A coffee roasting state on-line real-time detection device and method

Abstract

This invention belongs to the field of visual inspection technology for coffee beans, specifically relating to an online real-time detection device and method for coffee roasting status. The detection device includes a controller electrically connected to a light source module, a sensor assembly, and an AD converter. The light source module emits different wavelengths of light. An optical component for filtering light is externally mounted on the light source module, and an optical component for focusing light is externally mounted on the sensor assembly. The light source module includes a laser light source and a monochromatic LED light source connected to the controller. The laser light source emits infrared laser light. The optical component includes a PID controller connected to the controller, and an objective lens is externally mounted on the light source module. The method includes image acquisition using the sensor assembly. This invention quantifies the amount of light reflected by coffee beans during roasting, deriving relevant quantitative indicators for determining the degree of coffee roasting, thus facilitating the detection of volume changes during coffee bean roasting to approximate true values.

Description

Technical Field

[0001] This invention belongs to the field of visual inspection technology for coffee beans, specifically relating to an online real-time detection device and method for coffee roasting status, which is applied to industrial coffee bean processing lines for online optical detection of coffee bean roasting status. Background Technology

[0002] Roasting is a crucial step in coffee production because it directly determines the quality of the coffee. Traditional methods of observing color with the naked eye can be subjective and inconsistent, as the human visual system is easily affected by the surrounding environment and light. Professional roasters usually use an Agtron value analyzer to measure the degree of roasting, but this measurement can only be performed after the coffee beans have cooled, making it impossible to adjust the roasting process in time.

[0003] Most existing coffee bean testing methods incorporate optical contrast analysis techniques. For example, a Chinese invention published under publication number CN120070368A, "Intelligent Detection System for Impurities and Defects in Green Coffee Based on Deep Learning," combines visual detection, odor detection, and optical detection to analyze impurities in green coffee. The optical detection techniques include spectral analysis and comparative analysis after image processing.

[0004] In specific optical detection processes, traditional spectral analysis usually relies on a single-wavelength light source or continuous spectral technology. In the roasting environment, light of a single wavelength is easily interfered with, which is not conducive to capturing the optical characteristics of coffee beans at different wavelengths. This causes the detection results of volume changes during coffee bean roasting to deviate from the true value. Therefore, this invention proposes an online real-time detection device and method for coffee roasting status. Summary of the Invention

[0005] The purpose of this invention is to provide an online real-time detection device and method for coffee roasting status, which quantifies the amount of light reflected by coffee beans of different wavelengths during the roasting process, and obtains relevant quantitative indicators for determining the degree of coffee roasting, so that the detection results of volume change during coffee bean roasting are close to the true value.

[0006] The specific technical solution adopted by this invention is as follows:

[0007] A coffee roasting status online real-time detection device includes a controller, which is electrically connected to a light source module, a sensor assembly, and an AD converter. The light source module is used to emit different light rays.

[0008] The light source module is externally equipped with a light source optical component for filtering light, and the sensor component is externally equipped with a sensor optical component for focusing light.

[0009] As an optional solution, the light source module includes a laser light source and an LED monochromatic light source connected to the controller;

[0010] The laser source is used to emit infrared laser light.

[0011] As an optional solution, the optical component of the light source includes a PID controller connected to the controller, and an objective lens is disposed outside the light source module. The two objective lenses respectively filter the output beam of the laser light source and the output beam of the LED monochromatic light source.

[0012] As an alternative, the sensor assembly includes a laser sensor, a CMOS camera, a CCD camera, an XYZ color sensor, or an RGB color sensor.

[0013] As an alternative, the sensor optical assembly includes a condenser cup disposed outside the sensor assembly and a focusing lens installed inside the condenser cup, wherein the focal point of the focusing lens coincides with the optical path of the sensor assembly.

[0014] As an alternative, the sensor assembly acquires image information of coffee beans, and the controller captures the boundaries of coffee beans in the image information, wherein the semi-major axis of the coffee beans is set as a, and the semi-minor axis of the coffee beans is set as b.

[0015] As an alternative, it also includes:

[0016] An inner tube surrounds the light source module and the sensor assembly, with the light source optical assembly and the sensor optical assembly passing through the inner tube;

[0017] A carrier plate is fixed inside the inner tube, and the light source module, the sensor assembly, the light source optical assembly, and the sensor optical assembly are all mounted on the surface of the carrier plate;

[0018] An aerogel layer is coated on the outer surface of the inner tube. The aerogel layer is used to seal the gap between the inner tube and the dustproof outer sheath. The aerogel layer can be made of 1250-mesh silica aerogel into a felt-like layer, which meets the safety requirements for food-grade use and has good sealing performance.

[0019] As an alternative, an air spring for shock absorption of the light source module and the sensor assembly is fixed on the top of the carrier plate, and a heat dissipation annular tube surrounding the air spring and an external fan facing the top of the heat dissipation annular tube are provided on the outside of the carrier plate.

[0020] The heat dissipation annular pipe contains a fixed amount of refrigerant.

[0021] As an alternative, the air spring is fitted with a heat-conducting pipe connected to the carrier plate, and a limiting clamp for limiting the heat dissipation annular tube is fixed on the outside of the heat-conducting pipe. The inner tube is connected to a dustproof outer sheath through the aerogel layer, and the bottom of the inner tube is connected to a light-transmitting tube for limiting the light source optical component and the sensor optical component.

[0022] A method for real-time online detection of coffee roasting status includes the following steps:

[0023] Step 1: Use sensor components to acquire images, and acquire images of coffee bean roasting at a preset frequency;

[0024] Step 2: Crop the image into several equal detection regions and adjust the contrast and brightness to improve the identifiable features of the coffee beans;

[0025] Step 3: Use edge detection and contour recognition to determine the boundaries of each coffee bean;

[0026] Step 4: Analyze the area, perimeter, or shape parameters of each identified coffee bean and estimate its size;

[0027] Step 5: Calculate the expansion rate using the expansion rate calculation formula;

[0028] Step Six: Display the data on dilation rate, contrast ratio, and brightness on the screen, update them in real time, and send the data to the host computer via Bluetooth module.

[0029] The technical effects achieved by this invention are as follows:

[0030] This invention captures the optical properties of coffee beans at different wavelengths, quantifies the amount of light reflected by coffee beans during the roasting process, and realizes real-time measurement of the degree of roasting. This device uses an algorithm to extract features from the data and generate reflectance output. Then, it processes the data through regression to obtain relevant quantitative indicators for determining the degree of coffee roasting, so that the detection results of volume change during coffee bean roasting are close to the true value.

[0031] This invention protects the optical system with inner and outer double-layer protection to isolate impurities inside the coffee roaster, and provides heat insulation and external circulation for the optical system to prevent overheating of the light source and sensor. Moreover, during the heat dissipation process, the inner and outer environments are relatively sealed to prevent external impurities from entering the coffee roaster.

[0032] This invention provides heat dissipation for the optical system while simultaneously installing a shock-absorbing structure near the heat dissipation airflow. The internal deformation of the shock-absorbing structure absorbs the shaking inside the coffee roaster and firmly holds the heat dissipation airflow in place, preventing unnecessary loosening or misalignment caused by temperature changes. Attached Figure Description

[0033] Figure 1 This is a front view of an online real-time detection device for coffee roasting status installed on a coffee roaster, according to Embodiment 1 of the present invention.

[0034] Figure 2 This is a schematic diagram of the structure of an online real-time detection device for coffee roasting status according to Embodiment 1 of the present invention;

[0035] Figure 3 This is a cross-sectional view of the inner tubing in Embodiment 1 of the present invention;

[0036] Figure 4 This is a schematic diagram of the carrier plate in Embodiment 1 of the present invention;

[0037] Figure 5 This is a partial structural schematic diagram of the carrier plate in Embodiment 1 of the present invention;

[0038] Figure 6 This is a partial cross-sectional view of the carrier plate in Embodiment 1 of the present invention;

[0039] Figure 7 This is a cross-sectional view of the coffee bean roasting state in Embodiment 1 of the present invention;

[0040] Figure 8 This is a system block diagram of the controller in Embodiment 1 of the present invention;

[0041] Figure 9 This is a flowchart of a method for online real-time detection of coffee roasting status according to Embodiment 2 of the present invention.

[0042] The attached diagram lists the components represented by each number as follows:

[0043] 1. Controller; 2. Light source module; 201. Laser light source; 202. LED monochromatic light source; 3. Sensor assembly; 4. AD converter; 5. Light source optical assembly; 501. PID controller; 502. Objective lens; 6. Sensor optical assembly; 601. Condenser cup; 602. Focusing lens; 7. Inner tube; 8. Carrier plate; 9. Aerogel layer; 10. Air spring; 11. Heat dissipation ring tube; 12. External fan; 13. Heat conduction pipe; 14. Limiting clamp; 15. Dustproof outer sheath; 16. Light transmission tube; 17. Observation window. Detailed Implementation

[0044] To make the objectives and advantages of this invention clearer, the invention will be specifically described below with reference to embodiments. It should be understood that the following text is merely used to describe one or more specific embodiments of the invention and does not strictly limit the scope of protection specifically claimed by the invention.

[0045] Example 1:

[0046] like Figures 1-8 As shown, a coffee roasting status online real-time detection device, in order to... Figure 1 Taking the coffee roaster shown as an example, it includes a controller 1, which is electrically connected to a light source module 2, a sensor assembly 3, and an AD converter 4. The controller 1 controls the light source module 2 to emit different light rays, providing light intensity inside the coffee roaster. Since the light source module 2 has a light source optical assembly 5 for filtering light, it is convenient to set the light to a certain wavelength. Then, the light is focused by the sensor optical assembly 6 on the outside of the sensor assembly 3, so that the sensor assembly 3 receives the light and obtains an electrical signal, which is sent to the controller 1. The algorithm deployed on the main control board of the controller 1 is used to calculate and measure the current volume of coffee beans. and measured at a certain frequency By comparing the volume over different time periods, the degree of volume change can be determined, and the degree of roasting of coffee beans can be judged based on multi-dimensional spectral information.

[0047] As an optional embodiment, the AD converter 4 can be a commercially available PCF8591 AD / DA conversion module, which can convert the electrical signal received by the sensor component 3 into an analog signal and transmit it to the controller 1.

[0048] See attached document Figure 1 , Figure 3 and Figure 6 To enrich the spectrum, the light source module 2 in this embodiment includes a laser light source 201 and an LED monochromatic light source 202 connected to the controller 1. The laser light source 201 is used to emit infrared laser light, such as 850nm infrared light. Since infrared light has penetrating power, it can sensitively detect changes in the roasting degree of coffee beans. A full-band light source can also be used, including a multi-spectral white light source or a halogen lamp full-band light source. In this way, a richer spectrum can be provided and the anti-interference ability is strong.

[0049] See attached document Figure 1 , Figure 5 and Figure 6 The optical component 5 of the light source includes a PID controller 501 connected to the controller 1. The PID controller 501 is electrically connected to the light source module 2. Under the control of the preset program, the light source light is stabilized, ensuring the stability of the measurement values ​​and preventing the intensity attenuation of the light source over a long period of time. After the light source light is calibrated, the light source emission achieves the appropriate energy and wavelength. An objective lens 502 is set outside the light source module 2. The two objective lenses 502 filter the output beam of the laser light source 201 and the output beam of the LED monochromatic light source 202 respectively. This can isolate the impurities flying inside the coffee bean roaster and prevent the laser light source 201 and the LED monochromatic light source 202 from being blocked by the accumulation of impurities. It has functions such as light bandwidth processing, light attenuation, light scattering, light focusing, divergence, and collimating lens.

[0050] See attached document Figure 2 and Figure 6 The sensor component 3 includes a laser sensor, a photoelectric sensor, a CMOS camera, a CCD camera, an XYZ color sensor or an RGB color sensor, etc., which capture light of different spectra and generate electrical signals. These signals are converted into analog signals by an AD converter 4 and transmitted to the controller 1. Among them, the single-point infrared laser sensor can emit single-point laser infrared light and has a receiver that can receive infrared signals.

[0051] See attached document Figure 3 and Figure 6 The sensor optical component 6 includes a condenser cup 601 disposed outside the sensor component 3 and a focusing lens 602 installed inside the condenser cup 601. Since the focal point of the focusing lens 602 coincides with the optical path of the sensor component 3, when the coffee bean reflects light, the reflected light is gathered by the condenser cup 601 and the focusing lens 602 and converged onto the sensor component 3 to sensitively capture the light. The focusing lens 602 has functions such as light bandwidth processing, light attenuation, light scattering, light focusing, divergence, and collimation lens, and enhances the signal-to-noise ratio of the signal by focusing.

[0052] See attached document Figure 1 Sensor component 3 acquires image information of coffee beans. Controller 1 uses a preset algorithm to crop the image into detection areas and adjust the contrast and brightness to improve the identifiable features of the coffee beans. For example, in a Python program, you can enter the following commands in sequence: #Read image, #Crop, #Adjust contrast and brightness, #Save result. You can also add interactive area selection or sliders to control the contrast and brightness parameters.

[0053] like Figure 7 The image shows coffee beans in their roasted state. The boundaries of the coffee beans are captured from the image information. The area, perimeter, or shape parameters of each identified coffee bean are analyzed using the outer envelope, and its size is estimated. Here, the semi-major axis of the coffee bean is set as 'a', the semi-minor axis as 'b', and the semi-thickness as a constant 'c'. The formula for the elliptical area of ​​the coffee bean is: The formula for the ellipsoidal volume of a coffee bean is: .

[0054] The specific algorithm for extracting the boundaries of coffee beans in image information is as follows:

[0055] Assuming a coffee bean can be modeled as a deformed ellipsoid, its surface equation can be expressed as:

[0056]

[0057] in, , , All are three-dimensional coordinate coefficients;

[0058] Projecting onto a two-dimensional plane, let it be I( The boundary points satisfy the following equation:

[0059]

[0060] The boundary condition equations are then derived as follows:

[0061]

[0062] when When = 0, the maximum projection boundary condition is satisfied, as follows:

[0063]

[0064] An improved edge response function based on geometric constraints, combining Canny's formula for edge detection and geometric constraints:

[0065]

[0066] in, It is a Gaussian kernel. It is the image gradient. It is a geometric distance function. and These are all geometric constraint parameters;

[0067] Then we have:

[0068]

[0069] in, The standard deviation parameter represents the Gaussian function;

[0070] The output after image processing is I' ( The calculation is as follows:

[0071]

[0072]

[0073] Then, geometrically weighted gradient calculation is performed on the output image:

[0074]

[0075]

[0076] After the above calculations, the strictness of the elliptical boundary constraint is controlled. The σ in the boundary detection formula allows for a certain degree of shape variation, such as to accommodate the small dents or defects of coffee beans.

[0077] Furthermore, images of coffee bean roasting are captured N times per minute at a preset frequency. By comparing the volume of coffee beans at different times, the degree of expansion is obtained as follows:

[0078]

[0079] Therefore, by monitoring the entire coffee bean roasting process, the function of real-time coffee bean roasting degree detection can be achieved.

[0080] See attached document Figure 2 , Figure 3 and Figure 4 In this embodiment, to reduce vibrations to the optical system during the operation of the coffee roaster, an inner tube 7, a carrier plate 8 bonded and fixed inside the inner tube 7, and an aerogel layer 9 coated on the outer surface of the inner tube 7 are also included. The inner tube 7 is used to surround the light source module 2 and the sensor assembly 3, isolating them from external impurities. The light source optical assembly 5 and the sensor optical assembly 6 pass through the inner tube 7 to allow light to enter. The light source module 2, the sensor assembly 3, the light source optical assembly 5, and the sensor optical assembly 6 are all glued to the surface of the carrier plate 8. By inserting the inner tube 7 in one step, the optical system can be installed in the coffee roaster, making the operation simple.

[0081] See attached document Figure 4 , Figure 5 and Figure 6 In this embodiment, an air spring 10 for shock absorption of the light source module 2 and sensor assembly 3 is bonded to the top of the carrier plate 8. The air spring 10 is filled with a sufficient amount of inert gas, such as argon, which does not expand significantly in the high-temperature environment inside the coffee roaster, thus maintaining its shape. In addition, its elastic properties allow it to absorb shock from the carrier plate 8 through its own deformation. Furthermore, a heat dissipation annular pipe 11 surrounding the air spring 10 and an external fan 12 facing the top of the heat dissipation annular pipe 11 are provided outside the carrier plate 8. Since the heat dissipation annular pipe 11 is filled with a certain amount of refrigerant, such as ammonia or non-halogenated hydrocarbons, it will evaporate after absorbing heat inside the inner tube 7, drift to the top of the heat dissipation annular pipe 11, and the heat will be carried away by the external fan 12. After cooling, it will liquefy and flow back to the inner tube 7 along the heat dissipation annular pipe 11, thus achieving cyclic cooling.

[0082] See attached document Figure 5 and Figure 6To straighten the air spring 10, this embodiment has a heat-conducting pipe 13 welded to the carrier plate 8 on the outside of the air spring 10. Since the top of the heat-conducting pipe 13 is attached to the inner tube 7, it can completely wrap the inner tube 7 and limit its position. At the same time, an arc-shaped limiting clamp 14 is welded to the outside of the heat-conducting pipe 13 to limit the heat dissipation annular tube 11. By clamping the heat dissipation annular tube 11 at multiple points by the limiting clamp 14, the heat dissipation annular tube 11 can be quickly positioned and limited, preventing the heat dissipation annular tube 11 from falling off due to changes in the form of the refrigerant. A dustproof outer sheath 15 is connected to the outside of the inner tube 7 through an aerogel layer 9. The gaps between the three are filled with glue for insertion into a coffee roaster to prevent flying coffee beans or impurities. In order to facilitate the entry and exit of light, a light-transmitting tube 16 is connected to the bottom of the inner tube 7 to limit the optical components of the light source 5 and the optical components of the sensor 6. Light can pass through, and the gaps in the inner wall of the light-transmitting tube 16 are sealed with glue.

[0083] The control signal transmission method of controller 1 in this embodiment is described in [reference]. Figure 8 It, along with the AD converter 4 and PID regulator 501, are all located outside the coffee roaster, for example, in a nearby low-voltage box or control cabinet. The coffee roaster has an observation window 17 to facilitate visual observation of the coffee beans inside.

[0084] Example 2:

[0085] To facilitate staff in quickly mastering the detection of coffee roasting status, this embodiment provides specific implementation steps for the online real-time detection of coffee roasting status based on the embodiment's online real-time detection device. This allows staff to capture light of different wavelengths through sensor vision and extract identifiable features from the acquired data using the algorithm in Embodiment 1, thereby detecting the expansion rate of coffee beans.

[0086] like Figure 9 As shown, a method for online real-time detection of coffee roasting status includes the following steps:

[0087] Step 1: Use sensor component 3 to acquire images, and acquire images of coffee bean roasting at a preset frequency; where the frequency is N images acquired per second.

[0088] Step 2: Crop the image into several equal detection regions and adjust the contrast and brightness to improve the identifiable features of the coffee beans;

[0089] Step 3: Use edge detection and contour recognition methods to identify the boundaries of each coffee bean;

[0090] Step 4: Analyze the area, perimeter, or shape parameters of each identified coffee bean and estimate its size;

[0091] Let the semi-major axis of the coffee bean be *a*, the semi-minor axis be *b*, and the semi-thickness be a constant *c*. Then the formula for the area of ​​the ellipse of the coffee bean is: The formula for the ellipsoidal volume of a coffee bean is: ;

[0092] Furthermore, based on whether coffee beans can pass through sieves of different sizes, the measured dimensions are converted into corresponding mesh counts. The minimum passing diameter of the coffee beans is determined graphically. Since the sieve opening size is fixed, if the minimum passing diameter of the coffee beans is smaller than the sieve opening size, they are assigned to the next mesh count level. By calculating the passing diameter of the coffee beans using the graph, the mesh count distribution can be calculated.

[0093] Step 5: Calculate the expansion rate using the expansion rate calculation formula;

[0094]

[0095] Therefore, by detecting the entire coffee bean roasting process, the function of real-time coffee bean roasting degree detection can be realized;

[0096] Step Six: Display the data on dilatation rate, contrast ratio, and brightness on the screen, update them in real time, and send the data to the host computer via Bluetooth module.

[0097] In summary, by quantifying the amount of light reflected by coffee beans of different wavelengths during the roasting process through steps one through six, the degree of roasting can be measured in real time, the expansion rate of coffee beans can be estimated, and data used to determine the degree of coffee roasting can be obtained.

[0098] This embodiment can also generate reflectance output from the data by threshold filtering, and then process the data using regression to obtain the target temperature T. The regression calculation is as follows:

[0099]

[0100] in, For reflectivity, The centroid of the spectrum represents the frequency domain characteristics. The standard deviation of the signal represents its time-domain characteristics. , , , All are preset model weights. This indicates the error term.

[0101] The above description is merely an optional embodiment of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention. Structures, devices, and operating methods not specifically described or explained in this invention, unless otherwise specified or limited, shall be implemented according to conventional means in the art.

Claims

1. A coffee roasting status online real-time detection device, comprising a controller (1), characterized in that: The controller (1) is electrically connected to a light source module (2), a sensor assembly (3), and an AD converter (4). The light source module (2) is used to emit different light rays. The light source module (2) is externally provided with a light source optical component (5) for filtering light, and the sensor component (3) is externally provided with a sensor optical component (6) for focusing light. Also includes: An inner tube (7) is used to surround the light source module (2) and the sensor assembly (3), and the light source optical assembly (5) and the sensor optical assembly (6) pass through the inner tube (7). The carrier plate (8) is fixed inside the inner tube (7). The light source module (2), the sensor assembly (3), the light source optical assembly (5) and the sensor optical assembly (6) are all mounted on the surface of the carrier plate (8). An air spring (10) for shock absorption of the light source module (2) and the sensor assembly (3) is fixed on the top of the carrier plate (8). A heat dissipation annular tube (11) surrounding the air spring (10) and an external fan (12) facing the top of the heat dissipation annular tube (11) are provided on the outside of the carrier plate (8). An aerogel layer (9) is coated on the outer surface of the inner tube (7); The heat dissipation annular pipe (11) contains a fixed amount of refrigerant. The air spring (10) is fitted with a heat-conducting pipe (13) connected to the carrier plate (8). A limiting clamp (14) for limiting the heat dissipation annular pipe (11) is fixed on the outside of the heat-conducting pipe (13). A dustproof outer sheath (15) is connected to the outside of the inner tube (7) through the aerogel layer (9). A light-transmitting tube (16) for limiting the light source optical component (5) and the sensor optical component (6) is connected to the bottom of the inner tube (7).

2. The online real-time detection device for coffee roasting status according to claim 1, characterized in that: The light source module (2) includes a laser light source (201) and an LED monochromatic light source (202) connected to the controller (1). The laser source (201) is used to emit infrared laser.

3. The online real-time detection device for coffee roasting status according to claim 2, characterized in that: The optical component of the light source (5) includes a PID regulator (501) connected to the controller (1). An objective lens (502) is provided on the outside of the light source module (2). The two objective lenses (502) respectively filter the emitted beam of the laser light source (201) and the emitted beam of the LED monochromatic light source (202).

4. The online real-time detection device for coffee roasting status according to claim 1, characterized in that: The sensor assembly (3) includes a laser sensor, a CMOS camera, a CCD camera, an XYZ color sensor, or an RGB color sensor.

5. The online real-time detection device for coffee roasting status according to claim 1, characterized in that: The sensor optical component (6) includes a condenser cup (601) disposed outside the sensor component (3) and a focusing lens (602) installed inside the condenser cup (601), the focal point of the focusing lens (602) coinciding with the optical path of the sensor component (3).

6. The online real-time detection device for coffee roasting status according to claim 1, characterized in that: The sensor component (3) acquires image information of coffee beans, and the controller (1) captures the boundary of coffee beans in the image information, wherein the semi-major axis of the coffee beans is set as a and the semi-minor axis of the coffee beans is set as b.

7. A method for online real-time detection of coffee roasting status, applied to the online real-time detection device for coffee roasting status as described in any one of claims 1-6, characterized in that, Includes the following steps: Step 1: Use sensor component (3) to acquire images and acquire coffee bean roasting images at a preset frequency; Step 2: Crop the image into several equal detection regions and adjust the contrast and brightness to improve the identifiable features of the coffee beans; Step 3: Use edge detection and contour recognition to determine the boundaries of each coffee bean; Step 4: Analyze the area, perimeter, or shape parameters of each identified coffee bean and estimate its size; Step 5: Calculate the expansion rate using the expansion rate calculation formula; Step Six: Display the data on dilation rate, contrast ratio, and brightness on the screen, update them in real time, and send the data to the host computer via Bluetooth module.

Images