A home curtain defect detection system based on infrared light

Abstract

The present invention discloses a household curtain defect detection system based on infrared light, which relates to the field of infrared technology and solves the technical problem that the existing technology does not consider that due to different curtain materials, the light source of the light box net may not fully penetrate the curtain, which will lead to the inability to accurately evaluate the curtain quality; the present invention uses infrared technology to detect the curtain quality through the curtain warehouse; analyzes the surface temperature difference of the curtain through the infrared transmitter and receiver, and uses the infrared camera to scan the internal defects if they exist; otherwise, external abnormality detection is performed; the image acquisition device captures the external abnormality of the curtain; the infrared camera converts the infrared radiation into an electrical signal, processes and obtains the internal image, and identifies the material unevenness, damage and other abnormalities; these abnormal area data constitute a data set for defect identification and classification. According to the identification results, the stained curtain is cleaned or marked as a defective product; at the same time, the qualified rate of the current batch of curtains is predicted using historical data, and early warning and database information are provided to the management personnel.

Description

Technical Field

[0001] The invention belongs to the field of infrared technology and relates to infrared imaging technology, in particular to a household curtain defect detection system based on infrared light. Background Art

[0002] Home curtains go through multiple processing links during the production process. If there are problems in some links, it may cause defects in the appearance or function of the curtains. For example, defects such as asymmetric patterns, uneven sewing, and uneven fabrics; home curtains may have defects such as wear, tear, and derailment during long-term use. These defects not only affect the appearance, but also the functionality and service life of the curtains; infrared thermal imagers are used to detect the infrared energy radiated by various parts of the curtains. Since infrared light is very sensitive to slight changes in the surface of the material, it can detect imperceptible defects on the curtains, such as cracks, damage, stains, etc., and infrared light detection is a non-destructive testing method that will not cause any physical or chemical damage to the curtain material to ensure that the quality and performance of the curtains meet the expectations of consumers. In summary, infrared light defect detection for home curtains is an important measure to ensure the quality of curtains and protect the health and safety of consumers. Through comprehensive testing and evaluation, problems with curtains can be discovered and solved in a timely manner to improve the quality and competitiveness of products.

[0003] The prior art (invention patent with application number CN110231348B) discloses a curtain quality detection platform and detection method, which comprises the following steps: hanging one end of the curtain in a curtain hanging slot, operating a signal transmitter button, and using a power device to lift one end of the curtain; when the curtain hanging slot moves to the position of the limit plate, the photoelectric switch is triggered, and the signal is transmitted to the controller, and the controller controls the motor to stop running; the light box net is powered on, and the size of the curtain is directly read out using the scale set on the light box net, and the curtain is detected to have holes or defects by means of the brightness of the light source of the light box net; after the measurement and detection are completed, the data is recorded; the signal transmitter is operated again, and the curtain is lowered to a suitable position by means of the power device, and the curtain is removed; wherein, by hanging the curtain for detection, a light box net is also used, and an LED lamp tube is arranged in the light box net. After the curtain is hung up, the quality of the curtain can be directly detected by irradiating the light source of the light box net; however, the prior art detects whether the curtain has holes or defects by means of the brightness of the light source of the light box net, but does not consider that due to the different materials of the curtain, the light source of the light box net may not be able to fully penetrate the curtain, which will result in the inability to accurately evaluate the quality of the curtain.

[0004] The present invention provides a household curtain defect detection system based on infrared light to solve the above technical problems. Summary of the invention

[0005] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes a household curtain defect detection system based on infrared light, which is used to solve the technical problem that the prior art detects whether the curtain has holes or defects by relying on the brightness of the light source of the light box net, but does not consider that due to the different materials of the curtain, the light source of the light box net may not be able to fully penetrate the curtain, resulting in the inability to accurately evaluate the quality of the curtain.

[0006] To achieve the above-mentioned object, a first aspect of the present invention provides a household curtain defect detection system based on infrared light, comprising: an infrared detection module, an image acquisition and processing module, and a defect processing and prediction module;

[0007] Infrared detection module: An infrared transmitter is set in the curtain warehouse to emit infrared light signals to the curtain surface through the infrared transmitter; an infrared receiver is set to obtain the infrared radiation intensity at different positions on the object surface through the infrared receiver; a temperature distribution data graph of the curtain surface is constructed according to the infrared radiation intensity, and whether there is a temperature difference on the curtain surface is determined according to the temperature distribution data graph. If yes, an infrared camera is used to scan and identify internal defects of the curtain; if no, an external abnormality detection of the curtain is performed;

[0008] Image acquisition and processing module: obtain the external image of the curtain through the image acquisition device, pre-process the external image of the curtain, and obtain the external abnormal area of ​​the curtain; wherein the external abnormal area includes: the texture, color difference and stains of the curtain surface;

[0009] An infrared camera is set up in the curtain warehouse. The infrared radiation received is converted into an electrical signal through an infrared sensor, and the electrical signal is pre-processed to obtain the internal image of the curtain. Based on the image processing algorithm, the abnormal area inside the curtain is obtained. Among them, the internal abnormal area includes: material unevenness and fabric damage, snagging, and scratches. The abnormal area outside the curtain and the abnormal area inside the curtain constitute the abnormal area image data set.

[0010] Defect processing and prediction module: retrieve the abnormal area image feature library, automatically identify and classify the abnormal area of ​​the curtain based on image recognition technology; obtain the classified abnormal area data of the curtain through the defect recognition module; determine whether the defect type is a stain on the surface of the curtain, if yes, clean the curtain; if not, mark the curtain as a defective product; obtain the qualified rate of the curtains of the historical batches, and predict the qualified rate of the curtains of this batch based on the qualified rate of the curtains of the historical batches and the image data set of the abnormal area; form a database and early warning information through the qualified rate prediction, and display the database and early warning information to the production management personnel;

[0011] The method of predicting the qualified rate of a batch of curtains based on the qualified rate of a historical batch of curtains and the abnormal area image dataset includes:

[0012] Retrieve the abnormal area image data set and divide it into N parts, where N is a positive integer greater than 1; obtain the sampling sample pass rate through sampling detection, retrieve the historical batch curtain pass rate from the database, and mark the sampling sample pass rate and the historical batch curtain pass rate as SQ and HQ respectively;

[0013] The predicted value of the current batch pass rate is calculated using the formula QF=SQ×α+HQ×β, where α and β are non-zero weight coefficients.

[0014] Preferably, the step of constructing a temperature distribution data graph of the curtain surface according to the infrared radiation intensity includes:

[0015] The infrared radiation intensity at different positions on the surface of the object is obtained through an infrared receiver, and the intensity value is converted into different colors or brightness in the image; the different colors and brightness are integrated to obtain the temperature distribution data map of the curtain surface.

[0016] It should be noted that when the target object is close to the infrared detection module, the infrared receiver will receive a stronger infrared signal; when the target object is far away, the received infrared signal will become weaker. When detecting the curtains, the curtains need to be placed in the preset position.

[0017] The present invention obtains the infrared radiation intensity at different positions on the surface of an object through an infrared receiver, can quickly scan the entire surface of the curtain, and generate a temperature distribution map in a short time, greatly improving the measurement efficiency; the infrared radiation intensity is converted into color and brightness in the image, so that the temperature distribution data is presented in an intuitive and visual manner, which is convenient for users to quickly understand the temperature conditions on the curtain surface.

[0018] Preferably, the step of determining whether a temperature difference occurs on the surface of the curtain includes:

[0019] Retrieve the temperature distribution data graph of the curtain surface; determine whether the color or brightness in the temperature distribution data graph is consistent. If yes, perform an external abnormality detection on the curtain; if not, use an infrared camera to scan and identify internal defects of the curtain.

[0020] The present invention determines whether it is necessary to perform internal defect identification on the curtain or external abnormality detection on the curtain by judging whether the color or brightness in the temperature distribution data graph is consistent. The temperature difference on the surface of the curtain can be accurately identified through the color or brightness difference of the temperature distribution data graph, and unnecessary defect detection operations can be reduced based on the preliminary judgment of the temperature distribution data graph, thereby reducing the false alarm rate and helping to improve the accuracy and efficiency of detection.

[0021] Preferably, the preprocessing of the curtain external image to obtain the abnormal area outside the curtain includes:

[0022] Retrieve the curtain external image; convert the color image into a grayscale image; use the Ostu maximum inter-class variance adaptive algorithm image processing technology to segment the grayscale image and distinguish the curtain from the background area; remove debris and noise in the image;

[0023] The largest connected domain part is extracted as the target area, wherein the target area includes: the abnormal curtain area and the normal curtain area; texture features are extracted from the target area, wherein the texture features include: texture, color and stains.

[0024] It should be noted that the installation position and angle of the camera to collect images must be able to clearly capture the curtains and any abnormal surface areas that may exist there; the maximum connected domain includes all abnormal curtain areas.

[0025] The present invention converts a color image into a grayscale image, which can simplify subsequent processing steps and reduce the amount of calculation because the grayscale image only contains brightness information; the grayscale image is segmented to distinguish the curtain and background areas, and the debris and noise in the image are effectively removed;

[0026] Preferably, the infrared sensor converts the received infrared radiation into an electrical signal, pre-processes the electrical signal, and obtains the curtain image, including:

[0027] The infrared sensor converts the received infrared radiation into an electrical signal and preprocesses the electrical signal. The preprocessing process includes: amplification, filtering, enhancement and noise reduction. An image is reconstructed based on the processed signal. Several reconstructed images are integrated to generate a reconstructed image data set.

[0028] It should be noted that the reconstructed image can display different temperature areas or specific objects of infrared reflection according to the intensity and distribution of infrared radiation; the reconstructed image is displayed on the display screen of the infrared camera;

[0029] The present invention helps to increase signal strength by amplifying the electrical signal through an amplifier circuit, which is convenient for subsequent signal processing and image reconstruction; the amplified signal will pass through a series of signal processing algorithms, including filtering, enhancement and noise reduction, which greatly reduces interference and helps to quickly extract the required information.

[0030] Preferably, the method of obtaining the abnormal area inside the curtain based on the image processing algorithm includes:

[0031] Retrieve the reconstructed image data set to obtain the original internal image of the sample curtain; compare the reconstructed image data set with the original internal image to obtain similarity;

[0032] It is determined whether the similarities of the reconstructed image data sets are all greater than the similarity threshold. If yes, the inner area of ​​the curtain is normal; if not, the corresponding reconstructed image data is annotated to obtain the abnormal area inside the curtain.

[0033] The present invention determines whether the similarities of the reconstructed image data sets are all greater than the similarity threshold, and then determines the abnormal area inside the curtain. When the similarity of the reconstructed image data sets is lower than the similarity threshold, the system can mark the abnormal area. In this way, the abnormal area can be accurately located, which greatly improves the efficiency of detection, reduces the time and labor cost of manual inspection, and improves the accuracy of detection.

[0034] Preferably, the automatic identification and classification of abnormal areas of the curtain based on image recognition technology includes:

[0035] Retrieving an abnormal region image dataset, and marking the abnormal region image dataset as an abnormality recognition sequence;

[0036] Retrieve the anomaly recognition model, input the anomaly recognition sequence into the anomaly recognition model, and obtain the anomaly label; match the corresponding anomaly type according to the anomaly label; match the corresponding pass rate type according to the anomaly label; wherein the anomaly prediction model is constructed based on the artificial intelligence model; the anomaly label is set to a positive integer.

[0037] The present invention can quickly scan the curtain surface through image recognition technology and identify abnormal areas in a short time. Compared with traditional manual detection methods, automatic recognition greatly improves detection efficiency and reduces manual labor costs.

[0038] Preferably, the abnormality prediction model is constructed based on an artificial intelligence model, including:

[0039] Acquire standard training data; wherein the standard training data includes standard input data consistent with the content attributes of the abnormal recognition sequence, and standard output data consistent with the content attributes of the abnormal label;

[0040] The artificial intelligence model is trained using standard training data, and the trained artificial intelligence model is marked as an abnormal prediction model; wherein the artificial intelligence model includes a convolutional neural network model or a long short-term memory neural network model.

[0041] Preferably, the method of forming a database and early warning information through the prediction of the qualified rate, and displaying the database and early warning information to the production management personnel includes:

[0042] By comparing the predicted value of the qualified rate of the current batch with the preset standard threshold, it is determined whether the predicted value of the qualified rate of the current batch exceeds the preset standard threshold. If yes, the test is continued; if not, an early warning message is immediately generated and displayed to the production management personnel; the preset standard threshold includes national standard values, international standard values ​​and standard values ​​specified by customers; the early warning information includes the abnormality type, severity and possible impact;

[0043] Store the forecast results and related historical data in the database; develop a visual interface to display the information and warning information in the database in the form of charts and reports; allow production managers to query historical data, forecast results and warning information through the interface.

[0044] It should be noted that when using a colorimeter to measure, the total color difference value △E is usually used to evaluate the color difference of curtains; the normal color difference value allowable range is as follows:

[0045] 0-0.25△E: very small or no, an ideal match;

[0046] 0.25-0.5△E: tiny, an acceptable match;

[0047] 0.5-1.0△E: Slight to moderate, acceptable in some applications;

[0048] 1.0-2.0△E: Medium, acceptable in certain applications;

[0049] 2.0-4.0△E: There is a gap, which is acceptable in certain applications;

[0050] 4.0 △E or above: Very large, unacceptable in most applications.

[0051] The present invention can establish an effective prediction and early warning system by storing the prediction results and related historical data in a database for subsequent query and analysis, helping production managers to promptly identify potential problems and make wise decisions, thereby improving production efficiency and reducing operational risks.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] 1. When detecting internal defects of curtains, an infrared camera and an infrared detector are combined. The infrared camera can form a video image to provide an intuitive monitoring effect, while the infrared detector can detect changes in infrared radiation, providing important information for the accuracy of detection. When detecting external defects of curtains, the color image is converted into a grayscale image, because the grayscale image only contains brightness information, which can simplify subsequent processing steps and reduce the amount of calculation; the grayscale image is segmented to distinguish between the curtain and background areas, and the debris and noise in the image are effectively removed.

[0054] 2. The infrared light-based household curtain defect detection system combines three modules: infrared detection, image acquisition and processing module, and defect processing and prediction module. Each module plays an indispensable role, ensuring that the system can efficiently and accurately detect curtain defects and make corresponding processing to achieve automatic detection and management of curtain defects; the present invention uses infrared light detection technology, which allows detection without direct contact with the curtain material, avoiding the risk of physical damage or contamination of the curtain; infrared light detection speed is fast, and large-area curtain detection can be completed in a short time, improving detection efficiency; infrared light detection technology can feedback the detection results in real time, so that producers or quality inspectors can quickly understand the quality status of the curtains. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0056] Figure 1 A schematic diagram of the module relationship of the present invention;

[0057] Figure 2 Schematic diagram of the steps of detecting internal defects of curtains according to the present invention;

[0058] Figure 3 It is a schematic diagram of the steps of detecting external defects of curtains according to the present invention. DETAILED DESCRIPTION

[0059] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0060] See also Figure 1 , the first aspect of the present invention provides a household curtain defect detection system based on infrared light, including: an infrared detection module, an image acquisition and processing module, and a defect processing and prediction module;

[0061] Infrared detection module: An infrared transmitter is set in the curtain warehouse to emit infrared light signals to the curtain surface through the infrared transmitter; an infrared receiver is set to obtain the infrared radiation intensity at different positions on the object surface through the infrared receiver; a temperature distribution data graph of the curtain surface is constructed according to the infrared radiation intensity, and whether there is a temperature difference on the curtain surface is determined according to the temperature distribution data graph. If yes, an infrared camera is used to scan and identify internal defects of the curtain; if no, an external abnormality detection of the curtain is performed;

[0062] Image acquisition and processing module: obtain the external image of the curtain through the image acquisition device, pre-process the external image of the curtain, and obtain the external abnormal area of ​​the curtain; wherein the external abnormal area includes: the texture, color difference and stains of the curtain surface;

[0063] An infrared camera is set up in the curtain warehouse. The infrared radiation received is converted into an electrical signal through an infrared sensor, and the electrical signal is pre-processed to obtain the internal image of the curtain. Based on the image processing algorithm, the abnormal area inside the curtain is obtained. Among them, the internal abnormal area includes: material unevenness and fabric damage, snagging, and scratches. The abnormal area outside the curtain and the abnormal area inside the curtain constitute the abnormal area image data set.

[0064] Defect processing and prediction module: retrieve the abnormal area image feature library, automatically identify and classify the abnormal area of ​​the curtain based on image recognition technology; obtain the classified abnormal area data of the curtain through the defect recognition module; determine whether the defect type is a stain on the surface of the curtain, if yes, clean the curtain; if not, mark the curtain as a defective product; obtain the qualified rate of the curtains of the historical batches, and predict the qualified rate of the curtains of this batch based on the qualified rate of the curtains of the historical batches and the image data set of the abnormal area; form a database and early warning information through the qualified rate prediction, and display the database and early warning information to the production management personnel;

[0065] The method of predicting the qualified rate of a batch of curtains based on the qualified rate of a historical batch of curtains and the abnormal area image dataset includes:

[0066] Retrieve the abnormal area image data set and divide it into N parts, where N is a positive integer greater than 1; obtain the sampling sample pass rate through sampling detection, retrieve the historical batch curtain pass rate from the database, and mark the sampling sample pass rate and the historical batch curtain pass rate as SQ and HQ respectively;

[0067] The predicted value of the current batch pass rate is calculated using the formula QF=SQ×α+HQ×β, where α and β are non-zero weight coefficients.

[0068] See also Figure 2-Figure 3, the steps of curtain defect detection are: set up an infrared transmitter in the curtain warehouse, and transmit infrared light signals to the curtain surface through the infrared transmitter; set up an infrared receiver, and obtain the infrared radiation intensity of different positions on the object surface through the infrared receiver; construct a temperature distribution data map of the curtain surface according to the infrared radiation intensity, and judge whether there is a temperature difference on the curtain surface according to the temperature distribution data map. If yes, use an infrared camera to scan and identify internal defects of the curtain; if not, perform external abnormality detection of the curtain; obtain the external image of the curtain through the image acquisition device, pre-process the external image of the curtain, and obtain the external abnormal area of ​​the curtain; set up an infrared camera in the curtain warehouse, convert the received infrared radiation into an electrical signal through an infrared sensor, pre-process the electrical signal, and obtain the internal image of the curtain; based on the image The image processing algorithm is used to obtain the abnormal area inside the curtain; the internal abnormal area includes: material unevenness and fabric damage, thread pulling, and scratches; the abnormal area outside the curtain and the abnormal area inside the curtain constitute an abnormal area image data set; the abnormal area image feature library is retrieved, and the abnormal area of ​​the curtain is automatically identified and classified based on image recognition technology; the classified abnormal area data of the curtain is obtained through the defect recognition module; it is determined whether the defect type is stains on the surface of the curtain, if yes, the curtain is cleaned; if not, the curtain is marked as a defective product; the qualified rate of curtains in historical batches is obtained, and the qualified rate of the batch of curtains is predicted based on the qualified rate of curtains in historical batches and the abnormal area image data set; a database and early warning information are formed through the qualified rate prediction, and the database and early warning information are displayed to production management personnel.

[0069] For example, there is a curtain warehouse. An infrared transmitter is set in the warehouse to emit infrared light signals to the curtain surface through the infrared transmitter; an infrared receiver is set to obtain the infrared radiation intensity at different positions on the surface of the object through the infrared receiver; a temperature distribution data map of the curtain surface is constructed according to the infrared radiation intensity. In the temperature distribution data map, the color and brightness of area A and area B are consistent, while the color and brightness of area C and area D are inconsistent; an infrared camera is set in the curtain warehouse, and the infrared sensor in the infrared camera converts the received infrared radiation into an electrical signal, and the electrical signal is pre-processed to obtain the internal abnormal area image of area A and area B; the abnormal area image feature library is retrieved, and the defects of area A and area B are automatically identified and classified based on image recognition technology, and the curtain is marked as defective, and external abnormality detection is performed on area C and area D; the abnormal area image feature library is retrieved, and the defects of area A and area B are automatically identified and classified based on image recognition technology, and the curtain is marked as defective, and external abnormality detection is performed on area C and area D; the abnormal area The domain image feature library is used to obtain the classified abnormal curtain area data through the defect recognition module; the defect types of area C and area D are judged, and it is concluded that the defect types of area C and area D are both stains on the surface of the curtain, so areas C and D are cleaned. Among them, the stains in area C can be cleaned and removed, so area C is not a defective product; the stains in area D cannot be cleaned and removed, so the curtain is marked as a defective product; and then the current batch pass rate prediction value is calculated by the formula QF=SQ×α+HQ×β. If the current batch pass rate prediction value is 97.5%, the preset standard threshold is 96%, and the current batch pass rate prediction value exceeds the preset standard threshold, then continue to detect; if the current batch pass rate prediction value is 95%, the preset standard threshold is 96%, and the current batch pass rate prediction value is lower than the preset standard threshold, then immediately generate a warning message and display it to the production management personnel.

[0070] Part of the data in the above formula is calculated by removing the dimension and taking its numerical value. The formula is a formula closest to the actual situation obtained by software simulation of a large amount of collected data; the preset parameters and preset thresholds in the formula are set by technical personnel in this field according to actual conditions or obtained through simulation of a large amount of data.

[0071] The working principle of the present invention is as follows: an infrared light signal is emitted to the surface of the curtain by an infrared transmitter; an infrared receiver is set, and the infrared radiation intensity at different positions on the surface of the object is obtained by the infrared receiver; a temperature distribution data map of the curtain surface is constructed according to the infrared radiation intensity, and whether a temperature difference occurs on the curtain surface according to the temperature distribution data map; an external image of the curtain is obtained by an image acquisition device, and the external image of the curtain is preprocessed to obtain an abnormal area outside the curtain; an infrared camera is set in a curtain warehouse, and the received infrared radiation is converted into an electrical signal by an infrared sensor, and the electrical signal is preprocessed to obtain an internal image of the curtain; based on an image processing algorithm, the abnormal area inside the curtain is obtained; by calling an abnormal area image feature library, the abnormal area of ​​the curtain is automatically identified and classified based on image recognition technology; the classified abnormal area data of the curtain is obtained by a defect recognition module; it is judged whether the defect type is a stain on the surface of the curtain; the qualified rate of the curtain of a historical batch is obtained, and the qualified rate of the curtain of the batch is predicted according to the qualified rate of the curtain of the historical batch and the image data set of the abnormal area; a database and early warning information are formed by the qualified rate prediction, and the database and early warning information are displayed to production management personnel.

[0072] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A household curtain defect detection system based on infrared light, characterized in that: include: Infrared detection module, image acquisition and processing module, and defect processing and prediction module; Infrared detection module: An infrared transmitter is set in the curtain warehouse to emit infrared light signals to the curtain surface through the infrared transmitter; An infrared receiver is set up to obtain the infrared radiation intensity at different positions on the surface of the object through the infrared receiver; The temperature distribution data graph of the curtain surface is constructed based on the infrared radiation intensity. Based on the temperature distribution data graph, it is determined whether there is a temperature difference on the curtain surface. If yes, an infrared camera is used to scan and identify the internal defects of the curtain; if no, an external abnormality detection of the curtain is performed. Image acquisition and processing module: obtain the external image of the curtain through the image acquisition device, pre-process the external image of the curtain, and obtain the external abnormal area of ​​the curtain; wherein the external abnormal area includes: the texture, color difference and stains of the curtain surface; An infrared camera is set up in the curtain warehouse. The infrared radiation received is converted into an electrical signal through an infrared sensor, and the electrical signal is pre-processed to obtain the internal image of the curtain. Based on the image processing algorithm, the abnormal area inside the curtain is obtained. Among them, the internal abnormal area includes: material unevenness and fabric damage, snagging, and scratches. The abnormal area outside the curtain and the abnormal area inside the curtain constitute the abnormal area image data set. Defect processing and prediction module: retrieve the abnormal area image feature library, automatically identify and classify the abnormal area of ​​the curtain based on image recognition technology; obtain the classified abnormal area data of the curtain through the defect recognition module; determine whether the defect type is a stain on the surface of the curtain, if yes, clean the curtain; if not, mark the curtain as a defective product; obtain the qualified rate of the curtains of the historical batches, and predict the qualified rate of the curtains of the batch based on the qualified rate of the curtains of the historical batches and the image data set of the abnormal area; form a database and early warning information through the qualified rate prediction, and display the database and early warning information to the production management personnel; The qualified rate of a batch of curtains is predicted based on the qualified rate of the historical batch of curtains and the abnormal area image dataset, including: Retrieve the abnormal area image data set and divide it into N parts, where N is a positive integer greater than 1; obtain the sampling sample pass rate through sampling detection, retrieve the historical batch curtain pass rate from the database, and mark the sampling sample pass rate and the historical batch curtain pass rate as SQ and HQ respectively; The predicted value of the current batch pass rate is calculated using the formula QF=SQ×α+HQ×β, where α and β are non-zero weight coefficients.

2. The infrared light-based household curtain defect detection system according to claim 1, characterized in that: The step of constructing a temperature distribution data graph of the curtain surface according to the infrared radiation intensity includes: The infrared radiation intensity at different positions on the surface of the object is obtained through an infrared receiver, and the intensity value is converted into different colors or brightness in the image; the different colors and brightness are integrated to obtain the temperature distribution data map of the curtain surface.

3. The infrared light-based household curtain defect detection system according to claim 2, characterized in that: The determining whether a temperature difference occurs on the surface of the curtain comprises: Retrieve the temperature distribution data graph of the curtain surface; determine whether the color or brightness in the temperature distribution data graph is consistent. If yes, perform an external abnormality detection on the curtain; if not, use an infrared camera to scan and identify internal defects of the curtain.

4. The infrared light-based household curtain defect detection system according to claim 1, characterized in that: The preprocessing of the curtain external image to obtain the curtain external abnormal area includes: Retrieve the curtain external image; convert the color image into a grayscale image; use the Ostu maximum inter-class variance adaptive algorithm image processing technology to segment the grayscale image and distinguish the curtain from the background area; remove debris and noise in the image; The largest connected domain part is extracted as the target area, wherein the target area includes: the abnormal curtain area and the normal curtain area; texture features are extracted from the target area, wherein the texture features include: texture, color and stains.

5. The infrared light-based household curtain defect detection system according to claim 1, characterized in that: The infrared sensor converts the received infrared radiation into an electrical signal, pre-processes the electrical signal, and obtains the curtain image, including: The infrared sensor converts the received infrared radiation into an electrical signal and preprocesses the electrical signal. The preprocessing process includes: amplification, filtering, enhancement and noise reduction. An image is reconstructed based on the processed signal. Several reconstructed images are integrated to generate a reconstructed image data set.

6. The infrared light-based household curtain defect detection system according to claim 5, characterized in that: The method of obtaining the abnormal area inside the curtain based on the image processing algorithm includes: Retrieve the reconstructed image data set to obtain the original internal image of the sample curtain; compare the reconstructed image data set with the original internal image to obtain similarity; It is determined whether the similarities of the reconstructed image data sets are all greater than the similarity threshold. If yes, the inner area of ​​the curtain is normal; if not, the corresponding reconstructed image data is annotated to obtain the abnormal area inside the curtain.

7. The infrared light-based household curtain defect detection system according to claim 1, characterized in that: The automatic identification and classification of abnormal areas of curtains based on image recognition technology includes: Retrieving an abnormal region image dataset, and marking the abnormal region image dataset as an abnormality recognition sequence; Retrieve the anomaly recognition model, input the anomaly recognition sequence into the anomaly recognition model, and obtain the anomaly label; match the corresponding anomaly type according to the anomaly label; match the corresponding pass rate type according to the anomaly label; wherein the anomaly prediction model is constructed based on the artificial intelligence model; the anomaly label is set to a positive integer.

8. The infrared light-based household curtain defect detection system according to claim 7, characterized in that: The abnormal prediction model is constructed based on an artificial intelligence model, including: Acquire standard training data; wherein the standard training data includes standard input data consistent with the content attributes of the abnormal recognition sequence, and standard output data consistent with the content attributes of the abnormal label; The artificial intelligence model is trained using standard training data, and the trained artificial intelligence model is marked as an abnormal prediction model; wherein the artificial intelligence model includes a convolutional neural network model or a long short-term memory neural network model.

9. The infrared light-based household curtain defect detection system according to claim 1, characterized in that: The method of forming a database and early warning information through the prediction of the qualified rate, and displaying the database and early warning information to the production management personnel, includes: By comparing the predicted value of the qualified rate of the current batch with the preset standard threshold, it is determined whether the predicted value of the qualified rate of the current batch exceeds the preset standard threshold. If yes, an early warning message is immediately generated and displayed to the production management personnel; if no, the test is continued; the preset standard threshold includes national standard value, international standard value and standard value specified by the customer; the early warning message includes the abnormality type, severity and possible impact; Store the forecast results and related historical data in the database; develop a visual interface to display the information and warning information in the database in the form of charts and reports; allow production managers to query historical data, forecast results and warning information through the interface.

Images