methods 0.000 description 2
- 239000011521 glass Substances 0.000 description 2
- 238000007726 management method Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000005192 partition Methods 0.000 description 2
- 229920003023 plastic Polymers 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Abstract
The instrument panel visual identification method, the instrument panel visual identification system and the storage medium based on the AI can realize remote data acquisition and data information anomaly detection; the invention has high recognition precision, has the capability of learning and memorizing dynamic self-service data information, can automatically recognize the data of the current instrument panel through an AI technology, feeds back the data to a control center, solves the problems of manual meter reading and missed detection, can match different recognition behaviors according to different environments so as to improve the recognition accuracy, can judge whether the current recognized instrument data is correct or not through the historical data and the data of the whole operation flow, and increases the user experience while improving the accuracy.
Description
Technical Field
The invention relates to the field of visual identification and intelligent identification, in particular to an AI-based instrument panel visual identification method, an AI-based instrument panel visual identification system and a storage medium.
Background
After the traditional factory is transformed to a digital factory or intelligent manufacturing, data acquisition is needed to be carried out on some industrial instruments in the production process, and when the instruments lack of data capacity, manual field visual detection record is needed, but the problems of missing detection and high false detection rate exist manually, and the detection requirements of modern industry on high speed and high accuracy are difficult to meet.
Accordingly, there is a need for improvement in the art.
Disclosure of Invention
In view of the above problems, the present invention aims to provide an AI-based instrument panel visual identification method, system and storage medium, which can automatically identify the data of the current instrument panel through AI technology and feed back the data to a control center to solve the problems of manual meter reading and missing detection, and can also match different identification behaviors according to different environments so as to improve the accuracy of identification, and can also judge whether the current identified instrument data is correct or not through historical data and the data of the whole operation flow, thereby improving the accuracy and increasing the user experience.
The first aspect of the invention provides an AI-based instrument panel visual identification method, which comprises the following steps:
acquiring image information of an instrument panel;
preprocessing the instrument panel image information to obtain first image information;
identifying the first image information through a preset image identification model to obtain data information;
and sending the data information to a preset terminal for display.
In this scheme, the preprocessing is performed on the dashboard image information to obtain first image information, specifically:
Carrying out graying, binarization, corrosion and expansion treatment on the instrument panel image information;
first image information is obtained.
In this scheme, the generation of the preset image recognition model specifically includes:
preprocessing historical image data to obtain a training data set;
inputting the training data set into an initialized neural network model for training;
obtaining result information of a neural network model;
comparing error rates of the result information;
and if the error rate is smaller than a preset error threshold value, ending training to obtain a preset image recognition model.
In this scheme, still include:
acquiring environmental information;
analyzing according to the environmental information to obtain an influence factor value;
judging whether the influence factor value is larger than a preset influence factor threshold value or not;
if the data is larger than the preset recognition behavior, acquiring the preset recognition behavior, and carrying out data recognition according to the recognition behavior.
In this scheme, obtain preset discernment action, carry out data identification according to discernment action, specifically be:
judging whether the influence factor is a water vapor influence factor or not;
if the water vapor influence factor is larger than a preset water vapor influence factor threshold value;
Acquiring n times of instrument panel image information;
analyzing the n times of instrument panel image information to obtain optimal image identification data;
and taking the optimal image identification data as final data information.
The time interval of the n times of instrument panel image information is a preset time interval.
In this scheme, still include:
acquiring data information in a preset time period to obtain a first numerical value change rate;
acquiring a historical numerical value change rate;
comparing the historical numerical value change rate with a first numerical value change rate to obtain a difference rate;
and if the difference rate is larger than a preset difference rate threshold value, sending warning information.
The second aspect of the present invention provides an AI-based dashboard visual recognition system, including a memory and a processor, where the memory includes an AI-based dashboard visual recognition method program, and the AI-based dashboard visual recognition method program when executed by the processor implements the following steps:
acquiring image information of an instrument panel;
preprocessing the instrument panel image information to obtain first image information;
identifying the first image information through a preset image identification model to obtain data information;
And sending the data information to a preset terminal for display.
In this scheme, the preprocessing is performed on the dashboard image information to obtain first image information, specifically:
carrying out graying, binarization, corrosion and expansion treatment on the instrument panel image information;
first image information is obtained.
In this scheme, the generation of the preset image recognition model specifically includes:
preprocessing historical image data to obtain a training data set;
inputting the training data set into an initialized neural network model for training;
obtaining result information of a neural network model;
comparing error rates of the result information;
and if the error rate is smaller than a preset error threshold value, ending training to obtain a preset image recognition model.
In this scheme, still include:
acquiring environmental information;
analyzing according to the environmental information to obtain an influence factor value;
judging whether the influence factor value is larger than a preset influence factor threshold value or not;
if the data is larger than the preset recognition behavior, acquiring the preset recognition behavior, and carrying out data recognition according to the recognition behavior.
In this scheme, obtain preset discernment action, carry out data identification according to discernment action, specifically be:
Judging whether the influence factor is a water vapor influence factor or not;
if the water vapor influence factor is larger than a preset water vapor influence factor threshold value;
acquiring n times of instrument panel image information;
analyzing the n times of instrument panel image information to obtain optimal image identification data;
and taking the optimal image identification data as final data information.
The time interval of the n times of instrument panel image information is a preset time interval.
In this scheme, still include:
acquiring data information in a preset time period to obtain a first numerical value change rate;
acquiring a historical numerical value change rate;
comparing the historical numerical value change rate with a first numerical value change rate to obtain a difference rate;
and if the difference rate is larger than a preset difference rate threshold value, sending warning information.
A third aspect of the present invention provides a computer-readable storage medium having an AI-based dashboard visual identification method program embodied therein, which when executed by a processor, implements the steps of an AI-based dashboard visual identification method as described in any one of the preceding claims.
According to the instrument panel visual identification method, system and storage medium based on the AI, the data of the current instrument panel can be automatically identified through the AI technology and fed back to the control center, the problems of manual meter reading and missing detection are solved, different identification behaviors can be matched according to different environments, the identification accuracy is improved, whether the current identified instrument data are correct or not can be judged through historical data and the data of the whole operation flow, and the user experience is improved while the accuracy is improved.
Drawings
FIG. 1 shows a flow chart of an AI-based dashboard visual recognition method of the present invention;
FIG. 2 is a diagram showing the current data of the instrument panel of the present invention;
FIG. 3 is a schematic diagram of the instrument panel data of the present invention after being preprocessed;
fig. 4 shows a block diagram of an AI-based dashboard visual recognition system of the present invention.
Detailed Description
In order that the above-recited objects, features and advantages of the present invention will be more clearly understood, a more particular description of the invention will be rendered by reference to the appended drawings and appended detailed description. It should be noted that, in the case of no conflict, the embodiments of the present application and the features in the embodiments may be combined with each other.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention, however, the present invention may be practiced in other ways than those described herein, and therefore the scope of the present invention is not limited to the specific embodiments disclosed below.
Fig. 1 shows a flowchart of an AI-based dashboard visual recognition method of the present invention.
As shown in fig. 1, the invention discloses an AI-based instrument panel visual identification method, which comprises the following steps:
s102, acquiring dashboard image information;
S104, preprocessing the instrument panel image information to obtain first image information;
s106, recognizing the first image information through a preset image recognition model to obtain data information;
s108, the data information is sent to a preset terminal for display.
It should be noted that, the instrument panel may be a digital instrument panel or a pointer instrument panel, the acquisition device in the present invention may be an image recognition device, for example, a camera, and the image acquisition device performs image acquisition on the existing instrument panel, then performs preprocessing on the image information, and sends the image information to a preset image recognition model for recognition, where the preset image recognition model is based on AI (Artificial Intelligence, abbreviated as AI, also referred to as smart machine or machine intelligence, and refers to intelligence represented by a machine manufactured by a person. And outputting data information by a preset image recognition model, wherein the data information is information of a dashboard to be acquired. According to the invention, the instrument is not required to be replaced, the acquisition of instrument data can be completed only by adding the image recognition device of the Internet of things, the instrument panel is recognized by the image recognition technology, the manual omission rate and false detection rate can be reduced, and the automatic detection level is improved.
For example, as shown in fig. 2, fig. 2 is current data of an instrument panel, an on-site industrial instrument panel image is shot through a camera, then a series of preprocessing operations are performed on the image, as shown in fig. 3, fig. 3 shows a schematic diagram of the preprocessed instrument panel data, the preprocessed instrument panel data is sent to a pre-trained digital recognition model, and the data model outputs data information of the output disc image and then uploads the data to a corresponding management system, and the data is stored and is used for personnel to check the instrument panel data.
According to an embodiment of the present invention, the preprocessing is performed on the dashboard image information to obtain first image information, specifically:
carrying out graying, binarization, corrosion and expansion treatment on the instrument panel image information;
first image information is obtained.
It should be noted that, as shown in fig. 3, after the image information of the dashboard is acquired, the image needs to be preprocessed, and then the image needs to be sent to a preset image recognition model for processing. The preprocessing comprises the processes of graying, binarizing, corroding and expanding, and the obtained image information can be used for identifying the neural network. The treatments of graying, binarizing, corroding and expanding are the prior art which can be realized by the person skilled in the art, and the invention is not repeated.
According to an embodiment of the present invention, the generation of the preset image recognition model specifically includes:
preprocessing historical image data to obtain a training data set;
inputting the training data set into an initialized neural network model for training;
obtaining result information of a neural network model;
comparing error rates of the result information;
and if the error rate is smaller than a preset error threshold value, ending training to obtain a preset image recognition model.
The instrument panel data is analyzed by a neural network model. Firstly, preprocessing historical image data, inputting the preprocessed historical image data into a preset neural network model for training, and obtaining a preset image recognition model. The more the historical image data information, the more accurate the training of the neural network. The neural network can also iterate continuously, that is, the neural network iterates automatically as long as data is input, and the result is more and more accurate. After the neural network is trained, a preset image recognition model is obtained, current image information is input into the model, data information is obtained through output, and the obtained data information is real data displayed by an instrument panel.
According to an embodiment of the present invention, further comprising:
acquiring environmental information;
analyzing according to the environmental information to obtain an influence factor value;
judging whether the influence factor value is larger than a preset influence factor threshold value or not;
if the data is larger than the preset recognition behavior, acquiring the preset recognition behavior, and carrying out data recognition according to the recognition behavior.
The influence factor includes various kinds, for example, a water vapor influence factor, a temperature influence factor, an illumination influence factor, and the like. The outermost layer of the instrument panel is made of glass or transparent plastic products, so that light rays can reflect light, water vapor can generate fog, image acquisition is not facilitated on the instrument panel by the image acquisition equipment, different environments can influence reading of the instrument panel, analysis is needed on the environments to obtain influence factor values, the influence factors can be a plurality of influence factors, whether the corresponding influence factors are larger than a preset influence factor threshold value is judged, if so, the acquisition of images is influenced, preset recognition behaviors are needed to be acquired, and image judgment is performed according to the preset recognition behaviors so as to improve the recognition accuracy.
According to the embodiment of the invention, the preset recognition behavior is obtained, and the data recognition is performed according to the recognition behavior, specifically:
judging whether the influence factor is a water vapor influence factor or not;
if the water vapor influence factor is larger than a preset water vapor influence factor threshold value;
acquiring n times of instrument panel image information;
analyzing the n times of instrument panel image information to obtain optimal image identification data;
and taking the optimal image identification data as final data information.
The time interval of the n times of instrument panel image information is a preset time interval.
It should be noted that, when the environmental factor is a water vapor factor, the readings of the instrument panel may be inaccurate, so that other methods are needed to be used for identification. Firstly, acquiring n times of instrument panel image information, wherein the time interval of the n times of instrument panel image information is a preset time interval, that is, the time interval of each acquired instrument panel image information is a time interval, and the time interval is set by a person skilled in the art according to actual needs, and of course, the preset time interval can also be dynamic. After the instrument panel image information is obtained for n times, analysis is performed to obtain optimal image identification data, and the optimal image identification data can be used as final data information.
It should be noted that the preset time interval may also be dynamic, and may be dynamically set through environmental information, specifically:
analyzing according to environmental information of a preset time period to obtain the data change rate of the current influence factor;
and determining a preset time interval according to the data change rate.
It should be noted that, the preset time period may be set by a person skilled in the art according to actual needs, and environmental data in the preset time period is obtained, so that a certain change rule may be obtained. For example, acquiring preset water vapor data within 5 hours, obtaining a law of water vapor change, and then determining a time interval for acquiring images according to the law of change. For example, if the change in the water vapor data is found to be weaker, it may be collected 1 time for the first 30 minutes, 1 time for 100 minutes, 1 time for 180 minutes, 1 time for 250 minutes, 1 time for 280 minutes, 1 time for 290 minutes, and 1 time for 295 minutes. The later data are more accurate due to the fact that the influence of water vapor is smaller, so that the later acquisition frequency is improved, and the accuracy of the whole acquisition is improved.
It should be noted that, the optimal image recognition data may be obtained by preprocessing the image information of each instrument panel and then inputting the preprocessed image information into a preset image recognition model to obtain a plurality of recognized data; and comparing the difference value of each identified data, selecting the data with a preset difference value range, and taking an average value to obtain the optimal image identification data.
It should be noted that, the optimal image recognition data may be optimal data obtained after multiple model analysis, and the optimal image is obtained from multiple acquired images, so that the accuracy is higher. Firstly, carrying out data preprocessing on the image information of the instrument panel collected each time to obtain data which can be analyzed and identified by a model; then, inputting the data into an image recognition model to obtain a plurality of recognized data; and comparing the data to obtain a difference value between each two data, if the difference value is too large, indicating that the data deviation is large and the data accuracy is poor, selecting a plurality of similar difference values for analysis, and carrying out average calculation on the data to obtain an optimal data value, namely the optimal image identification data. The preset difference range is set by a person skilled in the art according to actual needs, for example, the number of acquired data is 100, the preset difference range is 10, wherein only 65 data differences are relatively close to each other and are within 10, and 65 data are selected for analysis to obtain an average value, namely optimal image identification data.
It should be noted that the present invention further includes:
if the influence factor is an illumination factor;
the subarea light intensity data in the panel board image information is obtained;
And if the partition light intensity data is larger than the preset light intensity threshold value, sending predicted data to a preset terminal, and displaying according to a preset display mode.
It should be noted that, if there are different directions of illumination, the dashboard may reflect light, and if the light-reflecting position is just in the meter pointer or the meter digital region, the collected data may be affected. At this time, the position of the reflection in the image needs to be judged, the regional light intensity data in the acquired image information is calculated, wherein the regional light intensity data is large, the position is precisely the reflection position, effective data identification cannot be performed, the data which cannot be identified in the region are analyzed, the data value is predicted, the predicted data value is sent to a background or a user side, the data is displayed as a predicted value when the predicted value is displayed, and workers are prevented from misunderstanding the predicted value as an accurate value. After the illumination direction is changed, the analysis and the identification of accurate values are performed. In this case, when data prediction is performed, the data and the change trend of the pointer can be used to determine through the data outside the reflective area, so as to obtain predicted data. For example, the upper part of the last 1 bit of the three-bit data 861 of the digital display tube cannot be correctly identified due to reflection, and the trend of the data change is slowly increased, and the predicted value may be 1, 4 or 7 according to the trend of the data change and the lower part without reflection. The specific manner of prediction varies from meter display to meter display.
According to an embodiment of the present invention, further comprising:
acquiring data information in a preset time period to obtain a first numerical value change rate;
acquiring a historical numerical value change rate;
comparing the historical numerical value change rate with a first numerical value change rate to obtain a difference rate;
and if the difference rate is larger than a preset difference rate threshold value, sending warning information.
It should be noted that the invention can also judge the accuracy rate of the acquired data, and can also alarm if the abnormality and deviation of the data occur. Firstly, acquiring data information in a preset time period to obtain a first numerical value change rate; and obtaining a historical numerical value change rate, and comparing the historical numerical value change rate with the historical numerical value change rate to obtain a difference rate, wherein the difference rate refers to the difference between a rate change curve and a historical rate curve. If the difference is too large, the possibly collected data is inaccurate, and warning information is sent to the user side or the background. The difference rate threshold may be set by those skilled in the art according to actual needs, for example, the difference rate threshold is 8%.
It should be noted that the present invention further includes:
acquiring other equipment information of a preset area;
acquiring environmental information acquired by other equipment;
Comparing the environmental information acquired by other equipment to obtain second environmental information;
judging whether the second environment information is a proper environment or not;
if not, sending warning information.
The invention can also judge whether the instrument panel image data is suitable for acquisition according to the environmental information acquired by other equipment. Firstly, other equipment information of a preset area is acquired, wherein the preset area can be a workshop area, and other equipment can be other operation equipment in a workshop or other image acquisition equipment. After other devices are determined, the environmental information they collect can be obtained. Comparing the acquired environmental information to obtain second environmental information; wherein the second environmental information may comprise water vapor influence information, temperature influence information, illumination influence information, etc. And judging whether the second environment information is a proper environment or not, if not, indicating that the current environment is not suitable for image acquisition, or indicating that the information error of image acquisition is relatively large, and sending warning information to a background or a user side.
It should be noted that the present invention further includes:
Acquiring a working node where a current instrument panel is located, and acquiring current working node information;
acquiring data information of the following working node information to obtain second data information;
judging whether the second data information contradicts the data information of the current working node or not;
if the conflict exists, sending warning information.
It should be noted that, a plurality of working nodes are often present in the current industrial workshop on a production line, and each node needs to perform instrument data acquisition. For example, when the data collected by the current manufacturing node is 80, and after the corresponding data of the following working node is obtained, the data 80 is not in a reasonable data range, which indicates that there may be a contradiction between the data, and then warning information is sent to remind.
It should be noted that the present invention further includes:
acquiring position information of acquisition equipment and the instrument panel;
analyzing according to the environment information, the position information and the acquired instrument panel image information to obtain optimal position information;
And sending the optimal position information to a preset terminal.
The invention can also assist the staff to place the image recognition device, change the image recognition device to the optimal acquisition position and improve the user experience. Firstly, acquiring position information of the acquisition equipment and the instrument panel, and then analyzing according to the environment information, the position information and the acquired instrument panel image information to obtain the optimal instrument panel position information. The cloud computing of the cloud server can be used for analysis, and the AI algorithm can be introduced for analysis, so that a person skilled in the art can use a corresponding technical means for analysis of the optimal position, and the invention is not repeated. After the optimal position information is obtained, the optimal position information is sent to a background or a user side so as to prompt a worker to change the position of the camera. By the method, the acquired data can be more accurate, the user is indirectly guided to put the camera, and the use experience is improved.
Fig. 4 shows a block diagram of an AI-based dashboard visual recognition system of the present invention.
As shown in fig. 4, the present invention shows an AI-based dashboard visual recognition system 4, including a memory 41 and a processor 42, wherein the memory includes an AI-based dashboard visual recognition method program, and the AI-based dashboard visual recognition method program when executed by the processor implements the following steps:
Acquiring image information of an instrument panel;
preprocessing the instrument panel image information to obtain first image information;
identifying the first image information through a preset image identification model to obtain data information;
and sending the data information to a preset terminal for display.
It should be noted that, the instrument panel may be a digital instrument panel or a pointer instrument panel, the acquisition device in the present invention may be an image recognition device, for example, a camera, and the image acquisition device performs image acquisition on the existing instrument panel, then performs preprocessing on the image information, and sends the image information to a preset image recognition model for recognition, where the preset image recognition model is based on AI (Artificial Intelligence, abbreviated as AI, also referred to as smart machine or machine intelligence, and refers to intelligence represented by a machine manufactured by a person. And outputting data information by a preset image recognition model, wherein the data information is information of a dashboard to be acquired. According to the invention, the instrument is not required to be replaced, the acquisition of instrument data can be completed only by adding the image recognition device of the Internet of things, the instrument panel is recognized by the image recognition technology, the manual omission rate and false detection rate can be reduced, and the automatic detection level is improved.
For example, as shown in fig. 2, fig. 2 is current data of an instrument panel, an on-site industrial instrument panel image is shot through a camera, then a series of preprocessing operations are performed on the image, as shown in fig. 3, fig. 3 shows a schematic diagram of the preprocessed instrument panel data, the preprocessed instrument panel data is sent to a pre-trained digital recognition model, and the data model outputs data information of the output disc image and then uploads the data to a corresponding management system, and the data is stored and is used for personnel to check the instrument panel data.
According to an embodiment of the present invention, the preprocessing is performed on the dashboard image information to obtain first image information, specifically:
carrying out graying, binarization, corrosion and expansion treatment on the instrument panel image information;
first image information is obtained.
It should be noted that, as shown in fig. 3, after the image information of the dashboard is acquired, the image needs to be preprocessed, and then the image needs to be sent to a preset image recognition model for processing. The preprocessing comprises the processes of graying, binarizing, corroding and expanding, and the obtained image information can be used for identifying the neural network. The treatments of graying, binarizing, corroding and expanding are the prior art which can be realized by the person skilled in the art, and the invention is not repeated.
According to an embodiment of the present invention, the generation of the preset image recognition model specifically includes:
preprocessing historical image data to obtain a training data set;
inputting the training data set into an initialized neural network model for training;
obtaining result information of a neural network model;
comparing error rates of the result information;
and if the error rate is smaller than a preset error threshold value, ending training to obtain a preset image recognition model.
The instrument panel data is analyzed by a neural network model. Firstly, preprocessing historical image data, inputting the preprocessed historical image data into a preset neural network model for training, and obtaining a preset image recognition model. The more the historical image data information, the more accurate the training of the neural network. The neural network can also iterate continuously, that is, the neural network iterates automatically as long as data is input, and the result is more and more accurate. After the neural network is trained, a preset image recognition model is obtained, current image information is input into the model, data information is obtained through output, and the obtained data information is real data displayed by an instrument panel.
According to an embodiment of the present invention, further comprising:
acquiring environmental information;
analyzing according to the environmental information to obtain an influence factor value;
judging whether the influence factor value is larger than a preset influence factor threshold value or not;
if the data is larger than the preset recognition behavior, acquiring the preset recognition behavior, and carrying out data recognition according to the recognition behavior.
The influence factor includes various kinds, for example, a water vapor influence factor, a temperature influence factor, an illumination influence factor, and the like. The outermost layer of the instrument panel is made of glass or transparent plastic products, so that light rays can reflect light, water vapor can generate fog, image acquisition is not facilitated on the instrument panel by the image acquisition equipment, different environments can influence reading of the instrument panel, analysis is needed on the environments to obtain influence factor values, the influence factors can be a plurality of influence factors, whether the corresponding influence factors are larger than a preset influence factor threshold value is judged, if so, the acquisition of images is influenced, preset recognition behaviors are needed to be acquired, and image judgment is performed according to the preset recognition behaviors so as to improve the recognition accuracy.
According to the embodiment of the invention, the preset recognition behavior is obtained, and the data recognition is performed according to the recognition behavior, specifically:
judging whether the influence factor is a water vapor influence factor or not;
if the water vapor influence factor is larger than a preset water vapor influence factor threshold value;
acquiring n times of instrument panel image information;
analyzing the n times of instrument panel image information to obtain optimal image identification data;
and taking the optimal image identification data as final data information.
The time interval of the n times of instrument panel image information is a preset time interval.
It should be noted that, when the environmental factor is a water vapor factor, the readings of the instrument panel may be inaccurate, so that other methods are needed to be used for identification. Firstly, acquiring n times of instrument panel image information, wherein the time interval of the n times of instrument panel image information is a preset time interval, that is, the time interval of each acquired instrument panel image information is a time interval, and the time interval is set by a person skilled in the art according to actual needs, and of course, the preset time interval can also be dynamic. After the instrument panel image information is obtained for n times, analysis is performed to obtain optimal image identification data, and the optimal image identification data can be used as final data information.
It should be noted that the preset time interval may also be dynamic, and may be dynamically set through environmental information, specifically:
analyzing according to environmental information of a preset time period to obtain the data change rate of the current influence factor;
and determining a preset time interval according to the data change rate.
It should be noted that, the preset time period may be set by a person skilled in the art according to actual needs, and environmental data in the preset time period is obtained, so that a certain change rule may be obtained. For example, acquiring preset water vapor data within 5 hours, obtaining a law of water vapor change, and then determining a time interval for acquiring images according to the law of change. For example, if the change in the water vapor data is found to be weaker, it may be collected 1 time for the first 30 minutes, 1 time for 100 minutes, 1 time for 180 minutes, 1 time for 250 minutes, 1 time for 280 minutes, 1 time for 290 minutes, and 1 time for 295 minutes. The later data are more accurate due to the fact that the influence of water vapor is smaller, so that the later acquisition frequency is improved, and the accuracy of the whole acquisition is improved.
It should be noted that, the optimal image recognition data may be obtained by preprocessing the image information of each instrument panel and then inputting the preprocessed image information into a preset image recognition model to obtain a plurality of recognized data; and comparing the difference value of each identified data, selecting the data with a preset difference value range, and taking an average value to obtain the optimal image identification data.
It should be noted that, the optimal image recognition data may be optimal data obtained after multiple model analysis, and the optimal image is obtained from multiple acquired images, so that the accuracy is higher. Firstly, carrying out data preprocessing on the image information of the instrument panel collected each time to obtain data which can be analyzed and identified by a model; then, inputting the data into an image recognition model to obtain a plurality of recognized data; and comparing the data to obtain a difference value between each two data, if the difference value is too large, indicating that the data deviation is large and the data accuracy is poor, selecting a plurality of similar difference values for analysis, and carrying out average calculation on the data to obtain an optimal data value, namely the optimal image identification data. The preset difference range is set by a person skilled in the art according to actual needs, for example, the number of acquired data is 100, the preset difference range is 10, wherein only 65 data differences are relatively close to each other and are within 10, and 65 data are selected for analysis to obtain an average value, namely optimal image identification data.
It should be noted that the present invention further includes:
if the influence factor is an illumination factor;
the subarea light intensity data in the panel board image information is obtained;
And if the partition light intensity data is larger than the preset light intensity threshold value, sending predicted data to a preset terminal, and displaying according to a preset display mode.
It should be noted that, if there are different directions of illumination, the dashboard may reflect light, and if the light-reflecting position is just in the meter pointer or the meter digital region, the collected data may be affected. At this time, the position of the reflection in the image needs to be judged, the regional light intensity data in the acquired image information is calculated, wherein the regional light intensity data is large, the position is precisely the reflection position, effective data identification cannot be performed, the data which cannot be identified in the region are analyzed, the data value is predicted, the predicted data value is sent to a background or a user side, the data is displayed as a predicted value when the predicted value is displayed, and workers are prevented from misunderstanding the predicted value as an accurate value. After the illumination direction is changed, the analysis and the identification of accurate values are performed. In this case, when data prediction is performed, the data and the change trend of the pointer can be used to determine through the data outside the reflective area, so as to obtain predicted data. For example, the upper part of the last 1 bit of the three-bit data 861 of the digital display tube cannot be correctly identified due to reflection, and the trend of the data change is slowly increased, and the predicted value may be 1, 4 or 7 according to the trend of the data change and the lower part without reflection. The specific manner of prediction varies from meter display to meter display.
According to an embodiment of the present invention, further comprising:
acquiring data information in a preset time period to obtain a first numerical value change rate;
acquiring a historical numerical value change rate;
comparing the historical numerical value change rate with a first numerical value change rate to obtain a difference rate;
and if the difference rate is larger than a preset difference rate threshold value, sending warning information.
It should be noted that the invention can also judge the accuracy rate of the acquired data, and can also alarm if the abnormality and deviation of the data occur. Firstly, acquiring data information in a preset time period to obtain a first numerical value change rate; and obtaining a historical numerical value change rate, and comparing the historical numerical value change rate with the historical numerical value change rate to obtain a difference rate, wherein the difference rate refers to the difference between a rate change curve and a historical rate curve. If the difference is too large, the possibly collected data is inaccurate, and warning information is sent to the user side or the background. The difference rate threshold may be set by those skilled in the art according to actual needs, for example, the difference rate threshold is 8%.
It should be noted that the present invention further includes:
acquiring other equipment information of a preset area;
acquiring environmental information acquired by other equipment;
Comparing the environmental information acquired by other equipment to obtain second environmental information;
judging whether the second environment information is a proper environment or not;
if not, sending warning information.
The invention can also judge whether the instrument panel image data is suitable for acquisition according to the environmental information acquired by other equipment. Firstly, other equipment information of a preset area is acquired, wherein the preset area can be a workshop area, and other equipment can be other operation equipment in a workshop or other image acquisition equipment. After other devices are determined, the environmental information they collect can be obtained. Comparing the acquired environmental information to obtain second environmental information; wherein the second environmental information may comprise water vapor influence information, temperature influence information, illumination influence information, etc. And judging whether the second environment information is a proper environment or not, if not, indicating that the current environment is not suitable for image acquisition, or indicating that the information error of image acquisition is relatively large, and sending warning information to a background or a user side.
It should be noted that the present invention further includes:
Acquiring a working node where a current instrument panel is located, and acquiring current working node information;
acquiring data information of the following working node information to obtain second data information;
judging whether the second data information contradicts the data information of the current working node or not;
if the conflict exists, sending warning information.
It should be noted that, a plurality of working nodes are often present in the current industrial workshop on a production line, and each node needs to perform instrument data acquisition. For example, when the data collected by the current manufacturing node is 80, and after the corresponding data of the following working node is obtained, the data 80 is not in a reasonable data range, which indicates that there may be a contradiction between the data, and then warning information is sent to remind.
It should be noted that the present invention further includes:
acquiring position information of acquisition equipment and the instrument panel;
analyzing according to the environment information, the position information and the acquired instrument panel image information to obtain optimal position information;
And sending the optimal position information to a preset terminal.
The invention can also assist the staff to place the image recognition device, change the image recognition device to the optimal acquisition position and improve the user experience. Firstly, acquiring position information of the acquisition equipment and the instrument panel, and then analyzing according to the environment information, the position information and the acquired instrument panel image information to obtain the optimal instrument panel position information. The cloud computing of the cloud server can be used for analysis, and the AI algorithm can be introduced for analysis, so that a person skilled in the art can use a corresponding technical means for analysis of the optimal position, and the invention is not repeated. After the optimal position information is obtained, the optimal position information is sent to a background or a user side so as to prompt a worker to change the position of the camera. By the method, the acquired data can be more accurate, the user is indirectly guided to put the camera, and the use experience is improved.
A third aspect of the present invention provides a computer-readable storage medium having an AI-based dashboard visual identification method program embodied therein, which when executed by a processor, implements the steps of an AI-based dashboard visual identification method as described in any one of the preceding claims.
According to the instrument panel visual identification method, system and storage medium based on the AI, the data of the current instrument panel can be automatically identified through the AI technology and fed back to the control center, the problems of manual meter reading and missing detection are solved, different identification behaviors can be matched according to different environments, the identification accuracy is improved, whether the current identified instrument data are correct or not can be judged through historical data and the data of the whole operation flow, and the user experience is improved while the accuracy is improved. The invention can realize remote data acquisition and abnormal detection of data information; the recognition accuracy is high, and the dynamic self-help data information learning and memory capacity is provided; the invention can improve the accuracy of data reading of the industrial instrument panel in the manufacturing industry factory, meets the detection requirements of high speed and high accuracy in the modern industry, is beneficial to avoiding the phenomena of easy missed detection and high false detection rate, replaces the manual meter reading mode, and is convenient for large-scale data acquisition by the automatic reading mode of the industrial instrument panel.
In the several embodiments provided in this application, it should be understood that the disclosed apparatus and method may be implemented in other ways. The above described device embodiments are only illustrative, e.g. the division of the units is only one logical function division, and there may be other divisions in practice, such as: multiple units or components may be combined or may be integrated into another system, or some features may be omitted, or not performed. In addition, the various components shown or discussed may be coupled or directly coupled or communicatively coupled to each other via some interface, whether indirectly coupled or communicatively coupled to devices or units, whether electrically, mechanically, or otherwise.
The units described above as separate components may or may not be physically separate, and components shown as units may or may not be physical units; can be located in one place or distributed to a plurality of network units; some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
In addition, each functional unit in each embodiment of the present invention may be integrated in one processing unit, or each unit may be separately used as one unit, or two or more units may be integrated in one unit; the integrated units may be implemented in hardware or in hardware plus software functional units.
Those of ordinary skill in the art will appreciate that: all or part of the steps for implementing the above method embodiments may be implemented by hardware related to program instructions, and the foregoing program may be stored in a computer readable storage medium, where the program, when executed, performs steps including the above method embodiments; and the aforementioned storage medium includes: a mobile storage device, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk or an optical disk, or the like, which can store program codes.
Alternatively, the above-described integrated units of the present invention may be stored in a computer-readable storage medium if implemented in the form of software functional modules and sold or used as separate products. Based on such understanding, the technical solutions of the embodiments of the present invention may be embodied in essence or a part contributing to the prior art in the form of a software product 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 methods described in the embodiments of the present invention. And the aforementioned storage medium includes: a removable storage device, ROM, RAM, magnetic or optical disk, or other medium capable of storing program code.
Claims (10)
1. An AI-based dashboard visual identification method, comprising:
acquiring image information of an instrument panel;
preprocessing the instrument panel image information to obtain first image information;
identifying the first image information through a preset image identification model to obtain data information;
and sending the data information to a preset terminal for display.
2. The AI-based dashboard visual recognition method of claim 1, wherein the preprocessing the dashboard image information to obtain first image information specifically comprises:
carrying out graying, binarization, corrosion and expansion treatment on the instrument panel image information;
first image information is obtained.
3. The AI-based dashboard visual recognition method of claim 1, wherein the generation of the preset image recognition model is specifically:
preprocessing historical image data to obtain a training data set;
inputting the training data set into an initialized neural network model for training;
obtaining result information of a neural network model;
comparing error rates of the result information;
and if the error rate is smaller than a preset error threshold value, ending training to obtain a preset image recognition model.
4. The AI-based dashboard visual identification method of claim 1, further comprising:
acquiring environmental information;
analyzing according to the environmental information to obtain an influence factor value;
judging whether the influence factor value is larger than a preset influence factor threshold value or not;
If the data is larger than the preset recognition behavior, acquiring the preset recognition behavior, and carrying out data recognition according to the recognition behavior.
5. The AI-based instrument panel visual recognition method of claim 4, wherein the acquiring a preset recognition behavior, and performing data recognition according to the recognition behavior, specifically comprises:
judging whether the influence factor is a water vapor influence factor or not;
if the water vapor influence factor is larger than a preset water vapor influence factor threshold value;
acquiring n times of instrument panel image information;
analyzing the n times of instrument panel image information to obtain optimal image identification data;
taking the optimal image identification data as final data information;
the time interval of the n times of instrument panel image information is a preset time interval.
6. The AI-based dashboard visual identification method of claim 1, further comprising:
acquiring data information in a preset time period to obtain a first numerical value change rate;
acquiring a historical numerical value change rate;
comparing the historical numerical value change rate with a first numerical value change rate to obtain a difference rate;
and if the difference rate is larger than a preset difference rate threshold value, sending warning information.
7. An AI-based dashboard visual identification system, comprising a memory and a processor, wherein the memory comprises an AI-based dashboard visual identification method program, and the AI-based dashboard visual identification method program when executed by the processor implements the steps of:
acquiring image information of an instrument panel;
preprocessing the instrument panel image information to obtain first image information;
identifying the first image information through a preset image identification model to obtain data information;
and sending the data information to a preset terminal for display.
8. The AI-based dashboard visual identification system of claim 7, further comprising:
acquiring environmental information;
analyzing according to the environmental information to obtain an influence factor value;
judging whether the influence factor value is larger than a preset influence factor threshold value or not;
if the data is larger than the preset recognition behavior, acquiring the preset recognition behavior, and carrying out data recognition according to the recognition behavior.
9. The AI-based dashboard visual recognition system of claim 8, wherein the acquiring a preset recognition behavior, and performing data recognition according to the recognition behavior, specifically comprises:
Judging whether the influence factor is a water vapor influence factor or not;
if the water vapor influence factor is larger than a preset water vapor influence factor threshold value;
acquiring n times of instrument panel image information;
analyzing the n times of instrument panel image information to obtain optimal image identification data;
taking the optimal image identification data as final data information;
the time interval of the n times of instrument panel image information is a preset time interval.
10. A computer readable storage medium, wherein the computer readable storage medium comprises an AI-based dashboard visual identification method program, and the AI-based dashboard visual identification method program, when executed by a processor, implements the steps of an AI-based dashboard visual identification method according to any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310531877.2A CN116563631A (en) | 2023-05-12 | 2023-05-12 | AI-based instrument panel visual identification method, system and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310531877.2A CN116563631A (en) | 2023-05-12 | 2023-05-12 | AI-based instrument panel visual identification method, system and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116563631A true CN116563631A (en) | 2023-08-08 |
Family
ID=87491218
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310531877.2A Pending CN116563631A (en) | 2023-05-12 | 2023-05-12 | AI-based instrument panel visual identification method, system and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116563631A (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102063622A (en) * | 2010-12-27 | 2011-05-18 | 天津家宇科技发展有限公司 | Two-dimensional barcode image binarization method based on wavelet and OTSU method |
| CN104748678A (en) * | 2015-03-08 | 2015-07-01 | 大连理工大学 | Method of compensating image quality during high-temperature object measurement |
| CN107701236A (en) * | 2017-10-13 | 2018-02-16 | 鄂尔多斯市营盘壕煤炭有限公司 | A kind of multiple information data monitoring system of skip loading pocket |
| CN110717462A (en) * | 2019-10-12 | 2020-01-21 | 上海市建筑科学研究院 | Digital instrument reading identification method, device, equipment and medium |
| CN111931620A (en) * | 2020-07-31 | 2020-11-13 | 北京奈伦机器人科技有限公司 | Instrument panel positioning and identifying method and device, electronic equipment and storage medium |
| CN112115893A (en) * | 2020-09-24 | 2020-12-22 | 深圳市赛为智能股份有限公司 | Instrument panel pointer reading identification method and device, computer equipment and storage medium |
| CN113758514A (en) * | 2021-07-29 | 2021-12-07 | 江铃汽车股份有限公司 | Automobile instrument automatic testing method and system, readable storage medium and vehicle |
| CN115933561A (en) * | 2022-12-16 | 2023-04-07 | 青岛鹏海软件有限公司 | Data acquisition method and system of factory automation equipment and electronic equipment |
| CN116046051A (en) * | 2022-09-30 | 2023-05-02 | 青岛鹏海软件有限公司 | Intelligent motor management method and system and electronic equipment thereof |
-
2023
- 2023-05-12 CN CN202310531877.2A patent/CN116563631A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102063622A (en) * | 2010-12-27 | 2011-05-18 | 天津家宇科技发展有限公司 | Two-dimensional barcode image binarization method based on wavelet and OTSU method |
| CN104748678A (en) * | 2015-03-08 | 2015-07-01 | 大连理工大学 | Method of compensating image quality during high-temperature object measurement |
| CN107701236A (en) * | 2017-10-13 | 2018-02-16 | 鄂尔多斯市营盘壕煤炭有限公司 | A kind of multiple information data monitoring system of skip loading pocket |
| CN110717462A (en) * | 2019-10-12 | 2020-01-21 | 上海市建筑科学研究院 | Digital instrument reading identification method, device, equipment and medium |
| CN111931620A (en) * | 2020-07-31 | 2020-11-13 | 北京奈伦机器人科技有限公司 | Instrument panel positioning and identifying method and device, electronic equipment and storage medium |
| CN112115893A (en) * | 2020-09-24 | 2020-12-22 | 深圳市赛为智能股份有限公司 | Instrument panel pointer reading identification method and device, computer equipment and storage medium |
| CN113758514A (en) * | 2021-07-29 | 2021-12-07 | 江铃汽车股份有限公司 | Automobile instrument automatic testing method and system, readable storage medium and vehicle |
| CN116046051A (en) * | 2022-09-30 | 2023-05-02 | 青岛鹏海软件有限公司 | Intelligent motor management method and system and electronic equipment thereof |
| CN115933561A (en) * | 2022-12-16 | 2023-04-07 | 青岛鹏海软件有限公司 | Data acquisition method and system of factory automation equipment and electronic equipment |
Non-Patent Citations (3)
| Title |
|---|
| 王蜂, 邓方林, 雷万云: "一种改善坦克图像质量提高识别能力的方法", 弹箭与制导学报, no. 03 * |
| 钟大伟，高铎，王鹏，宋超著: "《数据应用工程 方法论与实践》", 机械工业出版社, pages: 157 * |
| 顾理, 庄镇泉, 万淑超, 蔡伟: "手形识别中的手形提取方法", 计算机仿真, no. 07 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111401609B (en) | Prediction method and prediction device for traffic flow time series | |
| CN112115927B (en) | Intelligent machine room equipment identification method and system based on deep learning | |
| CN111950812B (en) | Method and device for automatically identifying and predicting rainfall | |
| CN108776694A (en) | A kind of time series abnormal point detecting method and device | |
| CN110059894B (en) | Equipment state evaluation method, device, system and storage medium | |
| CN114266944B (en) | Rapid model training result checking system | |
| CN112288021A (en) | Medical wastewater monitoring data quality control method, device and system | |
| CN112183906B (en) | Machine room environment prediction method and system based on multi-model combined model | |
| CN112613569A (en) | Image recognition method, and training method and device of image classification model | |
| CN115471487A (en) | Insulator defect detection model construction and insulator defect detection method and device | |
| CN113672600A (en) | Anomaly detection method and system | |
| CN110738272B (en) | Method for labeling visual mechanical continuous alarm samples of power transmission line channel | |
| CN109242165A (en) | A kind of model training and prediction technique and device based on model training | |
| CN117372424B (en) | Defect detection method, device, equipment and storage medium | |
| CN109211564B (en) | Self-adaptive threshold detection method for health assessment of ball screw pair | |
| CN116932523B (en) | Platform for integrating and supervising third party environment detection mechanism | |
| CN116563631A (en) | AI-based instrument panel visual identification method, system and storage medium | |
| CN114593898B (en) | Lens quality analysis method, device, equipment and medium based on refraction data | |
| CN115965625A (en) | Instrument detection device based on visual identification and detection method thereof | |
| CN115293735A (en) | Unmanned factory industrial internet platform monitoring management method and system | |
| CN110455370B (en) | Flood-prevention drought-resisting remote measuring display system | |
| CN113268525A (en) | Cognitive assessment method and system for automatically optimizing norm | |
| CN114510994A (en) | Training method and device for urban underground gas leakage recognition model | |
| CN112632469A (en) | Method and device for detecting abnormity of business transaction data and computer equipment | |
| CN113378921A (en) | Data screening method and device and electronic equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |