A high-throughput smoke density test experimental system and method based on image recognition

Abstract

The present invention discloses a high-throughput smoke density test experimental system and method based on image recognition. The system includes a front-end high-throughput smoke density test device, a middle-end processing unit, and a back-end analysis unit that are interconnected and communicate with each other. Among them, the front-end high-throughput smoke density test device includes a background lighting device, a sample box, a photographing device, and a cone heater. The middle-end processing unit includes an image recognition module, a lighting data acquisition module, and a heating data acquisition module. The back-end analysis unit includes a computer and its built-in analysis program. By visualizing the characterization of smoke density testing, the present invention realizes the continuous characterization of smoke in space and time during the entire testing process, and then the computer performs secondary processing and automatic grading evaluation on the data collected, processed, and classified by the middle-end processing unit to generate a test report with detailed data and grades, which can improve the detection efficiency and accuracy of the sample smoke density.

Description

Technical Field

[0001] The present invention belongs to the technical field of material property testing, and particularly relates to a high-throughput smoke density testing experimental system and method based on image recognition. Background Art

[0002] Polymer materials are extremely flammable, and the toxic smoke generated during the combustion process can cause serious damage to human health and greatly reduce the escape time of people during a fire. Therefore, it is crucial to effectively reduce the smoke density released by polymer materials during combustion.

[0003] The existing methods for characterizing the smoke density of polymers mainly include the "single-chamber method for measuring the smoke density experimental method" (GB / T8323.2-2008 / ISO 5659-2:2006). In this method, the test specimen is placed horizontally in the test chamber, and the upper surface of the specimen is exposed to a heat radiation source with a constant irradiance set at a certain value for testing. The generated smoke is collected in the test chamber equipped with a photometer, and the attenuation of the measurement beam passing through the smoke is measured. The result is expressed in terms of specific optical density. This traditional test method is time-consuming, inefficient, and highly complex, and cannot meet the screening of large-scale low-smoke-density materials. Providing a high-throughput smoke density detection device and method is the key to solving such problems.

[0004] The high-throughput concept and method are widely used in fields such as medicine and metals, but relatively less in the polymer field. High-throughput will become a research hotspot in the future of polymer research. In order to effectively improve the screening of polymer composites with low smoke density, there is an urgent need for a high-throughput experimental device and method that is easy to operate, low-cost, and highly efficient to provide technical support for ordinary experimental detection equipment. Summary of the Invention

[0005] To overcome the above-mentioned deficiencies of the existing technology, the purpose of the present invention is to provide a high-throughput smoke density testing experimental system and method based on image recognition. By using visualization and digital means, the quantitative test of smoke density is converted from the initial specific optical density to a picture recognition method, and multiple samples are tested and analyzed simultaneously to achieve high-throughput, rapid, and efficient screening of polymer composites.

[0006] To achieve the above purpose, the present invention provides the following technical solutions:

[0007] A high-throughput smoke density testing experimental system based on image recognition, characterized in that it includes a front-end high-throughput smoke density testing device, a middle-end processing unit, and a back-end analysis unit that are interconnected and communicate with each other;

[0008] Among them, the front-end high-throughput smoke density testing device includes a background lighting device, multiple sample boxes with visual windows, a photographing device, and multiple cone heaters, which are used to visualize the characterization of smoke density testing and realize the continuous characterization of smoke in space and time during the whole testing process;

[0009] The middle-end processing unit includes an image recognition module, an illumination data acquisition module, and a heating data acquisition module, which are used to initially collect, classify, and record data;

[0010] The back-end analysis unit includes a computer and its built-in analysis program. The computer performs secondary processing and automatic grading evaluation on the data collected, processed, and classified by the middle-end processing unit, and generates a test report with detailed data and grades;

[0011] The background lighting device is provided with a cylindrical background plate with openings on both the upper and lower end faces. Inside the background plate in the thickness direction, there are multiple uniformly emitting LED light sources, and the output of the light sources is stable and uniform;

[0012] On the upper end face of each cone heater, there is a conical heating port;

[0013] Each sample box is provided with a box-shaped housing. Inside the box-shaped housing, there is a two-sided or four-sided removable glass slide. The hollow part surrounded by each glass slide contains the sample to be detected; on the box-shaped housing, there is at least one removable visual window and an observation window. Through the glass slide corresponding to the visual window and the observation window, the sample to be detected inside and the smoke emitted by it can be observed; the bottom of the box-shaped housing is a conical plug, which is inserted into the heating port of the cone heater (5) and fixed by the heating port;

[0014] The multiple cone heaters together with the sample boxes fixed by them are arranged in sequence along a set path inside the cylindrical background plate;

[0015] The photographing device is set at the center of the background plate of the background lighting device. It is a camera that takes pictures at a 360° angle of view. During the whole process of smoke density testing of the samples in all sample boxes, the images presented are photographed, recorded, and stored by the camera using the monitoring and photographing function;

[0016] The photographing device is connected to the image recognition module, and transmits the images of the whole process of smoke density testing taken to the image recognition module. The image recognition module sorts and stores the obtained information to obtain a detailed record in terms of space and time, and finally transmits the data to the back-end analysis unit for secondary processing. The back-end analysis unit converts the quantitative test of smoke density from the initial specific optical density to the way of picture recognition, and through its processing and automatic grading evaluation, automatically generates a test report with detailed data and grades.

[0017] The multiple sample boxes with visual windows are arranged in a set circular, square or any other arbitrary path with respect to each other, and the sample components in each sample box are the same or different; the number of visual windows and observation windows on each sample box can also be one or more to obtain multiple images of different parts for analysis.

[0018] On the visual window and the observation window of each sample box, there is also a baffle for covering to block some of the windows to reduce the interference of light on the camera taking pictures.

[0019] The front-end high-throughput smoke density testing device further includes light-shielding plates; there are multiple light-shielding plates, which are respectively arranged on the inner side of the background plate of the background lighting device and at the gap positions between the sample boxes after they are arranged in sequence to avoid the interference of the light on the side of the sample boxes on the generated images and improve the accuracy of image recognition.

[0020] The conical heater is arranged directly below each sample box, and a heating component is provided inside it to provide the temperature required for sample testing; the conical heater is connected to the heating data acquisition module, and the heating data acquisition module (8) collects and records the temperature change and transmits it to the backend analysis unit for analysis and processing.

[0021] Inside each conical heater, there is also a microelectronic scale, which is arranged directly below the connected sample box to measure the change data of the sample weight during the test as the thermogravimetric information of the sample, and preliminarily store the obtained thermogravimetric information through the heating data acquisition module (8), and finally transmit it to the backend analysis unit for secondary classification, storage and analysis.

[0022] A high-throughput smoke density testing method based on image recognition using the foregoing testing system, characterized in that it includes the following steps:

[0023] S1: Prepare a batch of samples, record the names and weights of each sample, place each sample into a corresponding sample box, then insert each sample box into the heating port of a conical heater, and then arrange the conical heaters together with the sample boxes in sequence on the inner side of the cylindrical background plate according to the set path;

[0024] S2: Power on the front-end high-throughput smoke density testing device, the middle-end processing unit, and the backend analysis unit, set the parameters in the backend system, and then turn on the photographing device and the background lighting device;

[0025] S3: Turn on each conical heater to heat the samples in each corresponding sample box respectively;

[0026] S4: The photographing device and the background lighting device work synchronously. The photographing device takes pictures. The middle-end processing unit, including an image recognition module, a lighting data acquisition module, and a heating data acquisition module, works synchronously to initially collect, classify, and record data, record the data of the phenomenon changes in space and time during the entire test process, and transmit this data to the middle-end processing unit.

[0027] S5: The middle-end processing unit initially stores and classifies the collected and received test data, and then packs and transmits the test data to the back-end analysis unit.

[0028] S6: After receiving the test data, the back-end analysis unit stores and analyzes it according to its built-in analysis program, and analyzes the process data of the test until the test ends, and summarizes and outputs the final grade result obtained after its analysis.

[0029] The built-in analysis program and parameters of the back-end analysis unit are based on the change range of the RGB values of multiple consecutive test pictures to analyze the image recognition rules for dynamically dividing the smoke density levels corresponding to multiple test pictures. The analysis program and parameters are as follows: when the RGB value is 215 - 255, output level L0; when the RGB value is 170 - 215, output level L1; when the RGB value is 125 - 170, output level L2; when the RGB value is 80 - 125, output level L3; when the RGB value is 0 - 80, output level L4.

[0030] The beneficial effects of the present invention are as follows:

[0031] (1) The high-throughput smoke density test experimental system and method based on image recognition provided by the present invention adopt visualization and digital means to convert the quantitative test of smoke density from the initial specific optical density to the way of picture recognition, synchronously conduct multiple sample tests and analyses, and realize the high-throughput, rapid, and efficient screening of polymer composites.

[0032] (2) The high-throughput smoke density test experimental system and method based on image recognition provided by the present invention convert the quantitative test of smoke density from the initial specific optical density to the way of picture recognition, synchronously conduct multiple sample (component) tests, realize the high-throughput, rapid, and efficient screening of polymer composites, and can realize the detailed digital recording of the entire smoke density test process in terms of space and time, with simple operation, low cost, high efficiency, and high detection accuracy.

[0033] (3) The high-throughput smoke density test experimental system and method based on image recognition provided by the present invention redefine the high-throughput smoke density level through visualization and digital means, provide a new method for the characterization of smoke density, which is more intuitive and easy to observe, making the visualization of the smoke density characterization possible.

[0034] (4) The high-throughput smoke density test experimental system and method based on image recognition provided by the present invention have high flexibility in use. Each part can be detachably assembled and combined to adapt to different numbers and batches of experimental samples. At the same time, it can also be multifunctional, providing a basis for high-throughput thermogravimetric testing. During the same test experiment process, not only digital information of high-throughput smoke density can be obtained, but also digital information of high-throughput thermogravimetry can be obtained. Description of the Drawings

[0035] Figure 1 It is a schematic diagram of the overall composition structure of the high-throughput smoke density test experimental system according to an embodiment of the present invention;

[0036] Figure 2 It is a top view of the device of the high-throughput smoke density test experimental system according to an embodiment of the present invention;

[0037] Figure 3 It is a schematic diagram of the external structure of the sample box of the high-throughput smoke density test experimental system according to an embodiment of the present invention;

[0038] Figure 4 It is a schematic diagram of the step flow of the high-throughput smoke density test experimental method according to an embodiment of the present invention;

[0039] Figure 5 It is a schematic diagram of representative photos of the high-throughput smoke density grades of different polypropylene composites obtained in Embodiment 2 of the present invention.

[0040] In the figure: 1. Background lighting device; 2. Sample box; 3. Photographing device; 4. Light-shielding plate; 5. Cone heater; 6. Image recognition module; 7. Lighting data acquisition module; 8. Heating data acquisition module; 9. Back-end analysis unit; 10. Conical plug; 11. Visual window; 12. Observation window; 13. Background board. Detailed Embodiments

[0041] The technical solutions of the present invention are further introduced in detail below in conjunction with embodiments, but the protection scope of the present invention is not limited thereto.

[0042] Embodiment 1:

[0043] Please refer to Figures 1-4 , the high-throughput smoke density test experimental system based on image recognition provided by an embodiment of the present invention includes a front-end high-throughput smoke density test device 1, a middle-end processing unit, and a back-end analysis unit 9 that are connected and communicate with each other;

[0044] Among them, the front-end high-throughput smoke density testing device includes a background lighting device 1, multiple sample boxes 2 with a visual window 11 and an observation window 12, a photographing device 3, and multiple cone heaters 5. Each part cooperates with each other to make the characterization of smoke density testing visual, and realizes the continuous characterization of smoke in space and time during the whole testing process;

[0045] The middle-end processing unit includes an image recognition module 6, an illumination data acquisition module 7, and a heating data acquisition module 8, which are used for preliminary data collection, classification, and recording;

[0046] The back-end analysis unit 9 includes a computer and its built-in analysis program. The computer performs secondary processing and automatic grading evaluation on the data collected, processed, and classified by the middle-end processing unit, and generates a test report with detailed data and grades;

[0047] The background lighting device 1 is provided with a cylindrical background plate 13 with openings on both the upper and lower end faces. Inside the thickness direction of the background plate 13, there are multiple uniformly emitting LED light sources, and the output of the light sources is stable and uniform;

[0048] On the upper end face of each cone heater 5, there is a conical heating port;

[0049] Each sample box 2 is provided with a box-shaped housing. Inside the box-shaped housing, there is a two-sided or four-sided removable glass slide. The hollow part surrounded by each glass slide contains the sample to be detected; on the box-shaped housing, there is at least one removable visual window 11 and observation window 12. Through the glass slides corresponding to the visual window 11 and the observation window 12, the sample to be detected inside and the smoke emitted after heating can be observed; the bottom of the box-shaped housing is a conical plug 10, which is inserted into the heating port of the cone heater 5 and fixed by the heating port;

[0050] The multiple cone heaters 5 together with the sample boxes 2 fixed by them are arranged in sequence along a set path inside the cylindrical background plate 13. The specific spacing distance is determined according to actual needs. They can be hung on the vertical surface of the inner side wall of the background plate 13 or placed on the ground;

[0051] The photographing device 3 is a panoramic camera, which is set at the center of the background plate 13 of the background lighting device 1. Specifically, it is a camera that takes pictures from a 360° perspective. During the whole process of smoke density testing of the samples in all sample boxes 2, the images presented are taken panoramically and continuously by using continuous photographing functions such as monitoring and photographing by this camera 3, and recorded and stored;

[0052] The described photographing device 3 is connected to the image recognition module 6 either wired or wirelessly, and transmits the images of the entire process of smoke density testing it captures to the image recognition module 6. The image recognition module 6 sorts, classifies, and stores the obtained information to obtain a detailed record in terms of space and time, and finally transmits the data to the backend analysis unit 9 for secondary processing. The backend analysis unit 9 converts the quantitative test of smoke density from the initial specific optical density into a method of image recognition. Through its processing and automatic grading and evaluation, it automatically generates a test report with detailed data and grades.

[0053] The described multiple sample boxes 2 with viewing windows 11 and observation windows 12 are arranged in a set circular, square, or any other arbitrary path with respect to each other. The sample components in each sample box 2 can be the same or different; the number of viewing windows 11 and observation windows 12 on each sample box can also be one or more to obtain multiple images of different parts for analysis.

[0054] On the viewing window 11 and observation window 12 of each sample box 2, there is also a baffle plate for covering (not shown in the figure), whose size and shape are basically the same as those of the viewing window 11 and observation window 13, and is used to block some of the windows to reduce the interference of external light or reflection on the photographing of the camera 3.

[0055] As an optimized solution, the described front-end high-throughput smoke density testing device further includes light-shielding plates 4; there are multiple light-shielding plates 4, whose shape and size are the same as those of the sample box 2, but they do not have transparent viewing windows; each light-shielding plate 4 is respectively arranged on the inner side of the background plate of the background lighting device (1), at the gap positions between the sample boxes 2 after they are arranged in sequence, and is also arranged on the same arrangement path as the sample boxes (filling in the vacant positions), to avoid the interference of the side light of the sample box on the image generated by the camera 3 when the number of sample boxes is insufficient to form a complete circle, and improve the accuracy of image recognition.

[0056] The described conical heater 5 is arranged directly below each sample box, and its interior is provided with a heating component to provide the temperature required for sample testing; the conical heater 5 is connected to the heating data acquisition module 8, and the heating data acquisition module 8 collects and records the temperature changes and transmits them to the described backend analysis unit 9 for analysis and processing.

[0057] Each conical heater 5 also has a microelectronic scale inside, which is arranged directly below the connected sample box to measure the change data of the sample weight during the test as the thermogravimetric information of the sample, and preliminarily stores the obtained thermogravimetric information through the heating data acquisition module 8, and finally transmits it to the backend analysis unit 9 for secondary classification, storage, and analysis.

[0058] A high-throughput smoke density testing method based on image recognition using the aforementioned testing system includes the following steps:

[0059] S1: Prepare a batch of samples, record the name and weight of each sample, place each sample into a corresponding sample box, then insert each sample box into the heating port of a conical heater 5, and then arrange the conical heaters 5 together with the sample boxes in sequence along the set path on the inner side of the cylindrical background board;

[0060] The amount of each sample to be tested is at the unit level of g or mg; when the system conducts experimental tests, the amount of samples used is small, and the unit level can be g or mg, which can save the preparation time and cost of composite materials, but a large amount of test data can be obtained;

[0061] Adjust the number of the multiple viewing windows 11 and observation windows 12 opened on two sides or four sides of the sample box 2, and only one can be retained, and the others are blocked with baffles; used to block some of the windows 11 or 12 to reduce the interference of light on the camera taking pictures;

[0062] S2: Power on and operate the front-end high-throughput smoke density test device, the middle-end processing unit, and the back-end analysis unit, set the parameters in the back-end system, and then turn on the photographing device, the panoramic camera 3, and the background lighting device 1;

[0063] S3: Turn on each conical heater and heat the samples in each sample box correspondingly;

[0064] S4: The photographing device 3 and the background lighting device 1 work synchronously. The photographing device 3 takes pictures, and the middle-end processing unit including the image recognition module 6, the lighting data acquisition module 7, and the heating data acquisition module 8 work synchronously to initially collect, classify, and record data, record the phenomenon change data in space and time during the whole test process, and transmit this data to the middle-end processing unit; the picture information obtained by the panoramic camera 3 includes RGB value data;

[0065] S5: The middle-end processing unit stores and classifies the collected and received test data initially, and then packs and transmits the test data to the back-end analysis unit;

[0066] S6: After receiving the test data, the back-end analysis unit stores and analyzes it according to its built-in analysis program, and analyzes the process data of the test until the test ends, and summarizes and outputs the final grade result obtained after its analysis.

[0067] The middle-end processing unit in step S5 and the back-end analysis unit in step S6 both perform classification and grading processing based on the RGB value information data of the picture.

[0068] The described back-end analysis unit 9 is built with an image recognition rule analysis program and parameters for dynamically dividing the smoke density levels corresponding to multiple test images based on the range of RGB values of multiple consecutive test images. The rule analysis program and parameters are as follows: when the RGB value is 215 - 255, output level L0; when the RGB value is 170 - 215, output level L1; when the RGB value is 125 - 170, output level L2; when the RGB value is 80 - 125, output level L3; when the RGB value is 0 - 80, output level L4.

[0069] In this embodiment, the described background lighting device 1 is placed surrounding the arrangement path of the sample boxes. An illumination background plate 13 is added on the back of the visual window and the observation window, and an LED light source device is built in the background plate to ensure stable and uniform light source output.

[0070] The sample box 2 with a visual window and an observation window is one of the key components of the high-throughput smoke density testing device. The number of sample boxes can be adjusted according to needs and can be arranged in a circular, square or any other arbitrary path for the purpose of facilitating photography and recording to achieve high-throughput testing. The number of visual windows and observation windows on the sample box is also correspondingly multiple and is evenly spaced vertically.

[0071] The panoramic camera 3 of the described photographing device is set at the center of the circle of the background plate 13 and has a 360° field of view. Its shape is not limited to the drawn shape and can be shown according to the actual situation. The purpose is to convert the quantification of smoke density from the initial specific optical density to the way of image recognition, and the phenomena presented by all sample boxes can be recorded and stored through functions such as monitoring and photographing by the camera.

[0072] The described light-shielding plate 4 is to avoid the adverse effects of the side light on the generated image when the number of sample boxes is insufficient or there are vacancies, so as to improve the accuracy of image recognition; when the number of sample boxes 2 is large enough and the sample boxes 2 can be directly connected to each other without gaps (vacancies) between two sample boxes 2, the light-shielding plate 4 can be not used.

[0073] The color of the light source output by the background lighting device 1 is not restricted, as long as it can maintain stable and uniform light source output, so that each sample box 2 is under the same illuminance, improving the comparability of the photographed pictures and thus the accuracy of the detection results.

[0074] See Figure 3 , the sample box 2 with a visual window and an observation window is fixed on the cone heater through its conical plug 10. The two sides or four sides of the glass slide are provided with detachable visual windows and observation windows, and the glass slide can be installed and cleaned according to the usage requirements. Each window can be covered with an optional baffle to reduce the influence of light on camera photography.

[0075] The opening sizes of the visual window and the observation window on the sample box are adjustable, and their shapes are also adjustable. When it is square, the opening size can also be adjusted by equipping with baffles of different sizes.

[0076] Embodiment 2:

[0077] Please refer to Figure 5 , the high-throughput smoke density test experimental system and method based on image recognition provided by the present invention are specifically applied to Embodiment 1. Weigh 0.5 g of polypropylene composite 1, polypropylene composite 2, polypropylene composite 3, polypropylene composite 4, and polypropylene composite 5 respectively, and test them through the high-throughput smoke density test experimental system and method of Embodiment 1. The test time is 15 minutes, the photographing interval of the panoramic camera 3 is set to 2 s, and only the smoke generation phenomenon presented in one visual window 11 is photographed to collect photos. Select the photo with the largest change in RGB value in the photo set to represent the grade. Finally, the backend analysis unit outputs the high-throughput smoke density grade results of the above-mentioned 5 polypropylene composites, and the representative photos of different grades are as Figure 5 shown.

[0078] The focus of the present invention is to make the characterization of smoke density testing visual and digital, realize the continuous characterization of smoke in space and time during the whole testing process, and then perform secondary processing and automatic grading evaluation on the data collected, processed and classified by the middle-end processing unit by a computer to generate a test report with detailed data and grades, which can improve the detection efficiency and accuracy of the smoke density of samples.

[0079] The above are only exemplary embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A high-throughput smoke density test experimental system based on image recognition, characterized in that: it includes a front-end high-throughput smoke density test device, a middle-end processing unit, and a back-end analysis unit (9) that are interconnected and communicate with each other; Among them, the front-end high-throughput smoke density test device includes a background lighting device (1), multiple sample boxes (2) with viewing windows (11), a photographing device (3), and multiple conical heaters (5), which are used to visualize the characterization of the smoke density test and realize the continuous characterization of the smoke in space and time during the entire test process; The middle-end processing unit includes an image recognition module (6), an illumination data acquisition module (7), and a heating data acquisition module (8), which are used to initially collect, classify, and record data; The back-end analysis unit (9) includes a computer and its built-in analysis program, and the computer performs secondary processing and automatic grading evaluation on the data collected, processed, and classified by the middle-end processing unit, and generates a test report with detailed data and grades; The background lighting device (1) is provided with a cylindrical background plate with openings on both the upper and lower end faces, and a plurality of uniformly emitting LED light sources are arranged inside the background plate in the thickness direction, and the output of the light source is stable and uniform; On the upper end face of each conical heater (5), a conical heating port is provided; Each sample box (2) is provided with a box-shaped housing, and a two-sided or four-sided, detachable glass slide is sleeved inside the box-shaped housing, and the hollow part surrounded by each glass slide is used to place the sample to be detected; on the box-shaped housing, at least one detachable viewing window is provided, and through the glass slide corresponding to the viewing window, the sample to be detected inside and the smoke emitted by it can be observed; the bottom of the box-shaped housing is a conical plug (10), which is inserted into the heating port of the conical heater (5) and fixed by the heating port; The multiple conical heaters (5) and the sample boxes (2) fixed by them are arranged in sequence along a set path inside the cylindrical background plate; The photographing device (3) is arranged at the center of the background plate of the background lighting device (1), and it is a camera that takes pictures at a 360° angle of view. During the entire process of the smoke density test of the samples in all sample boxes, the images presented are photographed, recorded, and stored by using the monitoring and photographing function of the camera; The photographing device (3) is connected to the image recognition module (6), and transmits the images of the entire process of the smoke density test taken to the image recognition module (6). The image recognition module (6) sorts and stores the obtained information to obtain a detailed record in terms of space and time, and finally transmits the data to the back-end analysis unit (9) for secondary processing. The back-end analysis unit (9) converts the quantitative test of the smoke density from the initial specific optical density to a picture recognition method, and through its processing and automatic grading evaluation, automatically generates a test report with detailed data and grades.

2. The high-throughput smoke density test experimental system based on image recognition according to claim 1, characterized in that: The described sample boxes (2) are arranged in a set circular, square or any other arbitrary path, and the sample components in each sample box (2) are the same or different; the number of viewing windows on each sample box can also be one or more to obtain multiple images of different parts for analysis.

3. The high-throughput smoke density test experimental system based on image recognition according to claim 1, characterized in that: Each of the sample boxes is further provided with an observation window, and a baffle for covering is also provided between the viewing window and the observation window to block some of the windows to reduce the interference of light on camera photography.

4. The high-throughput smoke density test experimental system based on image recognition according to claim 1, characterized in that: The front-end high-throughput smoke density test device further includes light-shielding plates (4); there are a plurality of the light-shielding plates (4), which are respectively arranged on the inner side of the background plate of the background lighting device (1) at the gap positions between the sample boxes (2) arranged in sequence to avoid the interference of the light on the side of the sample box on the generated image and improve the accuracy of image recognition.

5. The high-throughput smoke density test experimental system based on image recognition according to claim 1, characterized in that: The cone heater (5) is arranged directly below each sample box, and a heating component is provided inside to provide the temperature required for sample testing; the cone heater (5) is connected to the heating data acquisition module (8), and the heating data acquisition module (8) collects and records the temperature change and transmits it to the backend analysis unit (9) for analysis and processing.

6. The high-throughput smoke density test experimental system based on image recognition according to claim 5, characterized in that: Each cone heater (5) is further provided with a microelectronic scale inside, which is arranged directly below the connected sample box to measure the change data of the sample weight during the test as the thermogravimetric information of the sample, and the obtained thermogravimetric information is preliminarily stored through the heating data acquisition module (8) and finally transmitted to the backend analysis unit (9) for secondary classification, storage and analysis.

7. A high-throughput smoke density test method based on image recognition using the test system according to any one of claims 1-6, characterized in that, it includes the following steps: S1: Prepare a batch of samples, record the names and weights of each sample, place each sample into a corresponding sample box, then insert each sample box into the heating port of a cone heater (5), and then arrange the cone heaters (5) together with the sample boxes in sequence on the inner side of the cylindrical background plate according to the set path; S2: Power on the front-end high-throughput smoke density test device, the middle-end processing unit, and the backend analysis unit, set the parameters in the backend system, and then turn on the photographing device (3) and the background lighting device (1); S3: Turn on each cone heater and heat the samples in each sample box correspondingly. S4: The photographing device (3) and the background lighting device (1) work synchronously. The photographing device (3) takes pictures. The middle-end processing unit includes an image recognition module (6), an illumination data acquisition module (7), and a heating data acquisition module (8) that work synchronously to initially collect, classify, and record data, record the change data of phenomena in space and time during the entire test process, and transmit this data to the middle-end processing unit; S5: The middle-end processing unit initially stores and classifies the collected and received test data, and then packages and transmits the test data to the back-end analysis unit; S6: After receiving the test data, the back-end analysis unit stores and analyzes it according to its built-in analysis program, and analyzes the process data of the test until the test ends, and summarizes and outputs the final grade result obtained after its analysis.

8. The high-throughput smoke density test method based on image recognition according to claim 7, wherein, the step S1 further includes the following steps: The amount of each sample to be tested is at the unit level of g or mg; Adjust the number of multiple visible windows opened on two or four sides of the sample box (2), and use a baffle to block some of the windows to reduce the interference of light on camera photographing.

9. The high-throughput smoke density test method based on image recognition according to claim 7, wherein, the picture information obtained by photographing with the photographing device (3) in the step S4 includes RGB value data; The middle-end processing unit in the step S5 and the back-end analysis unit in the step S6 both perform classification and grading processing based on the RGB value information data of the picture.

10. The high-throughput smoke density test method based on image recognition according to claim 9, wherein, The back-end analysis unit (9) has an image recognition rule analysis program and parameters for dynamically dividing the smoke density levels corresponding to multiple test pictures based on the change range of RGB values of multiple consecutive test pictures. The rule analysis program and parameters are: When the RGB value is 215 - 255, output level L0; when the RGB value is 170 - 215, output level L1; when the RGB value is 125 - 170, output level L2; when the RGB value is 80 - 125, output level L3; when the RGB value is 0 - 80, output level L4.

Images