Device for detecting hard disk

Abstract

The invention discloses a device for detecting a hard disk, relates to the technical field of storage systems, integrates an AI visual defect detection model of a neural processing unit, performs scratch recognition and optical character recognition character verification, and realizes automatic judgment of the quality of the hard disk in combination with two-dimensional data comparison logic of FIO / SMART function test. The detection state and result are displayed in real time, the universality is high, the method adapts to a complex hard disk environment, the detection efficiency is high, the detection precision is high, and the method is particularly suitable for batch detection including scenes with different protocol interfaces.

Description

Technical Field

[0001] The present application relates to the technical field of storage systems, and particularly to a device for detecting a hard disk. Background Art

[0002] Existing technologies, such as manually reading hard drive surface information or visually inspecting for defects, carry the risk of missed detections and misjudgments, lack traceability, and often rely on hard drive inspection tools specifically designed for a single interface specification. With the increase in QC (incoming inspection), the traditional, case-by-case inspection method significantly limits quality control efficiency. Summary of the Invention

[0003] The present application provides a device for detecting hard disks, which at least solves the problems in the related art of low efficiency and poor accuracy of manual identification, a wide variety of interfaces, and difficulty in universal application of traditional tools.

[0004] The present application also provides a device for detecting a hard disk, comprising: Hard disk bays, each hard disk bay having a hard disk detection slot configured to receive a hard disk; A hard disk visual inspection component comprises an upper visual module, a lower visual module, and a first controller. The upper visual module is located above the hard disk inspection slot and faces the upper surface of the hard disk connected to the hard disk bay, and is configured to capture an image of the upper surface of the hard disk. The lower visual module is located below the hard disk inspection slot and faces the lower surface of the hard disk, and is configured to capture an image of the lower surface of the hard disk. The first controller is configured to perform visual defect inspection on the hard disk based on the upper and lower surface images. A hard disk parameter detection component includes a parameter reading module and a second controller, wherein the parameter reading module is configured to obtain hard disk parameters of the hard disk, and the second controller is configured to perform functional defect detection on the hard disk based on the hard disk parameters; The interactive interface component is electrically connected to the hard disk bay, the hard disk visual detection component and the hard disk parameter detection component, and is configured to receive detection instructions about the hard disk and display detection results corresponding to the detection instructions.

[0005] Through this application, based on internal parameter information and appearance information, a logical and physical architecture for hard disk detection is constructed, the quality of the hard disk is efficiently detected from multiple angles, and the detection process is automated, thereby solving the problems in the existing technology of complicated quality inspection work, numerous equipment and tools, inaccurate manual identification, and difficult operation and use. BRIEF DESCRIPTION OF THE DRAWINGS

[0006] In order to more clearly illustrate the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0007] Figure 1 A schematic block diagram of a device for detecting a hard disk provided in an embodiment of the present application; Figure 2 A schematic diagram of a hard disk detection device provided in an embodiment of the present application from a main viewing angle; Figure 3 A schematic diagram of a side view of a device for detecting a hard disk provided in an embodiment of the present application; Figure 4 This is a schematic diagram of a device for detecting a hard disk provided in an embodiment of the present application, viewed from the rear side, opposite to the main viewing angle. DETAILED DESCRIPTION

[0008] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0009] It should be noted that, in the description of this application, the terms "comprises," "includes," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. The terms "first," "second," etc., in this application are used to distinguish similar objects, and are not used to describe a particular order or sequence.

[0010] Existing technologies include various interface specifications, including SATA, SAS, M.2, U.2, E1.S, and E3.S. Traditional tools are often specialized for a specific interface type, leading to frequent equipment changes during the inspection process. Furthermore, manually reading hard drive surface information or visually inspecting for defects carries the risk of missed detections and misjudgments, lacks traceability, and defects such as scratches, stains, and deformations are difficult to identify and compare in a standardized manner, resulting in subjective testing that impacts inspection effectiveness. Furthermore, as the number of hard drive purchase batches increases, the traditional one-by-one testing method significantly limits quality inspection efficiency, and a large number of test results are not systematically archived, hindering subsequent quality tracking and problem tracing.

[0011] In order to enable those skilled in the art to better understand the present application, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0012] To address this, an embodiment of the present application provides a device for detecting a hard disk.

[0013] The following combination Figures 1 to 4 The present invention provides an apparatus for detecting a hard disk, which is described in conjunction with a specific application environment architecture or a specific hardware architecture on which the method for detecting a hard disk depends.

[0014] Figure 1 This is a schematic block diagram of a device for detecting a hard disk provided in an embodiment of the present application. Figure 2 This is a schematic diagram of a hard disk detection device provided by an embodiment of the present application from a main viewing angle. Figure 3 A schematic diagram of a side view of a device for detecting a hard disk provided in an embodiment of the present application, and Figure 4 This is a schematic diagram of a device for detecting a hard disk provided in an embodiment of the present application, viewed from the rear side, opposite to the main viewing angle.

[0015] like Figure 1 As shown, an embodiment of the present application further provides a device for detecting a hard disk, which may at least include: Hard disk bays, each hard disk bay having a hard disk detection slot configured to receive a hard disk; A hard disk visual inspection component comprises an upper visual module, a lower visual module, and a first controller. The upper visual module is located above the hard disk inspection slot and faces the upper surface of the hard disk connected to the hard disk bay, and is configured to capture an image of the upper surface of the hard disk. The lower visual module is located below the hard disk inspection slot and faces the lower surface of the hard disk, and is configured to capture an image of the lower surface of the hard disk. The first controller is configured to perform visual defect inspection on the hard disk based on the upper and lower surface images. A hard disk parameter detection component includes a parameter reading module and a second controller, wherein the parameter reading module is configured to obtain hard disk parameters of the hard disk, and the second controller is configured to perform functional defect detection on the hard disk based on the hard disk parameters; The interactive interface component is electrically connected to the hard disk bay, the hard disk visual detection component and the hard disk parameter detection component, and is configured to receive detection instructions about the hard disk and display detection results corresponding to the detection instructions.

[0016] Specifically, the hard disk detection slot (i.e., hard disk slot) included in at least one hard disk bay can be configured as a quick and easy fixing device to facilitate the fixing of SSDs in forms such as M.2, 2.5 / 3.5, etc.

[0017] Furthermore, in some embodiments, with respect to the hard disk bay, it should be noted that the device for detecting the hard disk may also include the following configuration: an extended adapter backplane, which is constructed to be electrically connected to the hard disk bay and to distribute and switch signal channels of different protocols on the same physical bus.

[0018] Furthermore, the storage interface unit can be configured to adopt a multi-protocol compatible hard disk adapter backplane to support mixed access of at least SATA / SAS, NVMe and PCIe storage devices. One example of its technical parameters is: Regarding interface compatibility, the SATA / SAS interface supports 7-pin SATA 3.0 (6Gbps) and 36-pin SAS 12Gbps, and is compatible with 2.5-inch / 3.5-inch hard drives. The NVMe interface provides dual PCIe 4.0 x4 physical slots and supports NVMe SSDs with U.2 (SFF-8639) standard interfaces. Regarding expansion adapter functions: equipped with an M.2 Key-M interface, supporting 2280 / 22110 specifications, and realizing protocol conversion from M.2 NVMe SSD to U.2 interface through a dedicated adapter board, compatible with enterprise-level U.2 form factor storage devices, and the backplane can also be replaced for identification of other form factors such as E1.S E3.S; Regarding electrical characteristics, the backplane integrates an independent power management module, supports 12V / 5V dual power supply, and the maximum power supply power of a single slot reaches 15W, meeting the power consumption requirements of high-performance NVMe devices.

[0019] In addition, the device for detecting hard drives can adopt a signal integrity optimization design, be equipped with an EMI shielding cover and high-speed differential signal lines to ensure the stability of PCIe signal transmission.

[0020] Furthermore, the hard drive detection device is configured to include a signal detection module and a protocol conversion chip within its hardware circuitry. This dedicated signal detection module monitors electrical parameters such as the pin level status and data transfer rate when the hard drive is connected, allowing for rapid determination of the hard drive type. Furthermore, when multiple hard drives share the same physical bus, the protocol conversion chip is used to adapt the protocols.

[0021] Furthermore, in some embodiments, with respect to the hard disk visual detection component, it should be noted that the device for detecting the hard disk may also include the following configuration: a fill light array, including an upper fill light and a lower fill light, the upper fill light being constructed to face the upper surface of the hard disk, and the lower fill light being constructed to face the lower surface of the hard disk.

[0022] Furthermore, in some embodiments, with respect to the hard disk visual detection component, it should be noted that the device for detecting the hard disk may also include the following configuration: a heat sink is provided on the side and / or rear of the hard disk visual detection component.

[0023] Specifically, for the hard disk visual inspection component, the following configuration can be included: industrial-grade macro camera modules are symmetrically deployed directly above and below the hard disk slot, and a single camera is configured with a 1 / 2.5-inch CMOS image sensor (for example, but not limited to, 5 million effective pixels, resolution 2592×1944), paired with an autofocus macro lens (for example, but not limited to, focal length 8-24mm, object distance 10-50mm adjustable), supporting full-resolution image acquisition at 30 frames per second.

[0024] Specifically, for the hard disk visual detection components, the following configurations can be included: the camera module can integrate but is not limited to the following functional components: a ring-shaped fill light system, using a high color rendering index (Ra≥90) LED array, 8 groups of independently controllable lamp beads arranged around the lens, supporting 0-100% brightness adjustment and 4000K-6500K color temperature compensation, and dynamically adjusting the fill light parameters through the main control platform to eliminate reflections and shadows on the hard disk surface, ensuring an imaging contrast ≥1000:1.

[0025] Specifically, for the hard disk visual detection component, the following configuration can be included: the dual cameras are directly connected to the main control platform through the USB 3.2 Gen1 interface, and support synchronous trigger shooting mode: when the hard disk is inserted into the slot, the sensor triggers a hardware interrupt, driving the dual cameras to simultaneously capture images of the front side (label side) and back side (circuit board side) of the hard disk, and the complete imaging time of a single hard disk is ≤200ms. In some embodiments of the present application, the upper surface of the hard disk can be set as the front side of the hard disk (i.e., the label side), and the lower surface of the hard disk can be set as the back side of the hard disk (i.e., the circuit board side). Of course, in some embodiments, the upper surface of the hard disk can also be set as the back side of the hard disk (i.e., the circuit board side), and the lower surface of the hard disk can also be set as the front side of the hard disk (i.e., the label side).

[0026] More specifically, one or more of the aforementioned master control platforms may include at least a first controller or a second controller. The first controller may be configured to control the hard disk visual inspection component using the following trained neural processing unit-based visual inspection model, and the second controller may be configured to control the hard disk parameter detection component based on predetermined script commands. Furthermore, one or more of the aforementioned master control platforms may be configured to integrate a Mali-G610 MP4 GPU and a neural processing unit neural network processor, support, but not limited to, Linux / Android dual-system drivers, and interconnect with the storage interface module via, but not limited to, a PCIe 4.0x4 bus to achieve a data throughput of at least 4GB per second. The platform also includes built-in 2GB of DDR4 memory and 32GB of eMMC storage, supporting local caching and offline analysis of inspection data. In other words, the core functions of one or more master control platforms may include: driving the visual inspection unit to complete image acquisition and preprocessing; running the OCR character recognition algorithm and AI visual defect detection model; controlling the storage interface unit to implement hard disk protocol conversion, hard disk information query and data verification, and hard disk FIO performance testing; and managing touch screen interaction logic and visual output of inspection results.

[0027] In some embodiments, with respect to the hard disk visual inspection component, it should be noted that the device for inspecting the hard disk may also include the following configuration: the first controller is configured to perform visual inspection on the upper surface image and the lower surface image according to a visual inspection model based on a neural processing unit, wherein the visual inspection model is a model obtained by training a machine learning model based on deep learning through training data and an end-to-end training strategy, and the training data at least includes data acquired based on a real scene acquisition strategy and data acquired based on a data enhancement synthesis strategy. In other words, the training data uses real scene acquisition and data enhancement synthesis to cover surface scratches of normal products and defective products, etc., and is collected and synthesized through multiple channels, and an end-to-end training strategy based on deep learning is adopted. Thus, the device for inspecting the hard disk performs scratch recognition and optical character recognition character verification based on the AI visual defect detection model of the neural processing unit, and combines the two-dimensional data comparison logic of the FIO / SMART functional test to achieve automated determination of the hard disk quality.

[0028] Specifically, in some embodiments, the real-scene acquisition strategy may refer to collecting hard drive image samples in a variety of actual industrial environments to ensure the robustness of the visual inspection model. For example, image samples of hard drives of different manufacturers, models, and appearance colors may be collected; for example, image samples of various typical scratches, stains, and missing parts in different locations and shapes may be collected; for example, image samples of hard drives may be collected under various ambient lighting conditions (including but not limited to daytime indoors, nighttime, strong light, and weak light); for example, image samples of hard drives may be collected at different shooting angles and distances. Of course, in this application, the real-scene acquisition strategy may also include other acquisition strategies involved in actual application scenarios.

[0029] Specifically, in some embodiments, the data augmentation and synthesis strategy may refer to performing various image enhancements and synthesis based on real-world scene data to expand the training dataset. For example, this may include geometric transformations such as random rotation, pixel translation, and scaling; lighting perturbations such as randomly adjusting brightness, contrast, and color temperature; noise injection such as by adding Gaussian noise; and defect synthesis such as superimposing scratch textures, missing parts, or stain maps on normal hard drive images to ensure that the defect locations are consistent with the real scene. For example, background replacement may be achieved by separating the hard drive body from various industrial backgrounds and then combining them to improve the model's adaptability to complex backgrounds.

[0030] Furthermore, in some embodiments, it should be noted that the device for detecting hard disks may also include the following configuration: the visual detection model based on the neural processing unit may include at least a defect detection layer, an optical character recognition layer, or an image preprocessing layer.

[0031] Furthermore, in some embodiments, it should be noted that the hard drive inspection apparatus may also include the following configuration: an image preprocessing layer may be configured to preprocess the hard drive image based on an image distortion correction algorithm and an automatic region of interest (ROI) location algorithm to obtain preprocessed image information. Specifically, after acquiring the original image, the image is subjected to feature extraction, classification, and location. Furthermore, the image distortion correction algorithm and ROI location technology (i.e., automatic region of interest (ROI) location technology) are integrated to automatically crop the hard drive's main area, eliminating interference from slot edges. This provides standardized input data for subsequent optical character recognition and appearance defect detection (including but not limited to scratches), facilitating subsequent defect detection.

[0032] Specifically, in some embodiments, the image distortion correction algorithm may refer to correction through fisheye correction or radial / tangential distortion based on the distortion coefficients involved in the camera.

[0033] Specifically, in some embodiments, the ROI automatic location algorithm may involve using the Canny operator to extract the hard drive edge, detecting rectangular outlines through the Hough transform, screening hard drive regions based on a preset size ratio, sorting the candidate regions by confidence, and selecting the one with the highest confidence as the region of interest (ROI). For another example, the ROI may be input into a lightweight convolutional neural network (including but not limited to the MobileNet series) to extract multi-scale feature maps, perform multi-scale feature fusion and normalization, and then be used for subsequent classification and recognition.

[0034] Furthermore, in some embodiments, it should be noted that the apparatus for detecting a hard disk may further include the following configuration: the optical character recognition layer may be configured to detect whether characters on the hard disk have defects based on pre-processed image information.

[0035] Furthermore, in some embodiments, it should be noted that the apparatus for detecting a hard disk may further include the following configuration: the defect detection layer may be configured to detect whether there are defects in the appearance of the hard disk based on preprocessed image information.

[0036] For example, the preprocessed image information may include at least one of the hard disk's scratch information, defect information, or character information. Of course, in this application, the preprocessed appearance information may also include other information that affects the hard disk quality judgment.

[0037] Specifically, optical character recognition and defect models are used to determine whether there are scratches or defects on the surface, and analysis and judgment are carried out. If so, it is classified into the defect library and returned to the factory for processing; the SMART information is read from the hard disk to determine whether it is consistent with the parameter information in the database. If not, the inconsistent parameters are marked, and R&D intervenes to analyze and determine the processing method; based on the photo of the hard disk surface, it is determined whether it is consistent with the photo in the specification sheet. If not, it is returned to the factory for processing.

[0038] In some embodiments, with respect to the hard disk visual detection component, it should be noted that the device for detecting the hard disk may also include the following configuration: the second controller is configured to perform functional defect detection on the hard disk based on whether at least one of the self-monitoring analysis report technical information, serial number information or firmware information among the hard disk parameters is abnormal, and if at least one of the self-monitoring analysis report technical information, serial number information or firmware information is abnormal, the second controller determines that the hard disk fails the functional defect detection.

[0039] For example, the parameter information may include at least one of the hard drive's Self-Monitoring Analysis and Reporting Technology (SMART) information, serial number information, or firmware information. Of course, in this application, the parameter information may also include other information that affects the determination of hard drive quality.

[0040] Specifically, in some embodiments, a hard disk parameter database is constructed. For the material number in the system library, such as the unique material number, according to the specification sheet in the system, the photo of the hard disk, the modlename in the SMART information, the SN serial number, the FW and other detailed parameters are entered into the database for subsequent quality inspection and joint verification.

[0041] Furthermore, in some embodiments, with respect to the hard disk visual detection component, it should be noted that the device for detecting the hard disk may also include the following configuration: the hard disk parameter detection component also includes: a read-write function test component, which is constructed to detect whether the hard disk has failed, and if the hard disk does not pass the performance benchmark set in the read-write function test, the read-write function test component determines that the hard disk has failed.

[0042] Specifically, for example, the read / write function test may include but is not limited to an FIO (flexible I / O tester) read / write function test.

[0043] Furthermore, in some embodiments, with respect to the interactive interface component, it should be noted that the apparatus for detecting a hard disk may further include the following configuration: a heat sink is provided on the side and / or rear of the interactive interface component.

[0044] Furthermore, in some embodiments, with respect to the interactive interface component, it should be noted that the apparatus for detecting a hard disk may further include the following configuration: the interactive interface component is constructed to display the detection information of the hard disk in real time.

[0045] Specifically, during the process of detecting whether a hard disk fails or has defects, the interactive interface component may present at least one of the hard disk's detection status information, detection result information, and abnormality type information in a visual manner.

[0046] More specifically, the human-computer interaction unit can be configured to be equipped with an industrial-grade capacitive touch display, using an IPS full-viewing angle high-definition panel, such as but not limited to a resolution of 1280×800, and not limited to supporting 10-point touch operation. The display can be directly connected to the main control platform through but not limited to an HDMI 2.0 high-speed interface, and present detection data, hard disk status and image recognition results in real time. The touch screen surface is covered with but not limited to 3H hardness scratch-resistant tempered glass to adapt to high-frequency operation requirements in industrial environments, support glove mode and wet hand operation, and ensure the stability and reliability of human-computer interaction.

[0047] In addition, in some embodiments, the device for detecting hard disks of the present application also includes a component with a data traceability and management mechanism. For example, the component can archive the detection results in a local database for offline analysis, export logs, etc., making it more suitable for industrial-grade application scenarios.

[0048] The following combination Figures 2 to 4 The present invention describes a device for detecting a hard disk provided in an embodiment of the present application.

[0049] Figure 2 This is a schematic diagram of a device for detecting a hard disk provided in an embodiment of the present application from a main viewing angle.

[0050] like Figure 2 As shown, the core architecture of the hard drive detection device can also be described as consisting of four parts: a human-computer interaction unit, a main control platform, a storage interface unit, and a visual detection unit. The accompanying drawings are marked with 101, a touch screen display, which can display hard drive inspection information, fault prompts, and log export operations; 102, an upper panel; 103, a power button and a recovery button; 104, an industrial macro camera, which is used to capture images of the hard drive's appearance and has built-in algorithms for image optimization, information extraction, scratch detection, etc.; 105, a hard drive slot, equipped with a quick and easy fixing device for conveniently fixing SSDs in the form of M.2, 2.5 / 3.5, etc.; 106, a fill light bead, which is used to provide fill light; 107, a hard drive bay; 108, a lower body; 109, a USB port; 110, an RJ45 network port; 111, a fill light bead; and 112, an industrial macro camera.

[0051] Figure 3 This is a schematic diagram of a side view of a device for detecting a hard disk provided in an embodiment of the present application.

[0052] like Figure 3 As shown, in which, from a side view, the device for detecting the hard disk may include: 201, a power connector; 202, a fan heat dissipation window; 203, a touch display heat dissipation window; 204, a lower body; 205, an upper display operation body.

[0053] Figure 4This is a schematic diagram of a device for detecting a hard disk provided in an embodiment of the present application, viewed from the rear side, opposite to the main viewing angle.

[0054] like Figure 4 As shown, at a rear viewing angle opposite to the main viewing angle, the device for detecting the hard disk may include: 301, a lower heat dissipation window; 302, a touch screen heat dissipation window; 303, a back panel; 304, a touch screen heat dissipation window.

[0055] During the operation of the hard disk detection device, the hard disk to be tested is inserted into the hard disk slot. Before the program automatically starts, the FIO read and write function test and SMART information query are performed, and the corresponding data already stored in the system database is verified to ensure the quality of the incoming hard disk. At the same time, the cameras located above and below the slot will capture images of the upper and lower parts of the hard disk, and then compare them with the corresponding image information in the database to provide comparison results. After the test is completed, the hard disk quality inspection results are synchronously displayed on the display screen and a log is retained. When using this hard disk detection device, the inspection efficiency can be greatly improved, the steps of human operation and subjective inspection can be reduced, and the SMART information and appearance inspection of the hard disk can be automatically performed, which greatly saves manpower and material resources.

[0056] In summary, the hard drive detection device of this application supports parallel processing of FIO testing, SMART reading and image analysis, and displays the detection status and results in real time. This shows that this application is highly versatile and adaptable to complex hard drive environments. It also has high detection efficiency and high detection accuracy, and is particularly suitable for batch detection, including scenarios with different protocol interfaces.

[0057] Those skilled in the art may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0058] The above is a detailed introduction to the hard disk detection device provided by the present application. This article uses specific examples to illustrate the principles and implementation methods of the present application. The description of the above examples is only intended to help understand the method and core concept of the present application. It should be pointed out that for ordinary technicians in this technical field, various improvements and modifications can be made to the present application without departing from the principles of the present application, and such improvements and modifications also fall within the scope of protection of the claims of the present application.

Claims

1. A device for detecting a hard disk, characterized in that: include: Hard disk bays, each of the hard disk bays having a hard disk detection slot, the hard disk detection slot being configured to receive the hard disk; A hard disk visual inspection component comprises an upper visual module, a lower visual module, and a first controller. The upper visual module is located above the hard disk inspection slot and faces the upper surface of the hard disk received in the hard disk bay, and is configured to capture an image of the upper surface of the hard disk. The lower visual module is located below the hard disk inspection slot and faces the lower surface of the hard disk, and is configured to capture an image of the lower surface of the hard disk. The first controller is configured to perform visual defect inspection on the hard disk based on the upper and lower surface images. a hard disk parameter detection component, comprising a parameter reading module and a second controller, wherein the parameter reading module is configured to obtain hard disk parameters of the hard disk, and the second controller is configured to perform functional defect detection on the hard disk based on the hard disk parameters; The interactive interface component is electrically connected to the hard disk bay, the hard disk visual detection component and the hard disk parameter detection component, and is configured to receive detection instructions about the hard disk and display detection results corresponding to the detection instructions.

2. The device according to claim 1, characterized in that The first controller is configured to perform visual detection on the upper surface image and the lower surface image according to a visual detection model based on a neural processing unit; Among them, the visual detection model is a model obtained by training a machine learning model based on deep learning through training data and an end-to-end training strategy, and the training data at least includes data obtained based on a real scene acquisition strategy and data obtained based on a data enhancement synthesis strategy.

3. The device according to claim 2, characterized in that The neural processing unit-based visual inspection model includes at least a defect detection layer, an optical character recognition layer, or an image preprocessing layer.

4. The device according to claim 3, characterized in that The image preprocessing layer is configured to preprocess the image of the hard disk based on an image distortion correction algorithm and an automatic region of interest positioning algorithm to obtain preprocessed image information.

5. The device according to claim 4, characterized in that The optical character recognition layer is configured to detect whether characters of the hard disk have defects based on the preprocessed image information, and the defect detection layer is configured to detect whether an appearance of the hard disk has defects based on the preprocessed image information.

6. The device according to claim 1, characterized in that The second controller is configured to perform functional defect detection on the hard disk based on whether at least one of the self-monitoring analysis report technical information, serial number information or firmware information among the hard disk parameters is abnormal, and when at least one of the self-monitoring analysis report technical information, serial number information or firmware information is abnormal, the second controller determines that the hard disk fails the functional defect detection.

7. The device according to claim 1, characterized in that The hard disk parameter detection component also includes: The read / write function test component is configured to detect whether the hard disk fails, and if the hard disk fails a performance benchmark set in the read / write function test, the read / write function test component determines that the hard disk fails.

8. The device according to claim 1, characterized in that The hard disk enclosure further includes: The expansion adapter backplane is constructed to be electrically connected to the hard disk bay and to distribute and switch signal channels of different protocols on the same physical bus.

9. The device according to claim 1, characterized in that The interactive interface component is configured to display the detection information of the hard disk in real time, and a heat sink is provided on the side and / or rear of the interactive interface component.

10. The device according to claim 1, characterized in that The hard disk visual detection component also includes: The fill light array includes an upper fill light and a lower fill light, wherein the upper fill light is configured to face the upper surface of the hard disk, and the lower fill light is configured to face the lower surface of the hard disk, and a heat sink is provided on the side and / or rear of the hard disk visual detection component.

Images