A front windshield glass heating detection method, system, product, medium and device

Abstract

The application discloses a front windshield glass heating detection method, system, product, medium and equipment, and relates to the technical field of vehicle accessories. The method comprises the following steps: acquiring a multi-angle image of a front windshield glass of a vehicle to be detected, a heating state thermal imaging image and heating loop parameters; performing micro-crack optical screening on the multi-angle image of the front windshield glass by using a visual detection model; performing static detection on the integrity of the heating loop; performing instantaneous interruption monitoring on the heating loop based on dynamic load simulation; comprehensively analyzing the heating process qualification of the front windshield glass based on the optical detection result, the static detection result of the heating loop, the dynamic detection result of the heating loop and the heating state thermal imaging image, and tracing abnormal conditions according to the correlation between the detection results. The application can more accurately and comprehensively detect and trace the heating function of the front windshield glass before the vehicle is delivered.

Description

Technical Field

[0001] This invention relates to the field of vehicle parts technology, and in particular to a method, system, product, medium and equipment for testing the heating of a windshield. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] After a period of use, electrically heated windshields are prone to localized heating malfunctions, necessitating pre-shipment heating testing of the windshield. However, with the increasing complexity of electrically heated windshield manufacturing technology, traditional testing methods have become severely limited, failing to detect anomalies under microscopic conditions and dynamic operating conditions in a timely manner. For example, for multi-layered coated glass or printed circuit heated glass, continuity testing with a multimeter alone is insufficient to detect impedance anomalies caused by minute cracks. Furthermore, while static contact resistance may appear normal due to fretting corrosion at connector terminals or poor solder joints, it can lead to open circuits under vehicle vibration or temperature changes. Summary of the Invention

[0004] In view of the shortcomings of the existing technology, the purpose of this invention is to provide a method, system, product, medium and equipment for testing the heating function of the windshield before the vehicle leaves the factory, which can perform more accurate and comprehensive testing and traceability of the heating function of the windshield before the vehicle leaves the factory.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solution: The first aspect of this invention provides a method for detecting the heating of a windshield, comprising the following steps: Acquire multi-angle images of the windshield of the vehicle under test, thermal imaging of the heating status, and heating circuit parameters; A visual inspection model was used to perform optical screening of microcracks on multi-angle images of the windshield to obtain optical inspection results; A static integrity test of the heating circuit is performed to obtain the static test results of the heating circuit. Based on dynamic load simulation, transient monitoring of the heating circuit is performed to obtain dynamic detection results of the heating circuit; The passability analysis of the windshield heating process was conducted by combining the results of optical inspection, static inspection of the heating circuit, dynamic inspection of the heating circuit, and thermal imaging of the heating status. The abnormal conditions were traced based on the correlation between the inspection results.

[0006] Furthermore, preprocessing operations are performed on the multi-angle images of the windshield of the vehicle to be tested, the thermal imaging images of the heating state, and the heating circuit parameters. The specific steps include: aligning the multi-angle images, thermal imaging image sequences, and circuit parameter sequences along the time axis; denoising and normalizing the brightness of the multi-angle images; calibrating the temperature values ​​of the thermal imaging images; and removing outliers and smoothing the heating circuit parameters.

[0007] Furthermore, the specific steps for using a visual inspection model to perform optical screening of microcracks in multi-angle images of the windshield are as follows: A lightweight visual model is used to perform coarse crack screening on the windshield image to analyze whether there are obvious cracks and sealing defects in the windshield. A Transformer model based on a global attention mechanism is used to screen for microcracks in images without macrocracks or in areas outside the macrocrack region, and to output defect location information. The results of coarse crack screening and micro crack screening are integrated to obtain optical detection results.

[0008] Furthermore, the specific steps for using a Transformer model based on a global attention mechanism to screen for microcracks in images without macrocracks or in areas outside the macrocrack region are as follows: The image without macro-cracks or the area outside the macro-crack area is divided into local regions and initial features are extracted; Based on the initial feature overlay position information; Based on the initial features obtained after overlaying positional information, a global contextual association is established through a self-attention mechanism to enhance features; The feature outputs the defect category and location based on the globally enhanced features.

[0009] Furthermore, based on dynamic load simulation, the specific steps for instantaneous monitoring of the heating circuit are as follows: Define transient events; The transient interruption events are counted according to the definition of transient interruption events, and the occurrence time, duration, and amplitude changes of the transient interruption events are recorded; Output dynamic detection results along with a list of transient interruption events.

[0010] Furthermore, the specific steps for tracing the source of abnormalities based on the correlation between test results are as follows: Construct a rule base for association based on the correlation between detection results; Match the detection results with features in the association rule base; Output the source tracing results and generate an anomaly source tracing report.

[0011] A second aspect of the present invention provides a windshield heating detection system, comprising: The data acquisition module is configured to acquire multi-angle images of the windshield of the vehicle under test, thermal images of the heating status, and heating circuit parameters. The optical screening module is configured to use a visual inspection model to perform optical screening of microcracks on multi-angle images of the windshield to obtain optical inspection results. The circuit static detection module is configured to perform static circuit integrity detection on the heating circuit and obtain the static detection result of the heating circuit. The loop dynamic detection module is configured to perform transient interruption monitoring on the heating loop based on dynamic load simulation, and obtain the dynamic detection results of the heating loop. The comprehensive analysis module is configured to perform a pass / fail analysis of the windshield heating process by combining optical inspection results, static inspection results of the heating circuit, dynamic inspection results of the heating circuit, and thermal imaging of the heating status, and to trace the source of abnormalities based on the correlation between the inspection results.

[0012] A third aspect of the present invention provides a computer-readable storage medium storing a computer program adapted to be loaded by a processor and to execute the steps of the windshield heating detection method as described in the first aspect of the present invention.

[0013] A fourth aspect of the present invention provides a computer device comprising: A processor, adapted to execute computer programs; A computer-readable storage medium storing a computer program, which, when executed by the processor, implements the windshield heating detection method as described in the first aspect of the present invention.

[0014] A fifth aspect of the present invention provides a computer program product or computer program comprising computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the steps in the windshield heating detection method as described in the first aspect of the present invention.

[0015] The above one or more technical solutions have the following beneficial effects: The present invention discloses a method, system, product, medium and device for detecting the heating of a front windshield. By comprehensively detecting the heating circuit through dynamic and static detection methods, it can accurately identify the micro-corrosion or false welding phenomena that cannot be discovered by the prior art, improving the safety level of the vehicle during driving. It can also trace the cause of abnormal conditions based on thermal imaging technology, accurately identify the difference in the abnormal causes between the micro-cracks in the front windshield and the problems in the heating circuit, overcoming the defect in the prior art that it can only detect whether it is qualified and cannot accurately locate the abnormal cause.

[0016] Advantages of additional aspects of the present invention will be partially given in the following description, partially become obvious from the following description, or be understood through the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0018] Figure 1 It is a flowchart of the method for detecting the heating of the front windshield in Embodiment 1 of the present invention; Figure 2 It is a schematic structural diagram of an electrically heated front windshield in Embodiment 1 of the present invention; Figure 3 It is a schematic structural diagram of the heating circuit in Embodiment 1 of the present invention; Figure 4 It is a schematic diagram of the bus bar welding sequence and method in Embodiment 1 of the present invention; Figure 5 It is a schematic diagram of the sealant coating area in Embodiment 1 of the present invention; Figure 6 It is a schematic diagram of the sealant coating partition in Embodiment 1 of the present invention; Among them, 1. Heating circuit, 11. Left heating area, 12. Right heating area, 2. Left negative terminal, 21. Left negative bus bar, 22. Right negative bus bar, 3. Right negative terminal, 31. Left positive bus bar, 32. Right positive bus bar, 4. Left positive terminal, 41. Glass lower side edge area, 42. Glass left side edge area, 43. Glass upper side edge area, 44. Glass right side edge area, 5. Right positive terminal, 51. Front windshield, 51a. Outer glass sheet, 51b. Inner glass sheet, 51c. PVB film. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0019] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0020] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof. The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0021] Example 1: Embodiment 1 of the present invention provides a method for detecting the heating of a windshield, taking an electrically heated windshield as an example, such as... Figure 1 As shown, it includes the following steps: S1: Acquire multi-angle images of the windshield of the vehicle under test, thermal imaging of the heating status, and heating circuit parameters.

[0022] S11: Obtain multi-source parameters.

[0023] S111: Acquire multi-angle images of the windshield of the vehicle to be inspected.

[0024] In one specific implementation, imaging of the outer and inner surfaces of the windshield is triggered sequentially according to a preset angle sequence, such as top view, side view, and 45° oblique view, ensuring coverage of the entire glass area, with a focus on edges, corners, and areas with dense heating wires. At least one high-resolution grayscale or color image is acquired for each angle, and the corresponding angle label and acquisition timestamp are recorded.

[0025] S112: Acquire a thermal image of the windshield heating status of the vehicle to be inspected.

[0026] In one specific implementation, while the heating circuit is in operation (powered on for heating), thermal images are continuously acquired at fixed time intervals until the heating process ends or the glass surface temperature stabilizes. The thermal images record the temperature distribution on the glass surface, including a temperature comparison between the heating wire area and the non-heated area.

[0027] S113: Obtain the parameters of the windshield heating circuit of the vehicle under test.

[0028] In one specific implementation, parameters such as loop voltage, loop current, heating power, ambient temperature, and heating duration are collected before, during, and after the heating process. The parameters are stored in time-series format, with each parameter accompanied by an identifier indicating the time of collection and the sampling frequency.

[0029] S12: Preprocess the data.

[0030] In one specific implementation, multi-angle images, thermal imaging sequences, and circuit parameter sequences are aligned along the time axis to ensure that multiple data points at the same time can be analyzed jointly. The multi-angle images are denoised and their brightness normalized; the thermal imaging sequences are calibrated for temperature values; and the heating circuit parameters undergo outlier removal and smoothing.

[0031] S2: Use a visual inspection model to perform optical screening of microcracks on multi-angle images of the windshield to obtain optical inspection results.

[0032] S21: A lightweight visual model is used to perform coarse crack screening on the windshield image to analyze whether there are obvious cracks and sealing defects in the windshield.

[0033] In one specific implementation, the image from each angle is sequentially input into a pre-trained lightweight convolutional neural network. The model outputs a binary classification result indicating whether there is a significant crack or not, along with a confidence score. If a significant crack is detected at any angle, the defect type is directly recorded as a macroscopic crack, and the corresponding angle and region coordinates are labeled.

[0034] Similarly, for sealing defects, a pre-trained lightweight convolutional neural network is used to output the result of whether a sealing defect exists, along with a confidence score. For example... Figure 5 As shown, in this embodiment, the edges of the windshield are sealed with GP402 sealant, starting from the outer glass pane 51a and extending to the inner surface of the inner glass pane 51b, ensuring that the PVB membrane 51c is also sealed with GP402 sealant. Through the sealing with GP402 sealant, the entire glass cross-section is effectively sealed at the glass edges, eliminating the tiny gaps between the outer glass pane 51a, the inner glass pane 51b, and the PVB membrane 51c, preventing corrosive liquids from the external environment from entering the glass.

[0035] like Figure 6As shown, this embodiment employs a method of sealing the glass edges using a primer coating. The coating area includes the lower edge region 41 of the glass (C1 to C2), where the distance from C1 to A1 is ≥60mm and the distance from C2 to A4 is ≥60mm; the coating area includes the left edge region 42 of the glass (C5 to C6), where the distance from C3 to B1 is ≥60mm and the distance from C6 to B6 is ≥60mm; the coating area includes the upper edge region 43 of the glass (C4 to C5); and the coating area includes the right edge region 44 of the glass (C3 to C4). By applying the GP402 primer coating process to all busbars, namely the left positive busbar 31, the right positive busbar 32, the left negative busbar 21, and the right negative busbar 22, the sealing performance of the glass edges is further improved.

[0036] It should be noted that existing models, such as the YOLO model, can be used for the detection of obvious cracks and sealing defects; the specific structure will not be described here. In this embodiment, C1, C2, C3, C4, C5, and C6 are coating key points set based on experience. The coating area can be determined based on these coating key points, and the specific location can be customized according to the actual situation.

[0037] S22: The Transformer model based on the global attention mechanism is used to screen for microcracks in images without macrocracks or in areas outside the macrocrack region, and the defect location information is output.

[0038] In one specific implementation, the microcrack screening targets include pinholes, scratches, or traces caused by excessive laser trimming, referred to as microcracks. These microcracks are so small and difficult to detect that a more accurate model is needed for target detection.

[0039] In this embodiment, an image without macroscopic cracks or the area outside the macroscopic crack region is input into a Transformer vision model based on a global attention mechanism. The Transformer vision model extracts features from each local region in the image, establishes global contextual association through a self-attention mechanism, and outputs the category and location of the defect. Each detected defect outputs a bounding box. and category probability .

[0040] The specific steps include: S221: Divide the image without macro-cracks or the area outside the macro-crack region into local regions and extract initial features.

[0041] In one specific implementation, the input image is first divided into multiple non-overlapping rectangular blocks, each corresponding to a local region in the image. For each block, the model transforms its pixel values ​​into an initial feature vector through a linear projection layer. At this point, each feature vector only represents the visual content of that local region, such as the presence of edges, textures, brightness variations, etc., without considering the relationship between that region and other parts of the image.

[0042] S222: Overlay position information based on initial features.

[0043] In one specific implementation, a positional code is added to each feature vector to determine the specific coordinates of the current local region in the image (e.g., top left, center, or bottom right). This way, even if two small patches have exactly the same visual content, the model can differentiate them based on their different positions.

[0044] S223: Based on the initial features after overlaying position information, establish global context association through a self-attention mechanism to perform feature enhancement.

[0045] In one specific implementation, all feature vectors with superimposed positional encodings are input together into the Transformer encoder. In the encoder's multi-head self-attention sublayer, the feature vectors of each local region simultaneously share the feature vectors of all other local regions. The model calculates the attention weights between each pair of local regions: If two regions are highly related semantically or spatially (e.g., the two breakpoints of a crack are located in different regions), then the attention weight between them will be high.

[0046] Each region aggregates information from other regions based on the attention weights of all other regions and updates its own feature vector.

[0047] Through multiple global interactions, vectors that originally contained only local information gradually incorporate the global context. For example, a small patch that initially appears to have only weak edges will be enhanced by the model as part of a crack after it notices that adjacent small patches also have continuous edges; an isolated noise point will be suppressed if it has no other supporting regions.

[0048] S224: Output defect category and location based on globally enhanced features.

[0049] In one specific implementation, after encoder processing, the feature vector of each local region already contains the contextual semantics of that region within the entire image. These enhanced feature vectors are then input into the detection head module. The detection head uses a set of learnable queries to filter out the spatial extent most likely to be the defect target from all regions.

[0050] For each detected defect, the detection head outputs two parts of information: Category and Confidence: Determine the category of the defect (microcrack, pinhole, scratch, laser trimming crack, etc.) and provide a probability value. This indicates the reliability of the judgment.

[0051] Location: Outputs a bounding box, using... It indicates. Among them. This typically represents the coordinates of the center point or top-left corner of the bounding box in the image coordinate system. Indicates width, This indicates the height. This bounding box precisely defines the area occupied by the defect in the image.

[0052] It should be noted that in actual processing, this embodiment uses images from multiple angles as model input. For detection results of the same physical location at different angles, the model will fuse them in post-processing: if multiple angles detect the same type of defect in similar spatial locations, they will be merged into one defect output and the confidence level will be increased; if only a single angle detects the defect, it will be marked as a suspicious defect and handed over to the subsequent comprehensive judgment stage for processing.

[0053] S23: Integrate the coarse crack screening results with the micro crack screening results to obtain the optical detection results.

[0054] In one specific implementation, the coarse screening results and fine detection results from all angles are merged, and duplicates (repeated detections of the same physical location at different angles) are removed to generate a set of optical detection results. .

[0055] in, This is a collection of optical detection results. Defect type To determine the position in the glass coordinate system, For model confidence, The image angle used to detect the defect.

[0056] S3: Perform static integrity testing on the heating circuit to obtain the static test results of the heating circuit.

[0057] In one specific implementation, conventional methods such as circuit testing and manual inspection are used to perform conventional heating circuit integrity checks to ensure that the heating circuit is connected and designed according to the design method of this embodiment, and that the overall circuit connection is complete. The specific design method of the heating circuit in this embodiment is as follows: The electrically heated windshield in this embodiment is as follows: Figure 2As shown, the system includes a windshield 51, on which are provided a left negative terminal 2, a right negative terminal 3, a left positive terminal 4, and a right positive terminal 5. The left negative terminal 2 and the right negative terminal 3 are connected to the vehicle's negative terminal, and the left positive terminal 4 and the right positive terminal 5 are connected to the vehicle's positive terminal, thus completing the wiring connection.

[0058] In this embodiment, the heating circuit 1 for the electrically heated windshield is as follows: Figure 3 As shown, the structure includes a left heating zone 11, a right heating zone 12, a left negative busbar 21, a right negative busbar 22, a left positive busbar 31, a right positive busbar 32, a left negative connector 2, a right negative connector 3, a left positive connector 4, and a right positive connector 5. The left heating zone 11 and the right heating zone 12 are symmetrically arranged, and each heating zone contains evenly distributed heating wires. When energized, the heating wires heat up, achieving uniform heating of the entire glass panel and enabling de-icing and snow removal.

[0059] like Figure 4 As shown, the busbar welding sequence in this embodiment is as follows: Welding at the lower busbars, namely the right positive busbar 32 and the left positive busbar 31: For the right half, start from point A1 and extend the welding to point A10, where the distance between point A1 and A6 is ≥20mm. At point A10, weld the right positive terminal connector 5 to the right positive busbar 32, and then weld from point A10 to point A2, where the distance between the right positive busbar 32 and A7 is ≥20mm.

[0060] For the left half, weld from point A4 to point A11, where the distance between points A4 and A9 is ≥20mm. At point A11, weld the left positive terminal connector 4 to the left positive busbar 31. Then weld from point A11 to point A3, where the distance between the left positive busbar 31 and A8 is ≥20mm.

[0061] Welding at the upper busbars, namely the left negative busbar 21 and the right negative busbar 22: For the right half, starting from position B1, weld the right negative terminal connector 3 to the right negative busbar 22, and then weld to point B2; the distance from point B1 to A7 is ≥20mm, and the distance from point B2 to B7 is ≥20mm; then weld from point B2 to point B3, where the distance from point B3 to B9 is ≥20mm.

[0062] For the left half, starting from position B6, weld the left negative terminal connector 2 to the left negative busbar 21, and then weld to point B5; the distance between point B6 and B8 is ≥20mm, and the distance between point B5 and B8 is ≥20mm; then weld from point B5 to point B4, where the distance between point B4 and B10 is ≥20mm.

[0063] By controlling the welding sequence and method of the busbar, issues such as incomplete welds and weld porosity between the busbar and heating wire can be controlled, improving the welding quality of the connection between the busbar and heating wire. Simultaneously, the distance between the busbar and the glass edge can be controlled, keeping the busbar away from the glass edge and reducing the risk of corrosion from external substances. Furthermore, immersion tests were conducted using a sodium chloride and sodium hydroxide solution as a de-icing agent. The solution contained 10% sodium chloride and 10% sodium hydroxide, 500g of de-icing agent, and 500ml of glass cleaner. After immersing the glass in the solution for 1200 hours with electricity applied, the tests showed that the glass met the requirements for windshield heating.

[0064] It should be noted that A1, A2, A3, A4, A5, A6, A7, A8, A9, A10 and B1, B2, B3, B4, B5, B6, B7, B8, B9, B10 in this embodiment are welding points set based on experience for welding positioning. As long as the above constraints are met, the specific positions of the other unspecified welding points can be customized according to the actual situation.

[0065] S4: Based on dynamic load simulation, instantaneous interruption monitoring of the heating circuit is performed to obtain the dynamic detection results of the heating circuit.

[0066] In one specific implementation, dynamic load simulation refers to constructing a load model based on an existing simulation software platform to simulate abnormal operating conditions that may occur during vehicle operation, such as voltage fluctuations, vibrations, and temperature changes. It involves applying varying voltage and current loads to the heating circuit according to a preset time sequence, while simultaneously recording the continuous waveform of the circuit current. During dynamic load simulation, a sliding window analysis is performed on the current waveform based on the occurrence of transient interruption events.

[0067] S41: Define transient interruption events.

[0068] In this embodiment, a transient event is defined as: in any continuous... Within a certain time period, the current value is lower than the normal operating current. of The time exceeds the threshold, and the duration exceeds the threshold. .

[0069] S42: Count transient events according to the definition of transient events and record the occurrence time, duration, and amplitude changes of transient events.

[0070] Calculate the reference current This represents the theoretically expected current value based on the load model.

[0071] Calculate real-time deviation .

[0072] like And continue , a momentary interruption event is determined.

[0073] Among them, is the allowable fluctuation ratio relative to the reference current.

[0074] Count the number of momentary interruptions during the entire dynamic test process and the occurrence time, duration, and amplitude change of each momentary interruption. If , the dynamic detection result is "There is a momentary interruption abnormality", otherwise it is "dynamically normal".

[0075] S43: Output the dynamic detection result , and attach a list of momentary interruption events.

[0076] S5: Analyze the qualification of the front windshield heating process by integrating the optical detection result, the static detection result of the heating circuit, the dynamic detection result of the heating circuit, and the thermal imaging map of the heating state, and trace the abnormal conditions based on the relevance between the detection results.

[0077] S51: Analyze the qualification of the front windshield heating process by integrating the optical detection result, the static detection result of the heating circuit, the dynamic detection result of the heating circuit, and the thermal imaging map of the heating state.

[0078] In a specific implementation, if any one of the optical detection result, the static detection result of the heating circuit, and the dynamic detection result of the heating circuit is unqualified, it is determined that the current front windshield heating process is unqualified, otherwise it is determined to be preliminarily qualified. For the preliminarily qualified front windshield and heating circuit, temperature distribution detection is performed based on the thermal imaging map of the heating state. If the temperature distribution detection is qualified, the final result is judged to be qualified, otherwise it is judged to be unqualified, and the cause of the abnormal condition of the unqualified front windshield and heating circuit is analyzed.

[0079] 1. Judgment of the optical detection result. If contains any type of defect (macro crack, micro crack, pinhole, scratch, laser trimming crack), and the defect confidence > threshold , the optical detection is unqualified. Otherwise, the optical detection is qualified.

[0080] 2. Judgment of the static detection of the circuit. If is "normal", the static detection of the circuit is qualified; otherwise, the static detection of the circuit is unqualified. Among them, is the static detection result of the circuit, representing integrity, circuit connectivity, insulation abnormal conditions, etc.

[0081] 3. Judgment of the dynamic detection of the circuit. If is "normal" and If it is equal to 0, the loop dynamic detection is qualified; otherwise, the loop dynamic detection is unqualified.

[0082] The specific steps for temperature distribution detection based on the thermal imaging map in the heating state are as follows: Extract the temperature distribution in the heating wire area of the thermal imaging map and calculate the highest temperature , the lowest temperature , and the temperature uniformity index , where is the temperature standard deviation, is the average temperature.

[0083] Compare with the theoretical heating temperature distribution model. If exceeds the safety temperature threshold or is lower than the uniformity threshold, it is heating abnormality, and the temperature distribution detection is judged as unqualified. Detect whether there is local overheating or cold area (the temperature of the continuous area deviates from the mean value by more than γ times the standard deviation). If so, mark the abnormal area and judge the temperature distribution detection as unqualified. In other cases, the temperature distribution detection is judged as qualified.

[0084] S52: Trace the source of the abnormal condition according to the correlation between the detection results.

[0085] S521: Construct an association rule library according to the correlation between the detection results.

[0086] In a specific implementation, predefine the logical relationship between common abnormal patterns and multi-source detection results based on the results of optical detection of coarse cracks, optical detection of micro-crack screening, static detection of the heating loop, dynamic detection of the heating loop, and thermal imaging map in the heating state, and construct an association rule library.

[0087] For example: If optical detection finds micro-cracks near the heating wire, and at the same time the thermal imaging map shows that the temperature in this area is low → it may be that the crack causes the heating wire to break or the resistance to increase.

[0088] If the static detection of the heating loop is normal, but there is an instantaneous break in the dynamic detection of the heating loop, and at the same time the thermal imaging map shows that a certain area does not heat intermittently → it may be that the local contact of the heating wire is poor.

[0089] If there is no abnormality in optical detection, but there is local overheating in the thermal imaging map → it may be that the local resistance of the heating loop is abnormal (such as a processing defect causing a smaller cross-section).

[0090] S522: Match the detection results with the features in the association rule library.

[0091] In one specific implementation, key features are extracted from the current detection results, including the location and type of optical defects, the direction and magnitude of static resistance deviation, the correspondence between dynamic transient time and load conditions, and the location and shape of abnormal temperature regions in the thermal imaging. The key features are matched with features in the association rule base by calculating similarity: .

[0092] in, For the first Weights of class features For indicator functions. (Selection) The highest rule is considered the most likely cause of the anomaly. For anomaly rule index, Indicates the first Exception rule characteristics, This represents the extracted k-th class feature value. For matching logic.

[0093] S523: Output the source tracing results and generate an anomaly source tracing report.

[0094] In one specific implementation, the anomaly tracing report includes: the abnormal phenomenon (such as uneven heating, local non-heating, instantaneous interruption), possible causes (such as cracks causing open circuits, poor contact, geometrical abnormalities of the heating wire), confidence level (based on matching similarity normalization), and suggested directions for further verification or repair.

[0095] It is important to note that if multiple anomalies exist simultaneously, check whether they can be explained by the same root cause (e.g., a single processing defect simultaneously causes microcracks and local resistance changes). If they cannot be unified, output multiple independent source tracing results.

[0096] Example 2: Embodiment 2 of the present invention provides a windshield heating detection system, comprising: The data acquisition module is configured to acquire multi-angle images of the windshield of the vehicle under test, thermal images of the heating status, and heating circuit parameters. The optical screening module is configured to use a visual inspection model to perform optical screening of microcracks on multi-angle images of the windshield to obtain optical inspection results. The circuit static detection module is configured to perform static circuit integrity detection on the heating circuit and obtain the static detection result of the heating circuit. The loop dynamic detection module is configured to perform transient interruption monitoring on the heating loop based on dynamic load simulation, and obtain the dynamic detection results of the heating loop. The comprehensive analysis module is configured to perform a pass / fail analysis of the windshield heating process by combining optical inspection results, static inspection results of the heating circuit, dynamic inspection results of the heating circuit, and thermal imaging of the heating status, and to trace the source of abnormalities based on the correlation between the inspection results.

[0097] Example 3: Embodiment 3 of the present invention provides a computer-readable storage medium storing a computer program adapted for loading by a processor and executing the steps of the windshield heating detection method as described in Embodiment 1 of the present invention, wherein the steps are: S1 acquires multi-angle images of the windshield of the vehicle under test, thermal images of the heating status, and heating circuit parameters.

[0098] S2: Use a visual inspection model to perform optical screening of microcracks on multi-angle images of the windshield to obtain optical inspection results.

[0099] S3: Perform static integrity testing on the heating circuit to obtain the static test results of the heating circuit.

[0100] S4: Based on dynamic load simulation, instantaneous interruption monitoring of the heating circuit is performed to obtain the dynamic detection results of the heating circuit.

[0101] S5: Conduct a pass / fail analysis of the windshield heating process by combining the results of optical inspection, static inspection of the heating circuit, dynamic inspection of the heating circuit, and thermal imaging of the heating status, and trace the source of abnormalities based on the correlation between the inspection results.

[0102] The detailed steps are the same as those for the windshield heating test method provided in Example 1, and will not be repeated here.

[0103] Example 4: Embodiment 4 of the present invention provides a computer device, the device comprising: A processor, adapted to execute computer programs; A computer-readable storage medium storing a computer program, which, when executed by the processor, implements the steps in the windshield heating detection method as described in Embodiment 1 of the present invention, wherein the steps are: S1 acquires multi-angle images of the windshield of the vehicle under test, thermal images of the heating status, and heating circuit parameters.

[0104] S2: Use a visual inspection model to perform optical screening of microcracks on multi-angle images of the windshield to obtain optical inspection results.

[0105] S3: Perform static integrity testing on the heating circuit to obtain the static test results of the heating circuit.

[0106] S4: Based on dynamic load simulation, instantaneous interruption monitoring of the heating circuit is performed to obtain the dynamic detection results of the heating circuit.

[0107] S5: Conduct a pass / fail analysis of the windshield heating process by combining the results of optical inspection, static inspection of the heating circuit, dynamic inspection of the heating circuit, and thermal imaging of the heating status, and trace the source of abnormalities based on the correlation between the inspection results.

[0108] The detailed steps are the same as those for the windshield heating test method provided in Example 1, and will not be repeated here.

[0109] Example 5: Embodiment 5 of the present invention provides a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. The processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the steps in the windshield heating detection method as described in Embodiment 1 of the present invention, wherein the steps are: S1 acquires multi-angle images of the windshield of the vehicle under test, thermal images of the heating status, and heating circuit parameters.

[0110] S2: Use a visual inspection model to perform optical screening of microcracks on multi-angle images of the windshield to obtain optical inspection results.

[0111] S3: Perform static integrity testing on the heating circuit to obtain the static test results of the heating circuit.

[0112] S4: Based on dynamic load simulation, instantaneous interruption monitoring of the heating circuit is performed to obtain the dynamic detection results of the heating circuit.

[0113] S5: Conduct a pass / fail analysis of the windshield heating process by combining the results of optical inspection, static inspection of the heating circuit, dynamic inspection of the heating circuit, and thermal imaging of the heating status, and trace the source of abnormalities based on the correlation between the inspection results.

[0114] The detailed steps are the same as those for the windshield heating test method provided in Example 1, and will not be repeated here.

[0115] The steps and methods involved in Examples 2, 3, 4 and 5 above correspond to those in Example 1. For specific implementation methods, please refer to the relevant description section of Example 1.

[0116] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed in this application can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0117] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the flow or function according to the embodiments of this application is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in or transmitted through a computer-readable storage medium. The computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless means. The computer-readable storage medium can be any available medium that a computer can access or a data processing device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium, an optical medium, or a semiconductor medium, etc.

[0118] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for detecting the heating of a windshield, characterized in that, Includes the following steps: Acquire multi-angle images of the windshield of the vehicle under test, thermal imaging of the heating status, and heating circuit parameters; A visual inspection model was used to perform optical screening of microcracks on multi-angle images of the windshield to obtain optical inspection results; A static integrity test of the heating circuit is performed to obtain the static test results of the heating circuit. Based on dynamic load simulation, transient monitoring of the heating circuit is performed to obtain dynamic detection results of the heating circuit; The passability analysis of the windshield heating process was conducted by combining the results of optical inspection, static inspection of the heating circuit, dynamic inspection of the heating circuit, and thermal imaging of the heating status. The abnormal conditions were traced based on the correlation between the inspection results.

2. The windshield heating detection method as described in claim 1, characterized in that, The preprocessing steps for the multi-angle images of the windshield of the vehicle under inspection, the thermal imaging images of the heating state, and the heating circuit parameters include: aligning the multi-angle images, thermal imaging image sequences, and circuit parameter sequences along the time axis; denoising and normalizing the brightness of the multi-angle images; calibrating the temperature values ​​of the thermal imaging images; and removing outliers and smoothing the heating circuit parameters.

3. The windshield heating detection method as described in claim 1, characterized in that, The specific steps for using a visual inspection model to perform optical screening of microcracks in multi-angle images of the windshield are as follows: A lightweight visual model is used to perform coarse crack screening on the windshield image to analyze whether there are obvious cracks and sealing defects in the windshield. A Transformer model based on a global attention mechanism is used to screen for microcracks in images without macrocracks or in areas outside the macrocrack region, and to output defect location information. The results of coarse crack screening and micro crack screening are integrated to obtain optical detection results.

4. The windshield heating detection method as described in claim 3, characterized in that, The specific steps for using a Transformer model based on a global attention mechanism to screen for microcracks in images without macrocracks or in areas outside the macrocrack region are as follows: The image without macro-cracks or the area outside the macro-crack area is divided into local regions and initial features are extracted; Based on the initial feature overlay position information; Based on the initial features obtained after overlaying positional information, a global contextual association is established through a self-attention mechanism to enhance features; The feature outputs the defect category and location based on the globally enhanced features.

5. The windshield heating detection method as described in claim 1, characterized in that, The specific steps for instantaneous monitoring of the heating circuit based on dynamic load simulation are as follows: Define transient events; The transient interruption events are counted according to the definition of transient interruption events, and the occurrence time, duration, and amplitude changes of the transient interruption events are recorded; Output dynamic detection results along with a list of transient interruption events.

6. The windshield heating detection method as described in claim 1, characterized in that, The specific steps for tracing the source of abnormalities based on the correlation between test results are as follows: Construct a rule base for association based on the correlation between detection results; Match the detection results with features in the association rule base; Output the source tracing results and generate an anomaly source tracing report.

7. A windshield heating detection system, characterized in that, include: The data acquisition module is configured to acquire multi-angle images of the windshield of the vehicle under test, thermal images of the heating status, and heating circuit parameters. The optical screening module is configured to use a visual inspection model to perform optical screening of microcracks on multi-angle images of the windshield to obtain optical inspection results. The circuit static detection module is configured to perform static circuit integrity detection on the heating circuit and obtain the static detection result of the heating circuit. The loop dynamic detection module is configured to perform transient interruption monitoring on the heating loop based on dynamic load simulation, and obtain the dynamic detection results of the heating loop. The comprehensive analysis module is configured to perform a pass / fail analysis of the windshield heating process by combining optical inspection results, static inspection results of the heating circuit, dynamic inspection results of the heating circuit, and thermal imaging of the heating status, and to trace the source of abnormalities based on the correlation between the inspection results.

8. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the windshield heating detection method as described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program adapted to be loaded by a processor and executed as described in any one of claims 1-6.

10. A computer device, characterized in that, include: A processor, adapted to execute computer programs; A computer-readable storage medium storing a computer program, which, when executed by the processor, implements the windshield heating detection method as described in any one of claims 1-6.

Images