Perovskite-coated foreign object detection system and method

Abstract

The application relates to a perovskite coating foreign matter detection system and method, which comprises a bearing platform, a coating die head, a processing assembly, and at least two 2.5D visual identification assemblies fixedly arranged along the width direction of a substrate and partially overlapping in a scanning area, wherein each 2.5D visual identification assembly comprises a visual scanning camera and a light source, both of which are oppositely arranged in the coating direction, and the camera collects a shadow image of a convex area formed under the oblique irradiation of the light source at a preset oblique angle. The processing assembly can determine the shadow length according to the shadow image, and determine a foreign matter identification result in combination with the shadow length. Through the scheme, the system can realize wide-width high-precision coverage by splicing multiple 2.5D visual identification assemblies in a large-span coating scene, and overcome the problems of light attenuation and precision decline caused by the lengthening of the distance of a traditional laser sensor. Meanwhile, the height of the foreign matter is indirectly measured through shadow imaging, which reduces the risk of direct interference on the tiny gap.

Description

Technical Field

[0001] This application relates to the field of solar cell technology, and in particular to a perovskite coating foreign object detection system and method. Background Technology

[0002] Currently, with the rapid development of science and technology, perovskite solar cells are gradually being developed and applied due to their high theoretical conversion efficiency. In the industrial production of perovskite solar cells, slot coating is the core process for preparing a uniform, large-area perovskite photosensitive layer. As production capacity increases, the coating area is also gradually increasing, which, while improving coating efficiency, also places more stringent requirements on coating cleanliness.

[0003] During the coating process, the gap between the coating die and the glass substrate is extremely small, as small as 50 μm. If hard protruding foreign objects are attached to the surface of the glass substrate during the cleaving, transfer or loading process, when the height of the foreign object exceeds 50 μm, it will directly interfere rigidly with the die, causing scratches on the die, coating line breaks, film layer Mura defects, etc., which may lead to the scrapping of the die and the shutdown of the production line in severe cases.

[0004] Therefore, there is an urgent need for a high-precision foreign object identification scheme to identify larger foreign object particles before coating. Summary of the Invention

[0005] Therefore, it is necessary to propose a foreign object detection system and method for perovskite coating to accurately identify large-sized foreign objects on the glass substrate before coating, so as to ensure the safe and stable coating process.

[0006] This application provides a foreign object detection system for perovskite coating, including a support platform, a coating die, a processing component, and at least two 2.5D vision recognition components. The support platform is used to place the substrate to be coated and to move the substrate along the coating direction. The coating die is fixedly disposed above the support platform. Each of the 2.5D vision recognition components is fixedly disposed above the support platform along the width direction of the substrate to be coated, and the central axes of each 2.5D vision recognition component along the width direction are parallel to each other and do not coincide with each other. The width direction is perpendicular to the coating direction.

[0007] The 2.5D visual recognition component includes a visual scanning camera and a light source. The visual scanning camera and the light source are arranged opposite to each other in the coating direction, and the scanning areas of two adjacent visual scanning cameras at least partially overlap in the width direction. The visual scanning camera acquires the shadow image of the raised area of ​​the substrate to be coated formed by the light source at a preset installation tilt angle adapted to the light source.

[0008] The processing component is connected to the visual scanning camera. The processing component is used to determine the shadow length based on the shadow image and to determine the foreign object recognition result based on the shadow length. The shadow length is the pixel length measured in the image coordinate system along the coating direction after the shadow area formed by the substrate to be coated under the illumination of the light source is acquired by the visual scanning camera.

[0009] In one embodiment, the preset installation tilt angle is 40-50 degrees; and / or, the preset tilt illumination angle is 40-50 degrees.

[0010] In one embodiment, the preset installation tilt angle is 45 degrees, and the preset tilt illumination angle is 45 degrees.

[0011] In one embodiment, in the same 2.5D visual recognition component, the first vertical distance from the visual scanning camera to the surface of the substrate to be coated is greater than the second vertical distance from the light source to the surface of the substrate to be coated.

[0012] In one embodiment, the first vertical distance is 20mm-80mm, and the second vertical distance is 5mm-15mm.

[0013] In one embodiment, the visual scanning camera is a line scan camera, and the light source is a line light source.

[0014] This application also provides a method for detecting foreign objects in perovskite coating based on the above-mentioned perovskite coating foreign object detection system, comprising: acquiring a shadow image formed by a raised area of ​​the substrate to be coated under a light source at a preset tilt angle; determining a shadow length based on the shadow image; wherein the shadow length is the pixel length of the shadow area formed by the substrate to be coated under the light source, measured along the coating direction in the image coordinate system after being acquired by the visual scanning camera; determining the equivalent protrusion height of the raised area based on the shadow length and a preset calibration coefficient; and determining the foreign object identification result based on the equivalent protrusion height.

[0015] In one embodiment, determining the foreign object identification result based on the equivalent protrusion height includes: verifying the equivalent protrusion height when the shadow length is less than or equal to a preset length threshold; and determining the presence of a target foreign object that affects the coating quality when the equivalent protrusion height is greater than a preset height threshold.

[0016] In one embodiment, determining the foreign object identification result based on the equivalent protrusion height further includes: determining the actual protrusion length and actual protrusion width of the protrusion region when the shadow length is greater than the preset length threshold; determining an actual calibration coefficient based on the shadow length and the actual protrusion length; determining the protrusion area based on the actual protrusion length and the actual protrusion width when the actual calibration coefficient is less than the preset calibration coefficient; and determining the presence of a target foreign object affecting coating quality when the protrusion area is greater than a preset area threshold.

[0017] In one embodiment, the method further includes: controlling the visual scanning camera to acquire a calibration shadow image formed by a standard steel ball with a known true height placed on the surface of the substrate to be coated at a preset installation tilt angle adapted to the light source; determining the calibration shadow length based on the calibration shadow image; and determining the preset calibration coefficient based on the calibration shadow length and the true height.

[0018] The aforementioned perovskite coating foreign object detection system and method include a support platform, a coating die, a processing component, and at least two 2.5D vision recognition components fixed along the width of the substrate with partially overlapping scanning areas. Each 2.5D vision recognition component includes a vision scanning camera and a light source, which are positioned opposite each other in the coating direction. The camera acquires shadow images of the raised areas under the tilted illumination of the light source at a preset tilt angle. The processing component determines the shadow length based on the shadow image and uses this length to determine the foreign object identification result. This approach enables the system to achieve high-precision coverage across a wide area in large-span coating scenarios by stitching together multiple 2.5D vision recognition components, overcoming the problems of light attenuation and accuracy reduction caused by increased distance in traditional laser sensors. Simultaneously, by indirectly measuring the height of foreign objects through shadow imaging, the risk of direct interference with minute gaps is reduced, thus efficiently and accurately identifying large particles that may damage the die or affect coating quality before coating. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application or the conventional technology, the drawings used in the description of the embodiments or the conventional technology will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a schematic diagram of the perovskite coating foreign object detection system in one embodiment of this application;

[0021] Figure 2This is a schematic diagram of the 2.5D visual recognition component structure in one embodiment of this application;

[0022] Figure 3 This is a schematic diagram of the acquisition of a 2.5D visual recognition component in one embodiment of this application;

[0023] Figure 4 This is a schematic diagram of a foreign object detection method for perovskite coating in one embodiment of this application;

[0024] Figure 5 This is a schematic diagram of a shadow image in one embodiment of this application;

[0025] Figure 6 This is a schematic diagram of a foreign object identification process in one embodiment of this application;

[0026] Figure 7 This is a schematic diagram of the foreign object identification process in another embodiment of this application;

[0027] Figure 8 This is a schematic diagram of coefficient calibration in one embodiment of this application. Detailed Implementation

[0028] To facilitate understanding of this application, a more complete description will be provided below with reference to the accompanying drawings. Preferred embodiments of this application are shown in the drawings. However, this application can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a thorough and complete understanding of the disclosure of this application.

[0029] Please see Figure 1 This application provides a foreign object detection system for perovskite coating, including a support platform (not shown), a coating die 110, a processing component 130, and at least two 2.5D vision recognition components 120. The support platform is used to place the substrate 200 to be coated and to move the substrate 200 to be coated along the coating direction (i.e., F in the figure). The coating die 110 is fixedly disposed above the support platform. Each 2.5D vision recognition component 120 is fixedly disposed above the support platform along the width direction of the substrate 200 to be coated. The central axes of each 2.5D vision recognition component 120 along the width direction are parallel to each other and do not coincide with each other. The width direction is perpendicular to the coating direction.

[0030] The 2.5D visual recognition component 120 includes a visual scanning camera 121 and a light source 122. The visual scanning camera 121 and the light source 122 are arranged opposite each other in the coating direction, and the scanning areas of two adjacent visual scanning cameras 121 at least partially overlap in the width direction. The visual scanning camera 121 acquires the shadow image formed by the light source 122 at a preset tilt angle adapted to the light source 122 on the raised area of ​​the substrate 200 to be coated.

[0031] The processing component 130 is connected to the visual scanning camera 121. The processing component 130 is used to determine the shadow length based on the shadow image and to determine the foreign object recognition result based on the shadow length. The shadow length is the pixel length measured along the coating direction in the image coordinate system after the shadow area formed on the substrate 200 to be coated under the illumination of the light source 122 is acquired by the visual scanning camera 121.

[0032] The perovskite coating foreign object detection system of this application is used to automatically identify protruding foreign objects on the surface of a glass substrate before the perovskite slot coating process. The coating direction refers to the direction in which the support platform moves the substrate 200 to be coated, and it is also the direction in which the coating die 110 coats the substrate 200. The coating die 110 refers to the slot coating die 110 fixedly mounted above the support platform.

[0033] The 2.5D visual recognition component 120 includes a visual scanning camera 121 and a high-brightness light source. It acquires images of the shadows cast by protruding foreign objects under illumination from the light source 122 at a specific angle, and calculates the height of the foreign object using the ratio of shadow length to the actual object height. The width direction refers to the direction perpendicular to the coating direction within the plane of the substrate 200 to be coated. The visual scanning camera 121, also known as a 2.5D high-precision scanning camera, is mounted at a specific tilt angle to acquire shadow images. The light source 122, also known as a high-brightness light source, illuminates the substrate surface at a specific tilt angle, causing the protruding foreign objects to produce shadows that can be recognized by the camera.

[0034] For further reference Figure 2 The installation tilt angle can be understood as the angle A1 between the central axis Q1 of the lens of the visual scanning camera 121 and the surface of the substrate 200 to be coated. The tilt illumination angle can be understood as the angle A2 between the illumination direction Q2 of the light source 122 and the surface of the substrate 200 to be coated.

[0035] The shadow image refers to the image captured by the camera after the shadow area formed behind the protruding foreign object is illuminated at an angle by the light source 122. The processing component 130 is a computer or controller connected to the visual scanning camera 121, used to run visual algorithms (such as template matching, blob analysis, calibration calculation, etc.). The shadow length refers to the pixel length of the shadow area measured along the coating direction in the image coordinate system; for details, please refer to [reference needed]. Figure 3 The image coordinate system refers to the two-dimensional pixel coordinate system established after the camera acquires an image, used to measure the pixel size of the shadow. The pixel length is the length of the shadow expressed in pixels, which is converted to the actual physical height through a calibration coefficient.

[0036] In long-distance slit coating, where the minimum gap between the coating die 110 and the glass is only 50µm, traditional laser triangular sensors, due to light attenuation and limitations in their operating principle, cannot achieve high sensitivity in identifying foreign objects over long spans. This embodiment addresses this by fixing at least two 2.5D visual recognition components 120 above the support platform along the width direction of the substrate. The visual scanning camera 121 and light source 122 in each component are positioned opposite each other in the coating direction, and the scanning areas of adjacent visual scanning cameras 121 partially overlap in the width direction, thereby achieving wide-area, blind-spot-free coverage.

[0037] A visual scanning camera 121 acquires a shadow image formed by light source 122 illuminating the object at a preset tilt angle (usually equal, but with some deviation within the allowable error range). The processing component 130 calculates the actual height of the foreign object based on the pixel length of the shadow. This scheme utilizes the shadow magnification effect, avoiding the defect of laser ranging attenuation with the square of the distance, and can accurately identify interference foreign objects with a height exceeding 50µm before coating.

[0038] The above solution enables high-precision wide-area coverage in large-span coating scenarios by splicing multiple 2.5D vision recognition components (120 units) together, overcoming the problems of light attenuation and accuracy reduction caused by increased distance in traditional laser sensors. Simultaneously, by indirectly measuring the height of foreign objects through shadow imaging, the risk of direct interference with minute gaps is reduced, thus efficiently and accurately identifying large foreign particles that may damage the die or affect coating quality before coating.

[0039] It is understood that the preset installation tilt angle and the preset tilt illumination angle are not unique. In one embodiment, in order to ensure that the accurate protrusion height is obtained by conversion based on the shadow length, the preset installation tilt angle and the preset tilt illumination angle need to be configured to be the same.

[0040] For example, in one embodiment, the preset installation tilt angle is 40-50 degrees; and / or, the preset tilt illumination angle is 40-50 degrees.

[0041] In one embodiment, the preset installation tilt angle and the preset tilt illumination angle are both configured to be 40-50 degrees. This angle range ensures that the light source 122 produces a sufficiently stretched shadow behind the protruding foreign object, while avoiding the shadow being too short or too flat due to an excessively steep angle.

[0042] The above scheme limits the preset installation tilt angle and preset tilt illumination angle to 40-50 degrees. This angle range can ensure that the shadow projected by the light source 122 is clearly distinguishable, and at the same time, it can make the shadow image acquired by the visual scanning camera 121 have sufficient geometric stretching effect, thereby improving the recognizability of small protrusions in the image and further enhancing the detection sensitivity and height measurement stability of the system.

[0043] In one embodiment, the preset installation tilt angle is 45 degrees, and the preset tilt illumination angle is 45 degrees. By precisely setting the installation tilt angle of the visual scanning camera 121 and the illumination tilt angle of the light source 122 to 45 degrees, this symmetrical angle design makes the geometry of the shadow as simple as possible, and the ratio of the actual height of the foreign object to the length of the shadow is approximately 1, which facilitates engineering installation and repeated positioning, and helps to maintain detection consistency in large-scale production.

[0044] The above scheme optimizes both the preset installation tilt angle and the preset tilt illumination angle to 45 degrees. This symmetrical angle design makes the geometric relationship of the shadow most linear and stable, which makes it easier for the subsequent visual algorithm to accurately convert the pixel shadow length into the actual height of the foreign object by fixing the calibration coefficient. This simplifies the calibration process while ensuring detection accuracy and improves the reliability and engineering feasibility of the system.

[0045] In one embodiment, in the same 2.5D visual recognition component 120, the first vertical distance from the visual scanning camera 121 to the surface of the substrate 200 to be coated is greater than the second vertical distance from the light source 122 to the surface of the substrate 200 to be coated.

[0046] Within the same 2.5D visual recognition component 120, the mounting height (first vertical distance) of the visual scanning camera 121 is greater than the mounting height (second vertical distance) of the light source 122. This height difference arrangement allows the light source 122 to be closer to the substrate surface, thereby illuminating protruding foreign objects at an angle, significantly elongating the shadows. A slightly higher camera position provides a wider field of view and reduces substrate reflection interference. This allows even small foreign objects (less than 50 μm) to form measurable shadow lengths in the image.

[0047] In the above scheme, within the same 2.5D visual recognition component 120, the first vertical distance from the visual scanning camera 121 to the substrate surface is greater than the second vertical distance from the light source 122 to the substrate surface. This height difference arrangement allows the light source 122 to illuminate the substrate surface at a more tilted angle, thereby forming a longer and more obvious shadow area behind the protruding foreign object. Even if the foreign object height is less than 50 μm, it can still produce a sufficiently resolvable shadow length in the image, effectively improving the system's ability to detect small foreign objects.

[0048] In one embodiment, the first vertical distance is 20mm-80mm, and the second vertical distance is 5mm-15mm.

[0049] This solution further optimizes the installation height of the visual scanning camera 121 to 20mm-80mm and the installation height of the light source 122 to 5mm-15mm. This ensures clear camera imaging and good focus while allowing the light source 122 to be tilted sufficiently to achieve optimal magnification of shadow length and object height. Simultaneously, this height combination avoids interference between the camera and the coating die 110 or other mechanisms.

[0050] The above scheme specifically configures the first vertical distance to be 20mm-80mm and the second vertical distance to be 5mm-15mm. This height range, while ensuring the imaging clarity of the visual scanning camera 121 and the illumination intensity of the light source 122, further amplifies the proportional relationship between the geometric length of the shadow and the actual height of the foreign object. This enables the system to stably identify interference foreign objects with insufficient height or exceeding 50um in long-span, high-precision pre-coating inspection scenarios, avoiding damage to the mold head or poor coating caused by blind spots in the detection.

[0051] In one embodiment, the visual scanning camera 121 is a line scan camera, and the light source 122 is a line light source 122.

[0052] The visual scanning camera 121 is a line scan camera, and the light source 122 is a line light source 122. The line scan camera, in conjunction with the uniform movement of the substrate, can continuously acquire high-resolution, distortion-free shadow images, making it particularly suitable for full-area scanning of substrates up to 2400mm in length. The line light source 122 can generate highly directional, sharp-edged linear light, making the boundaries of foreign object shadows clear and facilitating the precise extraction of the pixel length of the shadow by the Blob tool.

[0053] The above scheme, with a line scan camera and a line light source 122, can acquire high-resolution shadow images line by line during the continuous movement of the substrate 200 to be coated, making it suitable for high-speed, continuous inspection of long-distance substrates. Simultaneously, the line light source 122 can form uniform, directional linear illumination, resulting in clear and sharp shadow edges, facilitating accurate extraction of shadow length in subsequent blob analysis. This ensures both inspection efficiency and meets the accuracy requirements for micron-level height measurement.

[0054] Please see Figure 4 This application also provides a method for detecting foreign matter in perovskite coating based on the above-mentioned perovskite coating foreign matter detection system, including steps 402, 404, 406 and 408.

[0055] Step 402: Obtain the shadow image formed by the light source at a preset tilt angle on the raised area of ​​the substrate to be coated.

[0056] Step 404: Determine the shadow length based on the shadow image.

[0057] Step 406: Determine the equivalent protrusion height of the protrusion area based on the shadow length and the preset calibration coefficient.

[0058] Step 408: Determine the foreign object identification result based on the equivalent protrusion height.

[0059] The perovskite coating foreign object detection system is shown in the above embodiments and figures. The shadow length is the pixel length measured along the coating direction in the image coordinate system after the shadow area formed on the substrate 200 under illumination by the light source 122 is acquired by the visual scanning camera 121.

[0060] The equivalent protrusion height refers to the estimated actual height of the object obtained by converting the shadow pixel length using a calibration coefficient. The preset calibration coefficient is a conversion coefficient pre-calibrated using a standard steel ball of known actual height, representing the proportional relationship between the shadow length and the actual height.

[0061] First, obtain the shadow image of the raised area of ​​the substrate 200 to be coated under the oblique illumination of the light source 122. For details, please refer to [reference needed]. Figure 5 Then, using visual algorithms (template matching to locate foreign objects, blob analysis to extract contours), the pixel length of the shadow along the coating direction in the image is determined.

[0062] Following this, the equivalent protrusion height is calculated based on a preset calibration coefficient. Specifically, the equivalent protrusion height is the product of the preset calibration coefficient and the shadow length. In this way, the accuracy attenuation of laser ranging over long distances is bypassed, and the geometric proportions of 2.5D imaging are directly utilized to achieve high-precision quantitative identification of oversized foreign objects before coating.

[0063] The above solution enables high-precision wide-area coverage in large-span coating scenarios by splicing multiple 2.5D vision recognition components (120 units) together, overcoming the problems of light attenuation and accuracy reduction caused by increased distance in traditional laser sensors. Simultaneously, by indirectly measuring the height of foreign objects through shadow imaging, the risk of direct interference with minute gaps is reduced, thus efficiently and accurately identifying large foreign particles that may damage the die or affect coating quality before coating.

[0064] Please see Figure 6 In one embodiment, step 408 includes steps 602 and 604.

[0065] Step 602: If the shadow length is less than or equal to a preset length threshold, verify the equivalent protrusion height.

[0066] Step 604: If the equivalent protrusion height is greater than the preset height threshold, it is determined that there is a target foreign object that affects the coating quality.

[0067] The preset length threshold refers to a pre-set critical value for the length of shadow pixels, used for initial screening. The preset height threshold is a pre-set critical value for the actual height of foreign objects (e.g., 50µm); exceeding this value is considered a risk of collision. Target foreign objects refer to protruding foreign objects that are determined to potentially affect coating quality or damage the die head.

[0068] For flat objects (with a small height-to-length ratio), their height is usually small while their length is large. If the equivalent protrusion height is calculated solely based on the shadow length and a preset calibration coefficient, the resulting height will typically be small, easily leading to the risk of missed areas. Furthermore, these flat objects will flip and shift due to the coating solution during the coating process, causing rigid interference with the coating die 110, also posing a risk. For objects with a large height-to-length ratio, the equivalent protrusion height calculated using the shadow length and a preset calibration coefficient can effectively characterize the object's impact on the coating process, thus allowing for direct foreign object identification based on the equivalent protrusion height.

[0069] Specifically, when the shadow length is less than or equal to a preset length threshold, the calculated equivalent protrusion height is further verified. If the equivalent height is greater than the preset height threshold, it is determined that a target foreign object exists.

[0070] In some embodiments, when it is determined that a target foreign object exists, the processing component 130 can also combine analysis such as shadow images to identify the location information, actual length and width dimensions of the target foreign object, which can be configured according to actual needs.

[0071] In the above scheme, if the shadow length is less than or equal to the preset length threshold, the equivalent protrusion height is further checked. If the height is greater than the preset threshold, it is determined that there is a target foreign object. By introducing height verification, excessively tall foreign objects can be accurately identified, thereby ensuring the safety of the coating die 110.

[0072] Please see Figure 7 In one embodiment, step 408 further includes steps 701, 702, 703 and 704.

[0073] Step 701: If the shadow length is greater than a preset length threshold, determine the actual protrusion length and actual protrusion width of the protrusion area.

[0074] Step 702: Determine the actual calibration coefficient based on the shadow length and the actual protrusion length.

[0075] Step 703: If the actual calibration coefficient is less than the preset calibration coefficient, determine the protrusion area based on the actual protrusion length and the actual protrusion width.

[0076] Step 704: If the area of ​​the protrusion is greater than a preset area threshold, it is determined that there is a target foreign object that affects the coating quality.

[0077] The actual protrusion length refers to the length of the foreign object itself on the substrate plane, that is, the dimension along the coating direction. The actual protrusion width refers to the width of the foreign object itself on the substrate plane, that is, the dimension along the width direction. The actual protrusion length and actual protrusion width can be determined by taking pictures of the protrusion with the visual scanning camera 121 and then using image recognition algorithms, which will not be elaborated here.

[0078] The actual calibration factor refers to the ratio calculated from the shadow length to the actual protrusion length. The protrusion area refers to the product of the actual protrusion length and width, used to assess the impact of flat foreign objects. The preset area threshold is a pre-set critical value for the area of ​​foreign objects; exceeding this value is considered risky even if the height does not exceed the standard.

[0079] For cases where the shadow length exceeds a preset length threshold (common in flat, low-height foreign objects), the actual protrusion length and width are obtained, and an actual calibration coefficient is calculated. If the actual calibration coefficient is less than the preset calibration coefficient, it indicates that the foreign object is flat, and relying solely on the shadow height may misjudge it as exceeding the limit. Therefore, the protrusion area is calculated based on the actual protrusion length and width. If the protrusion area exceeds a preset area threshold, the presence of a target foreign object is determined.

[0080] The above solution, when the shadow length exceeds a preset length threshold, further obtains the actual protrusion length and width of the protruding area, calculates the actual calibration coefficient, and determines whether there is a risk based on the protrusion area when this coefficient is less than the preset calibration coefficient (i.e., the height-to-length ratio is small and the foreign object is flat). This solution effectively solves the problem of false alarms that may occur when flat foreign objects (such as thin sheets, low but large areas) are detected solely based on shadow height. By comprehensively judging the area, unnecessary downtime for maintenance is avoided, while still retaining risk warnings for large-area foreign objects that may flip or move, thus improving the intelligence of detection and process continuity.

[0081] In one embodiment, the method further includes: controlling a visual scanning camera 121 to acquire a calibration shadow image formed by a standard steel ball with a known true height placed on the surface of the substrate 200 to be coated at a preset mounting tilt angle adapted to the light source 122; determining the calibration shadow length based on the calibration shadow image; and determining a preset calibration coefficient based on the calibration shadow length and the true height.

[0082] A standard steel ball refers to a spherical metal particle with a known actual height (diameter) used for calibration. For example, a steel ball with a diameter of 0.2 mm can be used. The actual height (H1) refers to the actual physical diameter of the standard steel ball. The calibration shadow image is a shadow image of the standard steel ball acquired under the same installation angle and illumination angle. The calibration shadow length refers to the pixel length of the steel ball's shadow in the calibration shadow image.

[0083] For further reference Figure 8 Before formal inspection, the visual scanning camera 121 is controlled to acquire a calibration shadow image of a standard steel ball (actual height H1 = 0.2 mm) placed on the surface of the substrate 200 to be coated, at the same preset installation tilt angle (e.g., 45 degrees) as the inspection. This image is then illuminated by the light source 122 at the same preset tilt angle. Afterward, the pixel length of this shadow (i.e., the calibration shadow length L1) is extracted using a Blob tool, and the preset calibration coefficient K = H1 / L1 is calculated. This method is simple and reliable, eliminates installation errors and optical differences, ensures accurate height conversion in each actual inspection scenario, and provides a traceable benchmark for long-distance coating foreign object detection.

[0084] The above scheme uses a visual scanning camera 121 to acquire a calibration shadow image of a standard steel ball of known true height under the same light source 122 at the same angle, determines the calibration shadow length, and then calculates a preset calibration coefficient. This calibration process is simple, reproducible, and can be completed online or offline, ensuring that the conversion relationship between shadow length and object height is accurate and reliable in each actual detection scenario. This provides a unified and traceable benchmark for all subsequent height calculations, thereby ensuring the consistency of the system's detection in different batches and under different environments.

[0085] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0086] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A foreign matter detection system for perovskite coatings, characterized in that, include: A support platform is used to place the substrate to be coated and to move the substrate along the coating direction. A coating die head is fixedly mounted above the support platform; At least two 2.5D vision recognition components are provided, each of which is fixedly disposed above the support platform along the width direction of the substrate to be coated. The central axes of each 2.5D vision recognition component along the width direction are parallel to each other and do not coincide with each other. The width direction is perpendicular to the coating direction. The 2.5D visual recognition component includes a visual scanning camera and a light source. The visual scanning camera and the light source are arranged opposite to each other in the coating direction, and the scanning areas of two adjacent visual scanning cameras at least partially overlap in the width direction. The visual scanning camera acquires the shadow image of the raised area of ​​the substrate to be coated formed by the light source at a preset installation tilt angle adapted to the light source. A processing component is connected to the visual scanning camera. The processing component is used to determine the shadow length based on the shadow image and to determine the foreign object recognition result based on the shadow length. The shadow length is the pixel length measured in the image coordinate system along the coating direction after the shadow area formed by the substrate to be coated under the illumination of the light source is acquired by the visual scanning camera.

2. The foreign matter detection system for perovskite coatings according to claim 1, characterized in that, The preset installation tilt angle is 40-50 degrees; and / or, the preset tilt irradiation angle is 40-50 degrees.

3. The foreign matter detection system for perovskite coatings according to claim 2, characterized in that, The preset installation tilt angle is 45 degrees, and the preset tilt irradiation angle is 45 degrees.

4. The foreign matter detection system for perovskite coatings according to claim 1, characterized in that, In the same 2.5D visual recognition component, the first vertical distance from the visual scanning camera to the surface of the substrate to be coated is greater than the second vertical distance from the light source to the surface of the substrate to be coated.

5. The foreign matter detection system for perovskite coatings according to claim 4, characterized in that, The first vertical distance is 20mm-80mm, and the second vertical distance is 5mm-15mm.

6. The foreign matter detection system for perovskite coatings according to any one of claims 1-5, characterized in that, The visual scanning camera is a line scan camera, and the light source is a line light source.

7. A method for detecting foreign matter in perovskite coatings based on the perovskite coating foreign matter detection system according to any one of claims 1-6, characterized in that, include: Obtain the shadow image formed by the raised area of ​​the substrate to be coated at a preset tilt angle of the light source; The shadow length is determined based on the shadow image; wherein, the shadow length is the pixel length of the shadow area formed by the substrate to be coated under the illumination of the light source, measured in the image coordinate system along the coating direction after being acquired by the visual scanning camera; The equivalent protrusion height of the protrusion area is determined based on the shadow length and the preset calibration coefficient. The foreign object identification result is determined based on the equivalent protrusion height.

8. The method for detecting foreign matter in perovskite coatings according to claim 7, characterized in that, The step of determining the foreign object identification result based on the equivalent protrusion height includes: If the shadow length is less than or equal to a preset length threshold, verify the equivalent protrusion height; If the equivalent protrusion height is greater than a preset height threshold, it is determined that there is a target foreign object that affects the coating quality.

9. The method for detecting foreign matter in perovskite coatings according to claim 8, characterized in that, The step of determining the foreign object identification result based on the equivalent protrusion height further includes: If the shadow length is greater than the preset length threshold, determine the actual protrusion length and actual protrusion width of the protrusion area; The actual calibration coefficient is determined based on the shadow length and the actual protrusion length; If the actual calibration coefficient is less than the preset calibration coefficient, the protrusion area is determined based on the actual protrusion length and the actual protrusion width. If the area of ​​the protrusion is greater than a preset area threshold, it is determined that there is a target foreign object that affects the coating quality.

10. The method for detecting foreign matter in perovskite coatings according to claim 7, characterized in that, The method further includes: The visual scanning camera is controlled to acquire a calibration shadow image formed by a standard steel ball with a known actual height placed on the surface of the substrate to be coated at a preset installation tilt angle adapted to the light source. The length of the calibrated shadow is determined based on the calibrated shadow image; The preset calibration coefficient is determined based on the calibrated shadow length and the actual height.

Images