Air conditioner appearance detection equipment and detection method

Abstract

The invention discloses an air conditioner appearance detection device and method, and belongs to the technical field of appearance detection.The air conditioner appearance detection device comprises a visual detection mechanism, a multi-axis moving mechanism and an auxiliary lamplight mechanism, and a conveying mechanism is configured to convey products from back to front; the visual inspection mechanism is arranged adjacent to the conveying mechanism and is configured to be used for detecting appearance defects of the products on the conveying mechanism; the multi-axis moving mechanism is connected with the visual inspection mechanism and is configured to drive the visual inspection mechanism to move around the product in multiple axes; the auxiliary lamplight mechanism is provided with a lamplight channel, the visual detection mechanism and the conveying mechanism are at least partially arranged in the lamplight channel, the surface, facing the lamplight channel, of the auxiliary lamplight mechanism is provided with an illuminating lamp, the conveying mechanism is configured to drive products to be conveyed to penetrate through the lamplight channel, and the appearance defects of the air conditioner outdoor unit can be accurately detected.

Description

Technical Field

[0001] This invention relates to the field of shape inspection technology, and in particular to air conditioner shape inspection equipment and inspection methods. Background Technology

[0002] During the manufacturing process, the surface of an air conditioner outdoor unit is prone to defects due to manufacturing issues, such as scratches, stains, or excessive deformation.

[0003] In related technologies, the current external appearance inspection of air conditioner outdoor units is generally carried out by human eyes, which makes it difficult to accurately detect surface defects and easily leads to defective products entering the market. Summary of the Invention

[0004] The present invention aims to at least solve the technical problems existing in related technologies. To this end, the present invention proposes an air conditioner shape inspection device, which is beneficial for accurately detecting shape defects of air conditioner outdoor units.

[0005] The present invention also proposes a detection method.

[0006] An air conditioner shape inspection device according to a first aspect of the present invention includes: a conveying mechanism configured to convey a product from back to front; a vision inspection mechanism adjacent to the conveying mechanism and configured to detect shape defects of the product located on the conveying mechanism; a multi-axis moving mechanism connected to the vision inspection mechanism and configured to drive the vision inspection mechanism to move around the product via multiple axes; and an auxiliary lighting mechanism having a lighting channel, wherein the vision inspection mechanism and the conveying mechanism are at least partially disposed within the lighting channel, the auxiliary lighting mechanism has a lighting lamp on its surface facing the lighting channel, and the conveying mechanism is configured to drive the product to be conveyed through the lighting channel.

[0007] The air conditioner shape inspection device according to an embodiment of the present invention has at least the following beneficial effects: the auxiliary lighting mechanism has a lighting channel, and a lighting lamp is provided on the surface of the auxiliary lighting mechanism facing the lighting channel. The conveying mechanism and the vision inspection mechanism are both at least partially located within the lighting channel. The conveying mechanism is configured to convey the product from back to front so that the product is conveyed through the lighting channel. The lighting lamp of the auxiliary lighting mechanism can illuminate or supplement the product, making the shape features of the product easier to inspect. The multi-axis moving mechanism is connected to the vision inspection mechanism to drive the vision inspection mechanism to move around the product in multiple axes, so that the vision inspection mechanism can inspect multiple shape surfaces of the product. The vision inspection mechanism is located adjacent to the conveying mechanism and can automatically detect shape defects of the product, reduce errors caused by manual inspection, and improve inspection efficiency. Under the illumination of the lighting lamp, the vision inspection mechanism can improve the accuracy of detecting shape defects of the air conditioner outdoor unit, and reduce the risk of defective products entering the market.

[0008] According to some embodiments of the present invention, the auxiliary lighting mechanism includes a plurality of lighting frames arranged at intervals along the front-rear direction, and a plurality of lighting lamps arranged at an angle are provided on the surface of the lighting frames facing the lighting channel.

[0009] According to some embodiments of the present invention, two visual inspection mechanisms are provided, both of which are located within the light channel and respectively located on the left and right sides of the conveying mechanism.

[0010] According to some embodiments of the present invention, a visual inspection mechanism includes a visible light camera array and a 3D line laser scanner, wherein the visible light camera array is configured to acquire 2D image data of a product, and the 3D line laser scanner is configured to acquire 3D point cloud data of the product.

[0011] According to some embodiments of the present invention, the visual inspection mechanism further includes a hyperspectral imager configured to detect spectral information of the product in the visible-near-infrared band.

[0012] The detection method of the second aspect of the present invention is applied to an air conditioner shape inspection device as described in any of the first aspects; the detection method includes:

[0013] Control the conveyor mechanism to transport products from back to front; When the conveyor drives the product to the preset position, it controls the operation of the auxiliary lighting mechanism, and the lighting lamp illuminates the surface of the product. The multi-axis movement mechanism is controlled to drive the vision inspection mechanism to move around the product along multiple axes. The system controls the visual inspection mechanism to acquire the product's shape data and determines whether the product's shape conforms to preset standards based on the shape data.

[0014] According to some embodiments of the present invention, controlling a vision inspection mechanism to acquire shape data of a product includes: By using an RGB-D camera and combining it with an instance segmentation-based learning model, the product's location and external features in space can be quickly identified and located. Generate a detection path to guide the visual inspection agency based on location and external features.

[0015] According to some embodiments of the present invention, the visual inspection mechanism includes a visible light camera array; The visual inspection mechanism acquires the product's shape data and determines whether the product's shape conforms to preset standards based on the shape data, including: Control the visible light camera array to acquire 2D image data of the product, and use residual image analysis to highlight high-frequency abnormal defects in the 2D image data; Algorithms are used to analyze and detect high-frequency abnormal defects.

[0016] According to some embodiments of the present invention, the visual inspection mechanism includes a 3D line laser scanner; The visual inspection mechanism acquires the product's shape data and determines whether the product's shape conforms to preset standards based on the shape data, including: Control a 3D line laser scanner to acquire 3D point cloud data of the product, perform plane fitting on the 3D point cloud data, and obtain the fitting plane; Calculate the abrupt change in surface curvature and normal vector of the fitted plane to accurately locate pits and deformations in the product.

[0017] According to some embodiments of the present invention, the visual inspection mechanism further includes a hyperspectral imager; the inspection method includes: Collect standard spectral curves of the product's standard casing, common surface treatment agents, and common dirt, and construct a feature library based on the standard spectral curves; The system controls the hyperspectral imager to detect the spectral information of the product in the visible-near infrared band. After the spectral information is downgraded using the main components, it is input into a vector machine or analyzer, and the material of the product is analyzed according to the feature library to determine whether it deviates from the original manufacturer's standard.

[0018] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0019] The present invention will be further described below with reference to the accompanying drawings and embodiments, wherein: Figure 1 This is a schematic diagram of the structure of an air conditioner shape inspection device according to an embodiment of the present invention; Figure 2 This is a top view of an air conditioner shape inspection device according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of an air conditioner shape detection device (hidden auxiliary lighting mechanism) according to an embodiment of the present invention; Figure 4 This is a schematic diagram of the structure of a multi-axis moving mechanism of an air conditioner shape detection device according to an embodiment of the present invention; Figure 5 This is a flowchart of a detection method according to an embodiment of the present invention; Figure 6 This is a flowchart illustrating the generation of a detection path in a detection method according to an embodiment of the present invention. Figure 7 This is a flowchart illustrating the acquisition of 2D image data and 3D point cloud data of a product according to an embodiment of the present invention. Figure 8 This is a flowchart illustrating the detection method of one embodiment of the present invention for analyzing whether the material of a product deviates from the original manufacturer's standard.

[0020] Icon labels: 100. Conveying mechanism; 200. Multi-axis moving mechanism; 300. Auxiliary lighting mechanism; 310. Lighting rig; 320. Lighting channel; 400. Visual inspection agencies. Detailed Implementation

[0021] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0022] In the description of this invention, it should be understood that the orientation descriptions, such as up, down, etc., are based on the orientation or positional relationship shown in the drawings and are only for the convenience of describing this invention and simplifying the description, and are not intended to indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this invention.

[0023] In the description of this invention, "multiple" refers to two or more. The use of "first" and "second" is for distinguishing technical features only and should not be construed as indicating or implying relative importance, or implicitly indicating the number of indicated technical features or their sequential relationship.

[0024] In the description of this invention, unless otherwise explicitly defined, terms such as "set up," "install," and "connect" should be interpreted broadly, and those skilled in the art can reasonably determine the specific meaning of the above terms in this invention in conjunction with the specific content of the technical solution.

[0025] Reference Figures 1 to 4 As shown, an air conditioner shape inspection device according to an embodiment of the present invention includes: a conveying mechanism 100, a vision inspection mechanism 400, a multi-axis moving mechanism 200, and an auxiliary lighting mechanism 300.

[0026] Reference Figure 1 , Figure 2 and Figure 3 As shown, the conveying mechanism 100 can specifically be a linear conveying mechanism such as a belt conveyor or a roller conveyor. The conveying mechanism 100 is configured to convey products from back to front to achieve the arrangement and conveying of multiple products. The product can specifically be an air conditioner outdoor unit.

[0027] Reference Figure 1 , Figure 2 and Figure 3As shown, the auxiliary lighting mechanism 300 has a lighting channel 320, and the vision inspection mechanism 400 and the conveying mechanism 100 are at least partially disposed within the lighting channel 320. The auxiliary lighting mechanism 300 has an illumination lamp on its surface facing the lighting channel 320, and the conveying mechanism 100 is configured to drive the product to be conveyed through the lighting channel 320.

[0028] Reference Figure 1 , Figure 2 and Figure 3 As shown, the lighting can illuminate or supplement the product that passes through the light channel 320, making the product's shape features easier to detect or identify, such as dents, scratches or stains on the product surface.

[0029] Reference Figure 1 , Figure 2 and Figure 3 As shown, the multi-axis moving mechanism 200 is connected to the vision inspection mechanism 400 to drive the vision inspection mechanism 400 to move around the product along multiple axes, so that the vision inspection mechanism 400 can perform shape inspection on multiple surfaces of the product, thereby improving the efficiency of shape inspection of the product.

[0030] Reference Figure 1 , Figure 2 and Figure 3 As shown, compared with the traditional manual inspection method, the visual inspection mechanism 400 of the air conditioner shape inspection equipment provided in this embodiment of the invention is located adjacent to the conveying mechanism 100. It can automatically detect the shape defects of the product, reduce the error caused by manual inspection, and improve the inspection efficiency. Under the illumination of the lamp, the visual inspection mechanism 400 can improve the detection accuracy of the shape defects of the air conditioner outdoor unit.

[0031] Reference Figure 1 , Figure 2 and Figure 3 As shown, it can be understood that the auxiliary lighting mechanism 300 includes a plurality of lighting frames 310 arranged at intervals along the front-back direction, and the surfaces of the lighting frames 310 facing the lighting channel 320 are provided with a plurality of lighting lamps arranged at an angle.

[0032] Reference Figure 1 , Figure 2 and Figure 3 As shown, each light fixture 310 includes multiple lighting lamps arranged at an angle, thereby providing illumination or supplemental lighting to the product from multiple angles, making the product's shape features easier to detect, and thus improving the detection accuracy and efficiency of the visual inspection agency 400.

[0033] Reference Figure 1 , Figure 2 and Figure 3As shown, multiple light holders 310 are arranged at intervals along the front-to-back direction, that is, the arrangement direction of the multiple light holders 310 is consistent with the product conveying direction. By arranging them at appropriate intervals, the manufacturing cost of the auxiliary light channel 320 can be effectively reduced while achieving supplementary lighting, and the lighting on the product surface can be made more uniform.

[0034] It should be understood that in some other embodiments, the auxiliary lighting channel 320 is equipped with lighting lamps of various specifications to enable the detection of different types of defects in the product. Multiple lighting lamps can be mounted on different lighting racks 310, allowing for various shape inspections of the product as it passes through multiple lighting racks 310.

[0035] Reference Figure 1 , Figure 2 and Figure 3 As shown, specifically, various lighting fixtures may include LEDs that integrate white light (adjustable color temperature), UV-A (365nm), and specific near-infrared bands (such as 850nm, 970nm) to excite different characteristics.

[0036] Reference Figure 1 , Figure 2 and Figure 4 As shown, it can be understood that in this embodiment, two visual inspection mechanisms 400 are provided, both of which are located within the light channel 320 and respectively located on the left and right sides of the conveying mechanism 100.

[0037] Reference Figure 1 , Figure 2 and Figure 4 As shown, specifically, two multi-axis moving mechanisms 200 are provided, and the two multi-axis moving mechanisms 200 are respectively connected to two vision inspection mechanisms 400 to drive the two vision inspection mechanisms 400 to operate synchronously, thereby realizing the shape inspection of multiple exterior surfaces of the product.

[0038] Reference Figure 1 , Figure 2 and Figure 4 As shown, for example, the visual inspection unit 400 located on the left can perform shape inspection on the left, front, and rear sides of the product, while the visual inspection unit 400 located on the right can perform shape inspection on the right and top sides of the product. This air conditioner shape inspection equipment can simultaneously inspect multiple exterior surfaces of the product through the synchronous inspection of the two visual inspection units 400, thereby improving inspection efficiency.

[0039] Reference Figure 1 , Figure 2 and Figure 4As shown, it is understood that the visual inspection agency 400 includes a visible light camera array and a 3D line laser scanner, the visible light camera array being configured to acquire 2D image data of the product and the 3D line laser scanner being configured to acquire 3D point cloud data of the product.

[0040] Reference Figure 1 , Figure 2 and Figure 4 As shown, specifically, the visual inspection agency 400 can use a high-resolution visible light camera array to acquire 2D image data of the product, thereby identifying 2D shape defects such as markings and corrosion detection.

[0041] Reference Figure 1 , Figure 2 and Figure 4 As shown, specifically, the visual inspection agency 400 can acquire 3D point cloud data of the product's shell with millimeter-level precision using a 3D line laser scanner, which can be used to detect morphological defects such as dents, deformations, and assembly gaps on the product's surface.

[0042] Reference Figure 1 , Figure 2 and Figure 4 As shown, it is understandable that the market contains "refurbished" products with yellowed casings bleached with chemical agents (such as deyellowing agents containing the carcinogen titanium dioxide) and arbitrarily affixed with counterfeit energy efficiency labels and production date labels. Consumers find it difficult to distinguish old or overdue air conditioners from their appearance, making them prone to purchasing high-energy-consuming products with safety and health hazards. This air conditioner appearance inspection equipment can provide a return-to-factory inspection and repair solution.

[0043] Reference Figure 1 , Figure 2 and Figure 4 As shown, the visual inspection unit 400 also includes a hyperspectral imager configured to detect spectral information of the product in the visible-near-infrared band.

[0044] Reference Figure 1 , Figure 2 and Figure 4 As shown, the hyperspectral imager can detect spectral information in the visible-near infrared band (400-1700nm). By using the unique spectra of different materials (such as virgin ABS plastic, refurbished coatings, and contaminants), it can non-destructively identify illegal refurbishments (such as paint cover-ups), abnormal materials, and types of dirt on condenser fins.

[0045] Reference Figure 1 , Figure 2 and Figure 4 As shown, this air conditioner appearance inspection equipment can efficiently detect whether a product has been refurbished using a hyperspectral imager, thereby assisting users in judging whether the product has been refurbished.

[0046] Reference Figures 5 to 8 As shown, the present invention provides a detection method applied to an air conditioner shape inspection device as described in any of the above embodiments. The detection method includes the following steps: Step S100: Control the conveyor mechanism 100 to convey the product from back to front; In step S200, when the conveying mechanism 100 drives the product to the preset position, the auxiliary lighting mechanism 300 is controlled to operate, and the lighting lamp provides supplementary lighting to the surface of the product. Step S300: Control the multi-axis moving mechanism 200 to drive the vision inspection mechanism 400 to move around the product in multiple axes; In step S400, the vision inspection mechanism 400 is controlled to acquire the shape data of the product and determine whether the shape of the product conforms to the preset standard based on the shape data.

[0047] Reference Figure 1 , Figure 2 and Figure 5 As shown, in step S100, the conveying mechanism 100 can specifically be a linear conveying mechanism such as a belt conveyor or a roller conveyor. The conveying mechanism 100 is configured to convey products from back to front to realize the arrangement and conveying of multiple products.

[0048] Reference Figure 1 , Figure 2 and Figure 5 As shown, in step S200, the auxiliary lighting mechanism 300 has a lighting channel 320, the visual inspection mechanism 400 and the conveying mechanism 100 are at least partially disposed in the lighting channel 320, the auxiliary lighting mechanism 300 has an illumination lamp on its surface facing the lighting channel 320, and the conveying mechanism 100 is configured to drive the product to be conveyed through the lighting channel 320.

[0049] Reference Figure 1 , Figure 2 and Figure 5 As shown, the lighting can illuminate or supplement the product that passes through the light channel 320, making the product's shape features easier to detect or identify, such as dents, scratches or stains on the product surface.

[0050] Reference Figure 1 , Figure 2 and Figure 5 As shown, in steps S300 and S400, the multi-axis moving mechanism 200 is connected to the vision inspection mechanism 400 to drive the vision inspection mechanism 400 to move around the product via multiple axes, so that the vision inspection mechanism 400 can perform shape inspection on multiple surfaces of the product, thereby improving the efficiency of shape inspection of the product.

[0051] Reference Figure 1, Figure 2 and Figure 5 As shown, compared with the traditional manual inspection method, the inspection method provided by the present invention can automatically detect the shape defects of the product, reduce the error caused by manual inspection, and improve the inspection efficiency. The visual inspection mechanism 400, under the illumination of the lighting lamp, can improve the detection accuracy of the shape defects of the air conditioner outdoor unit.

[0052] Reference Figure 1 , Figure 2 and Figure 6 As shown, it can be understood that in step S400 of this detection method, the vision inspection mechanism 400 is controlled to acquire the shape data of the product, including the following steps: Step S410: Using an RGB-D camera and a learning model based on instance segmentation, quickly identify and locate the product's position and external features in space. Step S420: Generate a detection path to guide the visual inspection mechanism 400 based on the location and external features.

[0053] In step S410, the air conditioner shape detection device can first use an RGB-D camera, combined with an instance segmentation-based deep learning model (such as Mask R-CNN), to quickly identify and locate the outdoor unit's position in space and its main components (host, fan, pipe interface).

[0054] In step S420, the detection method can generate a preliminary 3D bounding box of the outdoor unit based on the acquired position of the product in space and its main components, thereby guiding the vision inspection agency 400 to perform a fine inspection path and establishing a connection with the device's digital passport.

[0055] Reference Figure 1 , Figure 2 and Figure 6 As shown, the visual inspection mechanism 400 synchronously triggers all sensors along the planned inspection path. The key is to perform precise temporal and spatial registration of visible light images, 3D point clouds, and hyperspectral cube data to ensure that each physical point corresponds one-to-one in each modal data.

[0056] Reference Figure 1 , Figure 2 and Figure 7 As shown, it is understood that the visual inspection mechanism 400 of the air conditioner shape inspection equipment includes a visible light camera array; Reference Figure 1 , Figure 2 and Figure 7 As shown, in step S400 of this detection method, the vision inspection mechanism 400 is controlled to acquire the shape data of the product, and the shape data is used to determine whether the shape of the product conforms to the preset standard. This includes the following steps: Step S430: Control the visible light camera array to acquire 2D image data of the product, and use residual image analysis to highlight high-frequency abnormal defects in the 2D image data; Step S440: The algorithm is used to analyze and detect high-frequency abnormal defects.

[0057] Reference Figure 1 , Figure 2 and Figure 7 As shown, specifically, this detection method can use an improved YOLO-v8 or DETR model to detect labels and screws on the product surface. For minor scratches, high-frequency residual image analysis is used (the background after Gaussian blurring is subtracted from the original image to highlight micro-texture anomalies).

[0058] Reference Figure 1 , Figure 2 and Figure 7 As shown, the essence of high-frequency residual image analysis is to use digital image processing technology to strip or suppress low-frequency "normal background" information in the image, thereby highlighting and enhancing high-frequency "abnormal defect" information, making it easy to identify with the naked eye and detectable by algorithms.

[0059] Reference Figure 1 , Figure 2 and Figure 7 As shown, high-frequency residual image analysis is a powerful method specifically designed to detect weak, small, and low-contrast defects in images, especially suitable for minor scratches, shallow dents, paint particles, uneven textures, and other imperfections that are difficult to detect visually.

[0060] Reference Figure 1 , Figure 2 and Figure 7 As shown, it is understood that the visual inspection agency 400 includes a 3D line laser scanner.

[0061] In step S400 of this detection method, the vision inspection mechanism 400 is controlled to acquire the product's shape data, and the product's shape is determined to meet the preset standard based on the shape data. This includes the following steps: Step S440: Control the 3D line laser scanner to acquire the 3D point cloud data of the product, perform plane fitting on the 3D point cloud data, and obtain the fitting plane. Step S450: Calculate the abrupt change in the surface curvature and normal vector of the fitted plane to accurately locate the pits and deformations of the product.

[0062] Reference Figure 1 , Figure 2 and Figure 7As shown, this detection method can accurately locate pits and deformations by acquiring 3D point cloud data of the product, performing plane fitting on the acquired 3D point cloud data to obtain the fitting plane, and then calculating the abrupt change in surface curvature and normal vector.

[0063] Reference Figure 1 , Figure 2 and Figure 7 As shown, the detection method can specifically identify pits and excessive deformations on the product surface, thereby improving the accuracy of detecting the product's shape features.

[0064] Reference Figure 1 , Figure 2 and Figure 8 As shown, it is understood that the visual inspection agency 400 also includes a hyperspectral imager; the inspection method further includes the following steps: Step S500: Collect standard spectral curves of the product's standard shell, common surface treatment agents, and common dirt, and construct a feature library based on the standard spectral curves; Step S600: Control the hyperspectral imager to detect the spectral information of the product in the visible-near infrared band, input the spectral information into a vector machine or analyzer after downsampling the main components, and analyze whether the material of the product deviates from the original manufacturer's standard based on the feature library.

[0065] Reference Figure 1 , Figure 2 and Figure 8 As shown, hyperspectral imagers can detect spectral information in the visible-near-infrared band (400-1700nm). Different materials (such as virgin ABS plastic, refurbished coatings, and contaminants) have unique spectra. This detection method, using hyperspectral imagers, can non-destructively identify unauthorized refurbishments (such as paint cover-ups), material abnormalities, and types of fouling on condenser fins.

[0066] Reference Figure 1 , Figure 2 and Figure 8 As shown, in step S500, the detection method can collect standard spectral reflectance curves of standard original materials (such as new ABS shells, original paint, clean aluminum fins, and standard sealant) under specific illumination in advance in the laboratory or on the production line, and store them in the feature library.

[0067] Reference Figure 1 , Figure 2 and Figure 8 As shown, in step S600, during detection, the hyperspectral imager acquires a continuous spectral curve from visible light to near infrared for each pixel, and compares this measured curve with the standard curve in the database using an algorithm.

[0068] Reference Figure 1 , Figure 2 and Figure 8 As shown, for example, in the case of illegal chemical refurbishment, the reflectivity characteristics of this detection method in a specific band (such as 700-1100nm near infrared) are systematically different from those of the original material, and the refurbished coating may mask the intrinsic spectrum of the material aging.

[0069] Reference Figure 1 , Figure 2 and Figure 8 As shown, for example, in the case of material abnormalities / inferior materials, the detection method shows differences in the color spectrum in the visible light band (400-700nm) and the absorption characteristics of organic molecules (such as CH bonding) in the short-wave near-infrared band.

[0070] Reference Figure 1 , Figure 2 and Figure 8 As shown, for example, in the case of oil pollution, this detection method exhibits a reflectance "concave" in the characteristic absorption bands of oil, such as 930nm, 1210nm, 1400nm, and 1720nm.

[0071] Reference Figure 1 , Figure 2 and Figure 8 As shown, the adaptive auxiliary lighting mechanism 300 can switch to the most suitable combination of bands to excite the target characteristics during detection. For example, to detect a specific refurbishing agent, it can switch to the UV and specific near-infrared bands that can excite its fluorescence or characteristic absorption.

[0072] Reference Figure 1 , Figure 2 and Figure 8 As shown, this detection method can improve the efficiency and accuracy of product detection by cooperating with the auxiliary lighting mechanism 300 and the hyperspectral imager.

[0073] It should be noted that the methods of using principal component de-positioning, the working principles of vector machines or analyzers are conventional techniques in this field and will not be elaborated here.

[0074] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited to the above embodiments. Within the scope of knowledge possessed by those skilled in the art, various changes can be made without departing from the spirit of the present invention.

Claims

1. An air conditioner shape inspection device, characterized in that, include: The conveying mechanism (100) is configured to convey products from back to front; A visual inspection unit (400), adjacent to the conveying mechanism (100), is configured to detect shape defects of the product located on the conveying mechanism (100); A multi-axis moving mechanism (200), connected to the vision inspection mechanism (400), is configured to drive the vision inspection mechanism (400) to move around the product via multiple axes; An auxiliary lighting mechanism (300) has a lighting channel (320), wherein the visual inspection mechanism (400) and the conveying mechanism (100) are at least partially disposed within the lighting channel (320), the auxiliary lighting mechanism (300) has an illumination lamp on its surface facing the lighting channel (320), and the conveying mechanism (100) is configured to drive the product to be conveyed through the lighting channel (320).

2. The air conditioner shape inspection equipment according to claim 1, characterized in that, The auxiliary lighting mechanism (300) includes a plurality of lighting frames (310) arranged at intervals along the front-back direction, and the surface of the lighting frame (310) facing the lighting channel (320) is provided with a plurality of lighting lamps arranged at an angle.

3. The air conditioner shape inspection equipment according to claim 1, characterized in that, Two visual inspection mechanisms (400) are provided, both of which are located within the light channel (320) and respectively located on the left and right sides of the conveying mechanism (100).

4. The air conditioner shape inspection equipment according to claim 1, characterized in that, The visual inspection unit (400) includes a visible light camera array and a 3D line laser scanner, the visible light camera array being configured to acquire 2D image data of the product and the 3D line laser scanner being configured to acquire 3D point cloud data of the product.

5. The air conditioner shape inspection equipment according to claim 1, characterized in that, The visual inspection mechanism (400) also includes a hyperspectral imager configured to detect spectral information of the product in the visible-near-infrared band.

6. The detection method, characterized in that, Applied to the air conditioner shape inspection equipment as described in any one of claims 1 to 5; the inspection method includes: The conveying mechanism (100) is controlled to convey the product from back to front; When the conveying mechanism (100) drives the product to a preset position, it controls the operation of the auxiliary lighting mechanism (300), and the lighting lamp provides supplementary lighting to the surface of the product; The multi-axis movement mechanism (200) is controlled to drive the vision inspection mechanism (400) to move around the product in a multi-axis manner; The visual inspection mechanism (400) is controlled to acquire the shape data of the product, and the shape data is used to determine whether the shape of the product conforms to the preset standard.

7. The detection method according to claim 6, characterized in that, The control of the vision inspection mechanism (400) to acquire the shape data of the product includes: Using an RGB-D camera and a learning model based on instance segmentation, the product's location and external features in space can be quickly identified and located. A detection path is generated based on the location and the external features to guide the visual inspection mechanism (400).

8. The detection method according to claim 6, characterized in that, The visual inspection mechanism (400) includes a visible light camera array; The process of controlling the visual inspection mechanism (400) to acquire the shape data of the product and determining whether the shape of the product conforms to a preset standard based on the shape data includes: The visible light camera array is controlled to acquire 2D image data of the product, and residual image analysis is used to highlight high-frequency abnormal defects in the 2D image data. An algorithm is used to analyze and detect the high-frequency abnormal defects.

9. The detection method according to claim 6, characterized in that, The visual inspection mechanism (400) includes a 3D line laser scanner; The process of controlling the visual inspection mechanism (400) to acquire the shape data of the product and determining whether the shape of the product conforms to a preset standard based on the shape data includes: The 3D line laser scanner is controlled to acquire 3D point cloud data of the product, and the 3D point cloud data is fitted to a plane to obtain a fitted plane. Calculate the abrupt change in the surface curvature and normal vector of the fitted plane to accurately locate the pits and deformations of the product.

10. The detection method according to claim 6, characterized in that, The visual inspection mechanism (400) further includes a hyperspectral imager; the inspection method includes: Standard spectral curves of the product's standard outer shell, common surface treatment agents, and common dirt are collected, and a feature library is constructed based on the standard spectral curves. The hyperspectral imager is controlled to detect the spectral information of the product in the visible-near infrared band. The spectral information is then downsampled using the main components and input into a vector machine or analyzer. Based on the feature library, the material of the product is analyzed to determine whether it deviates from the original manufacturer's standard.

Images