Multi-dimensional vision inspection system and method

Abstract

The application belongs to the technical field of image analysis, and particularly relates to image data processing, and more particularly to a multi-dimensional visual detection system and a detection method, which comprises that both sides of the shooting module are provided with sound collecting modules electrically connected with an upper computer; the upper computer is configured to control the shooting module to shoot slurry on the material to obtain slurry image data, and then identify defects according to the slurry image data, that is, when the gray value of the slurry area in the slurry image data is greater than a preset threshold value, the upper computer directly identifies defects, when the gray value of the slurry area in the slurry image data is less than the preset threshold value, the upper computer controls the sound collecting module to collect sound data, embeds the sound data in the slurry image data after processing to form composite image data, and judges the fault position of the nozzle according to the composite image data, thereby realizing rapid and accurate judgment of the slight abnormality of the nozzle and determination of the fault position of the nozzle.

Description

Technical Field

[0001] This invention belongs to the field of image analysis technology, specifically relating to image data processing, and more particularly to a multidimensional visual detection system and detection method. Background Technology

[0002] On high-speed spraying production lines, slurry is extruded from the micron-level gaps of nozzles under constant pressure. Minor anomalies in its flow due to nozzle malfunctions, such as uneven flow rates caused by agglomerates, gels, or slight blockages, can directly damage the uniformity of coating thickness, resulting in defects such as dark spots, streaks, or thickness deviations after drying. While pure vision systems can capture macroscopic defects in the coating that have already appeared, they cannot identify minute anomalies in a timely and accurate manner. By the time the slurry dries and reveals the flaws caused by these minute anomalies, several meters to tens of meters of raw material already have defects.

[0003] Therefore, since pure visual inspection cannot identify minute anomalies in a timely manner, there is a technical problem that when a minute anomaly is detected, there are already several meters to tens of meters of raw materials in the actual production line. It is necessary to design a multi-dimensional visual inspection system and inspection method for nozzles.

[0004] It should be noted that the information disclosed in this background section is only for understanding the background technology of the present application concept, and therefore, the above description is not considered to constitute prior art information. Summary of the Invention

[0005] This disclosure provides at least one multidimensional visual inspection system and inspection method.

[0006] In a first aspect, embodiments of this disclosure provide a multi-dimensional visual inspection system, including:

[0007] A host computer, and several imaging modules electrically connected to the host computer;

[0008] Both sides of the shooting module are equipped with sound acquisition modules that are electrically connected to the host computer.

[0009] The host computer is configured to control the imaging module to capture images of the slurry on the material to obtain slurry image data. Then, it identifies defects based on the slurry image data. Specifically, when the gray value of the slurry area in the slurry image data is greater than a preset threshold, the host computer directly identifies the defect. When the gray value of the slurry area in the slurry image data is less than the preset threshold, the host computer controls the sound acquisition module to acquire sound data. After processing the sound data, it embeds it into the slurry image data to obtain composite image data. Based on the composite image data, it determines the fault location of the nozzle.

[0010] In one alternative embodiment, the imaging module is positioned on both sides of the nozzle in the width direction for spraying slurry onto the material;

[0011] The line connecting the two sound acquisition modules corresponding to the shooting module is parallel to the length of the nozzle.

[0012] To obtain various sound data, embed the sound data into the slurry image data to form composite image data, and determine the fault location of the nozzle based on the composite image data.

[0013] In one optional implementation, the host computer is further configured to perform resolution calibration on the image data, i.e.,

[0014] After the nozzles start spraying the slurry, the resolution of all shooting modules is obtained. The image captured by the lowest resolution shooting module is used as the reference frame. The slurry image data captured by each shooting module is standardized in size and resolution according to the reference frame, and a unique timeline is established to obtain several time packets.

[0015] By constructing time packets, the slurry image data (3) is synchronized with the corresponding sound data;

[0016] The time packet corresponds to a preset time length.

[0017] The time packet includes: slurry image data and sound data; wherein

[0018] The length of each audio data segment is the preset time-domain data interval.

[0019] In one optional implementation, the host computer is further configured to verify each time packet, that is, to determine whether the number of valid data in each time packet is consistent. When the number of complete slurry image data and the number of complete sound data in each time packet are the same, the time packet verification is determined to be successful. Otherwise, the preset time length is increased, and then the time packet is reacquired and verified again until the time packet verification is successful.

[0020] In one optional implementation, the host computer is further configured to adjust the valid data within the time packet according to the response speed requirement after successful time packet verification; when a fast response is required, the valid data within the time packet is reduced; when a fast response is not required, the valid data within the time packet remains unchanged.

[0021] After adjusting the effective data, incomplete data within the time packet is removed, that is, slurry image data that is different from the reference frame size and sound data that is smaller than the preset time domain data interval are removed to form the initial time packet, and the time period corresponding to the removed data is marked.

[0022] In an optional implementation, the host computer is further configured to, during the process of spraying slurry through the nozzle, when it is determined that the gray value of the slurry area in the slurry image data is less than a preset threshold, collect data according to the initial time packet. If the actual time-domain data spacing corresponding to the collected sound data at this time is the same as the preset time-domain data spacing, then add barcode areas on the left and right sides of each slurry image data in the valid data, convert the sound data corresponding to the slurry image data into a spectrogram and record it in the corresponding barcode area to form composite image data. Based on the spectrogram in the composite image data, it is determined whether the sound data collected by the corresponding sound acquisition module is abnormal, and if abnormal, it is determined that the nozzle at the corresponding sound acquisition module is faulty, and then the fault location of the nozzle is determined.

[0023] In one optional implementation, the host computer is further configured to, during the process of spraying slurry from the nozzle, when it is determined that the gray value of the slurry area in the slurry image data is less than a preset threshold, collect data according to the initial time packet, and if the actual time domain data spacing corresponding to the collected sound data at this time is not the same as the preset time domain data spacing, then adjust the initial time packet to form a new time packet.

[0024] In one optional implementation, the method of adjusting the initial time packet to form a new time packet includes:

[0025] The host computer re-inputs the removed time segments into the initial time packets to form a complete time length. It then re-divides the complete time length of the initial time packets according to the actual time domain data intervals. If the number of audio data corresponding to the complete actual time domain data intervals in each initial time packet is the same after division, the incomplete time data is removed to form a new time packet. If the number of audio data corresponding to the complete actual time domain data intervals in each initial time packet is different after division, the complete time length of the initial time packets is increased until the number of audio data corresponding to the complete actual time domain data intervals in each initial time packet is the same. Then, the incomplete time data is removed to form a new time packet.

[0026] In one optional implementation, the host computer is further configured to collect data according to the new time packet, add barcode areas on both sides of each slurry image data in the valid data, convert the sound data corresponding to the slurry image data into a spectrogram and record it in the corresponding barcode area to form composite image data.

[0027] Secondly, this disclosure also provides a detection method using the above-described multidimensional vision detection system, comprising:

[0028] The host computer controls the imaging module to capture images of the slurry on the material to obtain slurry image data. Then, defects are identified based on the slurry image data. Specifically, when the gray value of the slurry area in the slurry image data is greater than a preset threshold, the host computer directly identifies the defect. When the gray value of the slurry area in the slurry image data is less than the preset threshold, the host computer controls the sound acquisition module to collect sound data. The sound data is then processed and embedded into the slurry image data to obtain composite image data. Based on the composite image data, the location of the nozzle fault is determined.

[0029] The beneficial effects of this invention are as follows: This multi-dimensional visual inspection system includes a host computer and several imaging modules electrically connected to the host computer; both sides of each imaging module are equipped with sound acquisition modules electrically connected to the host computer; the host computer is configured to control the imaging modules to capture images of the slurry on the material to obtain slurry image data, and then identify defects based on the slurry image data. Specifically, when the gray value of the slurry area in the slurry image data is greater than a preset threshold, the host computer directly identifies the defect; when the gray value of the slurry area in the slurry image data is less than the preset threshold, the host computer controls the sound acquisition modules to collect sound data, processes the sound data and embeds it into the slurry image data to obtain composite image data, and determines the fault location of the nozzle based on the composite image data, thereby achieving rapid and accurate judgment of minor nozzle anomalies and determination of the nozzle fault location.

[0030] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention are realized and obtained through the structures particularly pointed out in the description and the drawings.

[0031] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described in detail below with reference to the accompanying drawings. Attached Figure Description

[0032] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0033] Figure 1 A schematic diagram of a multidimensional vision inspection system provided in this disclosure embodiment;

[0034] Figure 2 This is a schematic diagram of the arrangement of a shooting module provided in an embodiment of the present disclosure;

[0035] Figure 3 A flowchart for the formation of an initial time packet is provided as an embodiment of this disclosure;

[0036] Figure 4 This is a schematic diagram of the barcode area provided in an embodiment of this disclosure.

[0037] In the picture:

[0038] 1 nozzle;

[0039] 2 shooting modules, 21 sound acquisition modules;

[0040] 3. Slurry image data, 31 barcode areas. Detailed Implementation

[0041] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0042] As used herein, the phrases “in one embodiment,” “according to one embodiment,” “in some embodiments,” etc., generally refer to the fact that a particular feature, structure, or characteristic following the phrase can be included in at least one embodiment of this disclosure. Therefore, a particular feature, structure, or characteristic can be included in more than one embodiment of this disclosure, such that these phrases do not necessarily refer to the same embodiment. As used herein, the terms “example,” “exemplary,” etc., are used to “serve as an example, instance, or illustration.” Any implementation, aspect, or design described herein as “example” or “exemplary” is not necessarily to be construed as preferred or superior to other implementations, aspects, or designs. Rather, the use of the terms “example,” “exemplary,” etc., is intended to present concepts in a specific manner.

[0043] On high-speed spraying production lines, slurry is extruded from the micron-level gaps in the nozzles under constant pressure. Minor anomalies in its flow state, such as uneven flow rate caused by agglomerates, gels, or slight blockages, can directly damage the uniformity of the coating thickness, resulting in defects such as dark spots, streaks, or thickness deviations after drying. While pure vision systems can capture macroscopic defects in the coating that have already appeared, they cannot identify minute anomalies in a timely and accurate manner. By the time the slurry dries and reveals the flaws caused by these minute anomalies, several meters to tens of meters of raw material already have defects.

[0044] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0045] The following detailed description of some embodiments of the present invention is provided in conjunction with the accompanying drawings. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0046] like Figure 1 As shown, at least one disclosed embodiment provides a multi-dimensional visual inspection system, including: a host computer, and several imaging modules 2 electrically connected to the host computer; both sides of the imaging modules 2 are provided with sound acquisition modules 21 electrically connected to the host computer; the host computer is configured to control the imaging modules 2 to capture images of the slurry on the material to obtain slurry image data 3, and then identify defects based on the slurry image data 3, that is, when the gray value of the slurry area in the slurry image data 3 is greater than a preset threshold, the host computer directly identifies the defect; when the gray value of the slurry area in the slurry image data 3 is less than the preset threshold, the host computer controls the sound acquisition modules 21 to collect sound data, processes the sound data and embeds it into the slurry image data 3 to obtain composite image data, and determines the fault location of the nozzle 1 based on the composite image data, thereby realizing the rapid and accurate judgment of minor abnormalities of the nozzle 1 and determining the fault location of the nozzle 1.

[0047] In this embodiment, the acoustic spectrum corresponding to the sound data is combined with the slurry image data 3. The host computer analyzes the spectrum of the nozzle 1 based on the acoustic spectrum to analyze the working state of the nozzle 1, and then determines whether there is a minor abnormality in the nozzle 1. In addition, each shooting module 2 is equipped with a pair of sound acquisition modules 21, which can better determine the location of the abnormality in the nozzle 1. By combining the image and the sound, the abnormal location of the nozzle 1 can be quickly and accurately determined.

[0048] like Figure 2 As shown, in an optional embodiment, the imaging module 2 is disposed on both sides of the nozzle 1 used for spraying slurry onto the material in the width direction; the line connecting the two sound acquisition modules 21 corresponding to the imaging module 2 is parallel to the length of the nozzle 1; to obtain various sound data, so as to embed the sound data into the slurry image data 3 to form composite image data, and to determine the fault location of the nozzle 1 based on the composite image data.

[0049] In this embodiment, by setting up multiple shooting modules 2 and sound acquisition modules 21, it is convenient to determine the abnormal position of the nozzle 1.

[0050] like Figure 3As shown, in one optional implementation, the host computer is further configured to perform resolution calibration on the image data. That is, after the nozzle 1 starts spraying the slurry, the resolution of all the shooting modules 2 is acquired. Taking the image captured by the lowest resolution shooting module 2 as the reference frame, the size and resolution of the slurry image data 3 captured by each shooting module 2 are unified according to the reference frame, and a unique time axis is established to acquire several time packets. By constructing time packets, the slurry image data 3 is synchronized with the corresponding sound data. The time length corresponding to the time packet is a preset time length. The time packet includes: slurry image data 3 and sound data. The length of each segment of sound data is a preset time domain data interval.

[0051] In this embodiment, the time packet is a dataset formed by slurry image data 3 and sound data within a preset time length.

[0052] In this embodiment, the establishment of a unique time axis enables accurate synchronization and correspondence between time data and slurry image data 3, facilitating accurate judgment of minor slurry abnormalities caused by minor abnormalities in nozzle 1.

[0053] In this embodiment, the host computer can establish a unique time axis based on the internal crystal oscillator.

[0054] In this embodiment, the required shooting module 2 and the corresponding sound acquisition module 21 are selected according to the length requirements of the nozzle 1 in different spraying equipment. Each shooting module 2 acquires the corresponding slurry image data 3, and the corresponding pair of sound acquisition modules 21 acquires the corresponding sound data.

[0055] like Figure 3 As shown, in an optional implementation, the host computer is further configured to verify each time packet, that is, to determine whether the number of valid data in each time packet is consistent. When the number of complete slurry image data 3 and the number of complete sound data in each time packet are the same, the time packet verification is determined to be successful. Otherwise, the preset time length is increased, and then the time packet is re-acquired and verified again until the time packet verification is successful.

[0056] In this embodiment, the valid data can be complete slurry image data 3 and complete sound data.

[0057] In one optional implementation, the host computer is further configured to adjust the valid data within the time packet according to the response speed requirement after successful time packet verification; when a fast response is required, the valid data within the time packet is reduced; when a fast response is not required, the valid data within the time packet remains unchanged.

[0058] After adjusting the effective data, incomplete data within the time packet is removed, specifically the slurry image data 3 that differs from the reference frame size and the sound data that is smaller than the preset time domain data interval, forming the initial time packet, and the time period corresponding to the removed data is marked.

[0059] In this embodiment, after the spraying equipment has been debugged, the initial time package can be constructed.

[0060] In this embodiment, the time period corresponding to the data removal can be used for self-testing of the shooting module 2 and the sound acquisition module 21. The self-test is performed within the self-testing time, and then the data is acquired to ensure the accuracy of the acquisition.

[0061] In this embodiment, in high-speed response scenarios, a fast response is required, and the time packet is reduced to only one valid data point per valid data point; when the speed requirement is moderate, the number of valid data points can remain unchanged or be slightly reduced.

[0062] In this embodiment, the host computer can process the slurry image data 3 in grayscale and compare the grayscale value with a preset threshold to determine whether it is a minor anomaly. When the grayscale value is greater than the preset threshold, it is a macroscopic defect of the coating and can be directly judged as a defect. When the grayscale value is less than the preset threshold, it is judged as a suspected minor anomaly. At this time, corresponding sound data is added to the slurry image data 3 to make a quick and accurate judgment.

[0063] like Figure 4 As shown, in an optional embodiment, the host computer is further configured to, during the process of spraying slurry from nozzle 1, when it is determined that the gray value of the slurry area in the slurry image data 3 is less than a preset threshold, collect data according to the initial time packet, and if the actual time domain data spacing corresponding to the collected sound data at this time is the same as the preset time domain data spacing, then add barcode areas 31 on the left and right sides of each slurry image data 3 in the valid data, convert the sound data corresponding to the slurry image data 3 into a spectrogram and record it in the corresponding barcode area 31 to form composite image data.

[0064] In this embodiment, the host computer can determine the location of minor anomalies on the nozzle 1 based on the acoustic spectrogram of the sound data and the position of the corresponding sound acquisition module 21.

[0065] In this embodiment, the two barcode areas 31 added to the slurry image data 3 can correspond to the sound acquisition modules 21 on both sides of the shooting module 2, which makes it easier to obtain the location corresponding to the sound data.

[0066] In this embodiment, when acquiring sound data, it can be determined that the intensity of the sound data acquired by one shooting module 2 corresponds to two sound acquisition modules 21. When the sound data is matched with the slurry image data 3, the corresponding spectrogram can be recorded according to different requirements. For example, according to the requirements, only the sound data with higher sound intensity among the two sound data of the same slurry image data 3 can be recorded. At this time, when the sound intensity of the two sound data is the same, the corresponding spectrogram of both sound data is recorded.

[0067] In an optional implementation, the host computer is further configured to, during the process of spraying slurry from nozzle 1, when it is determined that the gray value of the slurry area in the slurry image data 3 is less than a preset threshold, collect data according to the initial time packet, and if the actual time domain data spacing corresponding to the collected sound data at this time is not the same as the preset time domain data spacing, then adjust the initial time packet to form a new time packet.

[0068] In this embodiment, when the actual time-domain data spacing is different from the preset time-domain data spacing, it is necessary to change the sampling rate. Lowering the sampling rate increases the time-domain data spacing, which is suitable for low-frequency sounds such as human voices and machine low frequencies. Increasing the sampling rate decreases the time-domain data spacing, restores more high-frequency details, and enables faster high-frequency signal analysis.

[0069] In one optional implementation, the method of adjusting the initial time packet to form a new time packet includes: the host computer re-inserting the removed time period into the initial time packet to form a complete time length, and re-dividing the complete time length of the initial time packet according to the actual time domain data interval. If the number of sound data corresponding to the complete actual time domain data interval in each initial time packet is the same after division, the incomplete time data is removed to form a new time packet. If the number of sound data corresponding to the complete actual time domain data interval in each initial time packet is not the same after division, the complete time length of the initial time packet is increased until the number of sound data corresponding to the complete actual time domain data interval in each initial time packet is the same, and the incomplete time data is removed to form a new time packet.

[0070] In an optional implementation, the host computer is further configured to collect data according to the new time packet, add barcode areas 31 on the left and right sides of each slurry image data 3 in the valid data, convert the sound data corresponding to the slurry image data 3 into a spectrogram and record it in the corresponding barcode area 31 to form composite image data, determine whether the sound data collected by the corresponding sound acquisition module 21 is abnormal according to the spectrogram in the composite image data, and determine whether the nozzle 1 at the corresponding sound acquisition module 21 is faulty when it is abnormal, and then determine the fault location of the nozzle 1.

[0071] At least one other disclosed embodiment also provides a detection method using the above-described multidimensional vision inspection system, comprising: controlling the imaging module 2 via a host computer to capture images of the slurry on the material to obtain slurry image data 3; then identifying defects based on the slurry image data 3, i.e., when the gray value of the slurry area in the slurry image data 3 is greater than a preset threshold, the host computer directly identifies the defect; when the gray value of the slurry area in the slurry image data 3 is less than the preset threshold, the host computer controls the sound acquisition module 21 to acquire sound data; processing the sound data and embedding it into the slurry image data 3 to obtain composite image data; and determining the fault location of the nozzle 1 based on the composite image data.

[0072] In summary, this multi-dimensional vision inspection system includes: a host computer and several imaging modules 2 electrically connected to the host computer; both sides of the imaging modules 2 are equipped with sound acquisition modules 21 electrically connected to the host computer; the host computer is configured to control the imaging modules 2 to capture images of the slurry on the material to obtain slurry image data 3, and then identify defects based on the slurry image data 3. That is, when the gray value of the slurry area in the slurry image data 3 is greater than a preset threshold, the host computer directly identifies the defect; when the gray value of the slurry area in the slurry image data 3 is less than the preset threshold, the host computer controls the sound acquisition modules 21 to collect sound data, processes the sound data and embeds it into the slurry image data 3 to obtain composite image data, and determines the fault location of the nozzle 1 based on the composite image data, thereby realizing the rapid and accurate judgment of minor abnormalities of the nozzle 1 and determining the fault location of the nozzle 1.

[0073] Based on the above-described preferred embodiments of the present invention, and through the foregoing description, those skilled in the art can make various changes and modifications without departing from the inventive concept. The technical scope of this invention is not limited to the contents of the specification, but must be determined according to the scope of the claims.

Claims

1. A multi-dimensional vision inspection system, characterized by, The application relates to a multi-dimensional visual detection system. The application comprises: a host computer and a plurality of shooting modules (2) electrically connected to the host computer; sound collection modules (21) are arranged on the two sides of the shooting module (2) and are electrically connected to the host computer; the host computer is configured to control the shooting module (2) to shoot slurry on a material to obtain slurry image data (3) and then identify defects according to the slurry image data (3); that is, when the gray value of the slurry region in the slurry image data (3) is greater than a preset threshold value, the host computer directly identifies defects; when the gray value of the slurry region in the slurry image data (3) is less than the preset threshold value, the host computer controls the sound collection module (21) to collect sound data, embeds the sound data in the slurry image data (3) after processing to form composite image data, and judges the fault position of the nozzle (1) according to the composite image data; the shooting module (2) is arranged on the two sides of the width direction of a nozzle (1) used for spraying slurry on a material; the two sound collection modules (21) corresponding to the shooting module (2) are connected in parallel to the length of the nozzle (1); so as to obtain sound data, embed the sound data in the slurry image data (3) to form composite image data, and judge the fault position of the nozzle (1) according to the composite image data; the host computer is further configured to calibrate the resolution of the image data; that is, the resolution of all the shooting modules (2) is obtained after the nozzle (1) starts to spray slurry, wherein the image shot by the shooting module (2) with the minimum resolution is taken as a reference frame, the slurry image data (3) shot by each shooting module (2) is unified in size and resolution according to the reference frame, a unique time axis is established, and a plurality of time packets are obtained; the slurry image data (3) is synchronized with the corresponding sound data through the construction of the time packet; the time length corresponding to the time packet is a preset time length; the time packet comprises the slurry image data (3) and the sound data; wherein the length of each piece of sound data is a preset time domain data interval.

2. The multi-dimensional visual detection system according to claim 1, wherein the host computer is further configured to check the time packet; that is, to judge whether the number of effective data in each time packet is consistent; when the number of complete slurry image data (3) in each time packet is the same and the number of complete sound data is also the same, it is judged that the time packet checking is successful, otherwise, the time length is increased by a preset time length, then the time packet is reacquired and checking is performed again until the time packet checking is successful.

3. The multi-dimensional visual detection system according to claim 2, wherein the host computer is further configured to adjust the effective data in the time packet according to the response speed requirement after the time packet checking is successful; when fast response is required, the effective data in the time packet is reduced; when fast response is not required, the effective data in the time packet remains unchanged; and after the effective data is adjusted, the incomplete data in the time packet is removed; that is, the slurry image data (3) with a size different from that of the reference frame and the sound data smaller than the preset time domain data interval are removed to form an initial time packet, and the time period corresponding to the removed data is marked.

4. The multi-dimensional visual inspection system of claim 3, wherein The host computer is further configured to, during the process of spraying slurry by the nozzle (1), when judging that the gray value of the slurry region in the slurry image data (3) is less than the preset threshold, collect data according to the initial time package, if the actual time domain data interval corresponding to the sound data collected at this time is the same as the preset time domain data interval, add a barcode region (31) on the left and right sides of each slurry image data (3) in the effective data, record the sound data corresponding to the slurry image data (3) in the corresponding barcode region (31) after converting the sound data into a spectrogram, and form composite image data.

5. The multi-dimensional visual inspection system of claim 3, wherein The host computer is further configured to, during the process of spraying slurry by the nozzle (1), when judging that the gray value of the slurry region in the slurry image data (3) is less than the preset threshold, collect data according to the initial time package, if the actual time domain data interval corresponding to the sound data collected at this time is not the same as the preset time domain data interval, adjust the initial time package to form a new time package.

6. The multi-dimensional visual inspection system of claim 5, wherein The method of adjusting the initial time package to form a new time package comprises: The host computer reinserts the removed time period into the initial time package to form a complete time length, and redivides the complete time length of the initial time package according to the actual time domain data interval, if the number of sound data corresponding to the complete actual time domain data interval in each initial time package after division is the same, removes the incomplete time data to form a new time package, if the number of sound data corresponding to the complete actual time domain data interval in each initial time package after division is not the same, increases the complete time length of the initial time package until the number of sound data corresponding to the complete actual time domain data interval in each initial time package is the same, and removes the incomplete time data to form a new time package.

7. The multi-dimensional visual inspection system of claim 6, wherein The host computer is further configured to, according to the new time package, collect data, add a barcode region (31) on the left and right sides of each slurry image data (3) in the effective data, record the sound data corresponding to the slurry image data (3) in the corresponding barcode region (31) after converting the sound data into a spectrogram, and form composite image data, judge whether the sound data collected by the corresponding sound collecting module (21) is abnormal according to the spectrogram in the composite image data, and when abnormal, judge that the nozzle (1) at the corresponding sound collecting module (21) has a fault, and further judge the fault position of the nozzle (1).

8. A method of inspection using the multi-dimensional vision inspection system of claim 1, wherein, including: The host computer controls the shooting module (2) to shoot the slurry on the material to obtain slurry image data (3), and then identifies defects according to the slurry image data (3), that is, when the gray value of the slurry area in the slurry image data (3) is greater than a preset threshold, the host computer directly identifies defects, and when the gray value of the slurry area in the slurry image data (3) is less than the preset threshold, the host computer controls the sound collection module (21) to collect sound data, embeds the sound data in the slurry image data (3) after processing to obtain composite image data, and judges the fault position of the nozzle (1) according to the composite image data.

Images