Icing sensor calibration method and system based on binocular vision method

Abstract

The invention relates to the technical field of aviation icing detection, in particular to an icing sensor calibration method and system based on a binocular vision method.The icing sensor calibration method based on the binocular vision method comprises the following steps that in an icing environment, an icing sensor continuously collects icing information of a detection surface of the icing sensor, and the icing information is obtained; obtaining a real-time photoelectric response value x of the icing sensor; a binocular vision system in a non-icing environment is adopted to continuously collect icing process images of a detection surface of an icing sensor, and a real-time icing image pair sequence of the detection surface of the icing sensor is obtained; on the basis of the binocular vision principle, three-dimensional reconstruction is carried out on real-time icing of the icing sensor detection surface, a real-time ice surface three-dimensional image of the icing sensor detection surface is obtained, and a real-time ice layer thickness value y of the icing sensor detection surface is obtained; and associating the real-time ice layer thickness value y of the detection surface of the icing sensor with the real-time photoelectric response value x of the icing sensor, and fitting to obtain a function association formula y = f (x).

Description

Technical Field

[0001] The present invention relates to the technical field of aviation icing detection, and in particular to an icing sensor calibration method and system based on binocular vision method. Background Art

[0002] Ice sensors are components commonly used in aircraft to detect ice information on wings, fuselages, engine air inlets, etc., and have functions such as ice warning and ice thickness detection. There are many types of ice sensors, among which the optical fiber ice sensor is embedded inside the detection surface without destroying the surface shape, and has the advantages of high detection accuracy, simple structure, and strong anti-interference.

[0003] The optical fiber ice sensor needs to be calibrated before it is put into use. The calibration method is generally to use an ice sensor to measure the thickness of the ice layer, obtain the photoelectric response curve of the ice sensor to the ice layer thickness, and use other methods to measure the actual ice layer thickness. Then, the response curve of the ice sensor is associated with the actual ice thickness to complete the calibration. The calibration environment is generally an ice wind tunnel. During calibration, after setting the airflow and cloud field parameters, the wind tunnel is turned on to make the ice sensor detection surface ice up, and the ice sensor measures the response curve. When the ice reaches a certain thickness value, the wind tunnel is closed, and the actual ice thickness is measured after the wind tunnel is closed. The method for measuring the actual ice thickness generally uses a ruler to measure. Since the wind tunnel needs to be closed when using a ruler to measure the actual ice thickness, this calibration method can generally only obtain the final ice thickness value after the wind tunnel is closed, and the ice thickness value in the intermediate process of ice formation cannot be obtained. Therefore, only the final ice thickness value is associated with the photoelectric response curve of the ice sensor. Since the growth rate of ice thickness during the freezing process is variable in most cases, only using the final ice thickness value for calibration will result in low calibration accuracy. If the wind tunnel is closed several times during the freezing process to measure the actual ice thickness with a ruler, and then the wind tunnel is reopened to continue freezing, although the actual ice thickness value in the middle of the freezing process can be obtained, the ice type formed after closing and reopening the wind tunnel often differs, resulting in calibration errors. Summary of the invention

[0004] The object of the present invention is to provide an ice sensor calibration method based on binocular vision method to solve the problem of low accuracy of calibration results obtained by the calibration method in the prior art.

[0005] In order to solve the above problems, the present invention proposes an ice sensor calibration method based on binocular vision method, and the technical solution adopted is: A method for calibrating an ice sensor based on binocular vision method comprises the following steps: Step S1: In an icing environment, the icing sensor continuously collects icing information on its detection surface to obtain a real-time photoelectric response value x of the icing sensor; at the same time, a binocular vision system in a non-icing environment is used to continuously collect icing process images of the icing sensor detection surface to obtain a real-time icing image pair sequence of the icing sensor detection surface; Step S2: combining the real-time ice image pair sequence of the ice sensor detection surface and based on the binocular vision principle, three-dimensionally reconstructing the ice on the real-time ice sensor detection surface to obtain a real-time three-dimensional image of the ice surface on the ice sensor detection surface; Step S3, obtaining a real-time ice thickness value y of the ice sensor detection surface based on the real-time three-dimensional image of the ice surface of the ice sensor detection surface; Step S4: Correlate the ice thickness value y of the real-time icing sensor detection surface with the photoelectric response value x of the real-time icing sensor, and obtain a function correlation formula y=f(x) by fitting.

[0006] Furthermore, in step S1, in the icing environment, the icing sensor continuously collects icing information on its detection surface to obtain a real-time photoelectric response value x of the icing sensor, including: In an icing environment, ice forms on the detection surface of the ice sensor. The ice sensor continuously collects ice information on its detection surface to obtain a real-time response curve of the ice sensor to the ice information, and then obtains a real-time photoelectric response value x of the ice sensor.

[0007] Further, in step S3, the real-time ice layer thickness value y of the ice sensor detection surface is obtained based on the real-time three-dimensional image of the ice surface of the ice sensor detection surface, including: A plurality of sampling points are selected in the three-dimensional image of the ice surface of the real-time icing sensor detection surface to obtain ice thickness values ​​of the plurality of sampling points, and the ice thickness values ​​of the plurality of sampling points are averaged to obtain the real-time ice thickness value y of the icing sensor detection surface.

[0008] Furthermore, the number of the sampling points is greater than or equal to 30.

[0009] Furthermore, in step S2, combining the real-time ice image pair sequence of the ice sensor detection surface and reconstructing the ice on the real-time ice sensor detection surface in three dimensions based on the binocular vision principle to obtain the real-time ice surface three-dimensional image of the ice sensor detection surface includes: The surface feature points of the ice image pair sequence of the real-time ice sensor detection surface are extracted and matched to obtain the matched surface feature points. Based on the principle of binocular vision, the ice on the real-time ice sensor detection surface is three-dimensionally reconstructed according to the matched surface feature points, external parameters of the binocular camera, and internal parameters of the binocular camera to obtain the real-time three-dimensional image of the ice surface on the ice sensor detection surface.

[0010] Furthermore, the three-dimensional reconstruction of the ice on the real-time ice sensor detection surface based on the matched surface feature points, the external parameters of the binocular camera, and the internal parameters of the binocular camera includes: A stereo matching algorithm is used to construct a local point cloud of the ice image of the real-time ice sensor detection surface. The spatial relationship between adjacent image pairs is obtained through the ice image matching of the real-time ice sensor detection surface, and the global point cloud of the real-time ice sensor detection surface is established, thereby realizing three-dimensional reconstruction of the ice on the real-time ice sensor detection surface.

[0011] Furthermore, before step S1, an icing sensor is arranged in an icing environment; and a binocular vision system is arranged outside the icing environment, that is, in a non-icing environment.

[0012] Furthermore, an icing component to be detected is also provided in the icing environment, and the detection surface of the icing sensor is in the same direction and position as the icing surface to be detected of the icing component to be detected.

[0013] The present invention also provides a system for executing the above-mentioned ice sensor calibration method based on binocular vision method, comprising: The data acquisition module is used for the ice sensor to continuously collect ice information on its detection surface in an icing environment to obtain a real-time photoelectric response value x of the ice sensor; at the same time, the binocular vision system in a non-icing environment is used to continuously collect ice process images of the ice sensor detection surface to obtain a real-time sequence of ice image pairs of the ice sensor detection surface; The ice surface three-dimensional image acquisition module of the ice sensor detection surface is used to combine the real-time ice image pair sequence of the ice sensor detection surface and reconstruct the ice on the real-time ice sensor detection surface in three dimensions based on the binocular vision principle to obtain the real-time three-dimensional image of the ice surface of the ice sensor detection surface; The ice thickness value acquisition module of the ice sensor detection surface is used to obtain the real-time ice thickness value y of the ice sensor detection surface based on the real-time three-dimensional image of the ice surface of the ice sensor detection surface; The fitting module is used to associate the ice thickness value y of the real-time ice sensor detection surface with the photoelectric response value x of the real-time ice sensor, and obtain the function correlation formula y=f(x) by fitting.

[0014] Furthermore, the data acquisition module includes a photoelectric response value acquisition module of the icing sensor and an icing image pair sequence acquisition module of the icing sensor detection surface. The photoelectric response value acquisition module of the icing sensor is used for the icing sensor to continuously collect icing information on its detection surface in an icing environment to obtain a real-time photoelectric response value x of the icing sensor; The icing image pair sequence acquisition module of the icing sensor detection surface is used to continuously collect icing process images of the icing sensor detection surface using a binocular vision system in a non-icing environment to obtain a real-time icing image pair sequence of the icing sensor detection surface.

[0015] Beneficial effects: The present invention is an improved invention. The icing sensor calibration method based on binocular vision method proposed in the present invention can obtain the real-time icing thickness value of the icing sensor detection surface by using the icing sensor in an icing environment to continuously collect the icing information of its detection surface. At the same time, the binocular vision system in a non-icing environment is used to continuously collect the icing process image of the icing sensor detection surface, that is, a plurality of real ice thickness values ​​in the icing process are continuously obtained under non-contact conditions for calibrating the icing sensor, and the relationship between the icing sensor response value and the real ice thickness value is fitted to improve the calibration accuracy, solving the problem that the previous calibration method only has the final ice thickness value as the real ice thickness. At the same time, the calibration method proposed in the present invention does not need to repeatedly open / close the icing environment, and can obtain the real ice thickness value without closing the icing environment, solving the calibration error problem caused by inconsistent ice types, and improving the calibration accuracy of the sensor. At the same time, it saves a lot of time, reduces the operating cost, and can avoid the problem of repeated opening / closing of the icing environment and inconsistent icing ice types during repeated opening / closing.

[0016] In step S3, the step of obtaining a real-time ice thickness value y of the ice layer on the detection surface of the ice sensor based on the real-time three-dimensional image of the ice surface on the detection surface of the ice sensor comprises: A plurality of sampling points are selected in the three-dimensional image of the ice surface of the real-time ice sensor detection surface to obtain ice thickness values ​​of the plurality of sampling points, and the ice thickness values ​​of the plurality of sampling points are averaged to obtain the real-time ice thickness value y of the ice sensor detection surface, thereby improving the accuracy of the real-time ice thickness value y of the ice sensor detection surface.

[0017] The number of the sampling points is greater than or equal to 30, which reduces the calculation cost while ensuring the accuracy of the ice thickness value y of the real-time icing sensor detection surface. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a flow chart of the ice sensor calibration method based on binocular vision method of the present invention. DETAILED DESCRIPTION

[0019] As cited in the background art, the calibration results obtained by the calibration methods in the prior art have low accuracy. Therefore, the present invention provides an ice sensor calibration method based on binocular vision method, comprising the following steps: first, in an icing environment, the ice sensor continuously collects icing information on its detection surface to obtain a real-time photoelectric response value x of the ice sensor, that is, obtains the photoelectric response value x of the ice sensor under different thicknesses; a binocular vision system is used in a non-icing environment to continuously collect images of the icing process on the detection surface of the ice sensor to obtain a real-time sequence of icing image pairs on the detection surface of the ice sensor, that is, obtains a sequence of icing image pairs on the detection surface of the ice sensor under different thicknesses; secondly, combined with the real-time icing The ice sensor detection surface is obtained by obtaining a sequence of ice images, and based on the binocular vision principle, the real-time ice on the ice sensor detection surface is reconstructed in three dimensions to obtain a real-time three-dimensional image of the ice surface on the ice sensor detection surface, which is used to calculate the real-time ice thickness value y of the ice sensor detection surface; then, based on the real-time three-dimensional image of the ice surface on the ice sensor detection surface, the real-time ice thickness value y of the ice sensor detection surface is obtained; finally, the real-time ice thickness value y of the ice sensor detection surface is associated with the real-time photoelectric response value x of the ice sensor, and the function correlation formula y=f(x) is obtained by fitting. The ice sensor calibration method based on binocular vision proposed in the present invention can improve the calibration accuracy and solve the problem that the previous calibration method only has the final ice thickness value as the real ice thickness. At the same time, there is no need to repeatedly turn on / off the ice environment, and the real ice thickness value can be obtained without turning off the ice environment, which solves the calibration error problem caused by inconsistent ice types and improves the calibration accuracy of the sensor. At the same time, a lot of time is saved, operating costs are reduced, and repeated opening / closing of the icing environment and the problem of inconsistent ice types during repeated opening / closing can be avoided.

[0020] Embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.

[0021] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0022] The following describes an ice sensor calibration method and system based on binocular vision method according to an embodiment of the present application with reference to the accompanying drawings.

[0023] Combine the following Figure 1 , the ice sensor calibration method based on binocular vision method provided in this application is described in detail.

[0024] Step S1: In an icing environment, an icing sensor continuously collects icing information on its detection surface to obtain a real-time photoelectric response value x of the icing sensor; at the same time, a binocular vision system in a non-icing environment is used to continuously collect icing process images on the detection surface of the icing sensor to obtain a real-time icing image pair sequence of the detection surface of the icing sensor.

[0025] Specifically, in an icing environment, the ice sensor continuously collects ice information on its detection surface to obtain a real-time photoelectric response value x of the ice sensor, including: in an icing environment, ice forms on the detection surface of the ice sensor, the ice sensor continuously collects ice information on its detection surface to obtain a real-time response curve of the ice sensor to the ice information, and then obtains the real-time photoelectric response value x of the ice sensor.

[0026] Here, before step S1, the icing sensor is set in the icing environment; the binocular vision system is set outside the icing environment, that is, in a non-icing environment. An icing component to be detected is also set in the icing environment, and the detection surface of the icing sensor is in the same direction and position as the icing surface to be detected of the icing component to be detected.

[0027] In this embodiment, an application example in an open icing wind tunnel environment is taken as an example. An icing sensor and a corresponding icing component to be tested are placed at the exit of the icing wind tunnel. In this embodiment, the icing component to be tested is a wing model, and the icing sensor is arranged inside the wing airfoil. The icing sensor detection surface is at the leading edge of the wing. The icing sensor detection surface is in line with the arc of the leading edge of the wing and is a part of the leading edge of the wing, that is, the icing sensor detection surface is in the same direction and position as the leading edge of the wing. In this way, when the wing is frozen, ice will also form on the icing sensor detection surface. A binocular vision system is arranged outside the icing wind tunnel. The binocular vision system includes two cameras and a light source. The light source is a yellow light source. The two cameras are placed at a certain angle and aimed at the icing sensor detection surface of the leading edge of the wing.

[0028] During calibration, certain parameters such as wind speed, temperature, droplet size, water flow rate, etc. are set, and the icing wind tunnel is turned on to freeze the wing model and the icing sensor detection surface. The icing sensor continuously obtains the icing information of its detection surface, obtains the response curve of the icing sensor to the icing information, and then obtains the real-time photoelectric response value x of the icing sensor. At the same time, two cameras are used to take real-time photos of the icing process of the icing sensor detection surface, that is, to take real-time photos of the ice surface shape of the icing sensor detection surface. Here, that is, during the operation of the icing wind tunnel, the non-contact binocular vision method is used to measure the ice thickness of the icing sensor detection surface, and the real-time multiple measurements of the ice thickness are achieved without closing the wind tunnel. Multiple real icing thickness values ​​are obtained during the opening of the wind tunnel for calibration of the icing sensor.

[0029] It should be noted that the two cameras scan and shoot the ice sensor detection surface through horizontal translation. The shooting range is 60mm of the wing leading edge and the scanning process is 1s. Although the ice process is an unsteady dynamic process, the maximum ice speed is about 0.1mm / s and the ice surface change is small during the scanning time, so it can be ignored, so that the whole process can be uninterrupted ice measurement and a series of images can be obtained. The ice sensor detection surface includes a light transmitting end and a receiving end. The receiving signal of the receiving end will be converted into a photoelectric response value. This photoelectric response value will change with the thickness of the ice layer. For example, as the ice thickness increases, the photoelectric response values ​​at n moments are obtained, and the list is [80, 136, 256, 335, 458, 537, 678, 790, 913, 1023,1356, 1400...], and there are n values ​​in the list.

[0030] Step S2: combining the real-time ice image pair sequence of the ice sensor detection surface and based on the binocular vision principle, three-dimensionally reconstructing the ice on the real-time ice sensor detection surface to obtain a real-time three-dimensional image of the ice surface on the ice sensor detection surface.

[0031] Specifically, in combination with a sequence of ice image pairs of a real-time ice sensor detection surface, and based on the principle of binocular vision, the real-time ice on the ice sensor detection surface is three-dimensionally reconstructed to obtain a real-time three-dimensional image of the ice surface on the ice sensor detection surface, including: extracting and matching surface feature points of the real-time ice image pair sequence of the ice sensor detection surface to obtain matched surface feature points, and based on the principle of binocular vision, according to the matched surface feature points, external parameters of the binocular camera, and internal parameters of the binocular camera, the real-time ice on the ice sensor detection surface is three-dimensionally reconstructed to obtain a real-time three-dimensional image of the ice surface on the ice sensor detection surface.

[0032] Here, based on the matched surface feature points, external parameters of the binocular camera and internal parameters of the binocular camera, the real-time ice detection surface of the ice sensor is three-dimensionally reconstructed, including: using a stereo matching algorithm to construct a local point cloud of the ice image of the real-time ice detection surface of the ice sensor, and obtaining the spatial relationship between adjacent image pairs through ice image matching of the real-time ice detection surface of the ice sensor, establishing a global point cloud of the real-time ice detection surface of the ice sensor, and then realizing the three-dimensional reconstruction of the ice on the real-time ice detection surface of the ice sensor.

[0033] For example, the real-time three-dimensional image of the ice surface of the ice sensor detection surface is determined as the three-dimensional image of the ice surface of the ice sensor detection surface at n moments. It should be noted that the n moments for determining the three-dimensional image of the ice surface of the ice sensor detection surface are the same as the n moments for determining the photoelectric response value in step S1.

[0034] Step S3: based on the real-time three-dimensional image of the ice surface on the detection surface of the icing sensor, obtain the real-time ice layer thickness value y on the detection surface of the icing sensor.

[0035] Since the ice surface of the ice sensor detection surface is not flat but uneven, it is necessary to calculate the average ice thickness value of the ice layer on the ice detection surface. Specifically, based on the real-time three-dimensional image of the ice surface of the ice sensor detection surface, the real-time ice layer thickness value y of the ice sensor detection surface is obtained, including: selecting multiple sampling points in the real-time three-dimensional image of the ice surface of the ice sensor detection surface, obtaining the ice layer thickness values ​​of the multiple sampling points, and averaging the ice layer thickness values ​​of the multiple sampling points to obtain the real-time ice layer thickness value y of the ice sensor detection surface. Here, the number of sampling points is greater than or equal to 30.

[0036] For example, in the three-dimensional images of the ice surface of the ice sensor detection surface at n moments in step S2, m sampling points (m ≥ 30) are selected within the range of the ice sensor detection surface for the three-dimensional images of the ice surface of the ice sensor detection surface at each moment, and the ice thickness values ​​of the m sampling points are averaged to obtain the ice thickness value of the ice sensor detection surface at each moment. In this way, the ice thickness values ​​of the ice sensor detection surface at n moments are obtained, which are [0.2, 0.5, 0.8, 1.1,1.4, 1.7, 2.0, 2.3, 2.6, 2.9, 3.2, 3.5, 3.8, 4.1……], and there are n values ​​in the list.

[0037] Step S4: Correlate the ice thickness value y of the real-time icing sensor detection surface with the photoelectric response value x of the real-time icing sensor, and obtain a function correlation formula y=f(x) by fitting.

[0038] For example, the ice thickness values ​​y of the detection surface of the ice sensor at n moments obtained in step S3 are correlated with the photoelectric response values ​​x of the ice sensor at n moments obtained in step S1, and the function correlation formula y=f(x) is obtained by fitting.

[0039] The present application also provides a system for executing the above-mentioned ice sensor calibration method based on binocular vision method, comprising: The data acquisition module is used for the ice sensor to continuously collect ice information on its detection surface in an icing environment to obtain a real-time photoelectric response value x of the ice sensor; at the same time, the binocular vision system in a non-icing environment is used to continuously collect ice process images of the ice sensor detection surface to obtain a real-time sequence of ice image pairs of the ice sensor detection surface; The ice surface three-dimensional image acquisition module of the ice sensor detection surface is used to combine the real-time ice image pair sequence of the ice sensor detection surface and reconstruct the ice on the real-time ice sensor detection surface in three dimensions based on the binocular vision principle to obtain the real-time three-dimensional image of the ice surface of the ice sensor detection surface; The ice thickness value acquisition module of the ice sensor detection surface is used to obtain the real-time ice thickness value y of the ice sensor detection surface based on the real-time three-dimensional image of the ice surface of the ice sensor detection surface; The fitting module is used to associate the ice thickness value y of the real-time ice sensor detection surface with the photoelectric response value x of the real-time ice sensor, and obtain the function correlation formula y=f(x) by fitting.

[0040] Among them, the data acquisition module includes a photoelectric response value acquisition module of the ice sensor and an ice image pair sequence acquisition module of the ice sensor detection surface. The photoelectric response value acquisition module of the ice sensor is used for the ice sensor to continuously collect ice information of its detection surface in an icing environment to obtain a real-time photoelectric response value x of the ice sensor; the ice image pair sequence acquisition module of the ice sensor detection surface is used for continuously collecting ice process images of the ice sensor detection surface using a binocular vision system in a non-icing environment to obtain a real-time ice image pair sequence of the ice sensor detection surface.

[0041] Here, those skilled in the art will appreciate that the specific method of the ice sensor calibration system based on binocular vision method has been described in the above reference. Figure 1 The description of the ice sensor calibration method based on binocular vision method has been introduced in detail, so its repeated description will be omitted.

[0042] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. The patent protection scope of the present invention shall be based on the claims. All equivalent structural changes made using the contents of the description and drawings of the present invention should also be included in the protection scope of the present invention.

Claims

1. A method for calibrating an ice sensor based on binocular vision, characterized in that: The following steps are involved: Step S1: In an icing environment, the icing sensor continuously collects icing information on its detection surface to obtain a real-time photoelectric response value x of the icing sensor; at the same time, a binocular vision system in a non-icing environment is used to continuously collect icing process images of the icing sensor detection surface to obtain a real-time icing image pair sequence of the icing sensor detection surface; Step S2: combining the real-time ice image pair sequence of the ice sensor detection surface and based on the binocular vision principle, three-dimensionally reconstructing the ice on the real-time ice sensor detection surface to obtain a real-time three-dimensional image of the ice surface on the ice sensor detection surface; Step S3, obtaining a real-time ice thickness value y of the ice sensor detection surface based on the real-time three-dimensional image of the ice surface of the ice sensor detection surface; Step S4: Correlate the ice thickness value y of the real-time icing sensor detection surface with the photoelectric response value x of the real-time icing sensor, and obtain a function correlation formula y=f(x) by fitting.

2. The ice sensor calibration method based on binocular vision method according to claim 1 is characterized in that: In step S1, in the icing environment, the icing sensor continuously collects icing information on its detection surface to obtain a real-time photoelectric response value x of the icing sensor, including: In an icing environment, ice forms on the detection surface of the ice sensor. The ice sensor continuously collects ice information on its detection surface to obtain a real-time response curve of the ice sensor to the ice information, and then obtains a real-time photoelectric response value x of the ice sensor.

3. The ice sensor calibration method based on binocular vision method according to claim 1 is characterized in that: In step S3, the step of obtaining a real-time ice thickness value y of the ice layer on the detection surface of the ice sensor based on the real-time three-dimensional image of the ice surface on the detection surface of the ice sensor comprises: A plurality of sampling points are selected in the three-dimensional image of the ice surface of the real-time icing sensor detection surface to obtain ice thickness values ​​of the plurality of sampling points, and the ice thickness values ​​of the plurality of sampling points are averaged to obtain the real-time ice thickness value y of the icing sensor detection surface.

4. The ice sensor calibration method based on binocular vision method according to claim 3 is characterized in that: The number of the sampling points is greater than or equal to 30.

5. The ice sensor calibration method based on binocular vision method according to claim 1, characterized in that: In step S2, combining the real-time ice image pair sequence of the ice sensor detection surface and reconstructing the ice on the real-time ice sensor detection surface in three dimensions based on the binocular vision principle to obtain a real-time three-dimensional image of the ice surface on the ice sensor detection surface includes: The surface feature points of the ice image pair sequence of the real-time ice sensor detection surface are extracted and matched to obtain the matched surface feature points. Based on the principle of binocular vision, the ice on the real-time ice sensor detection surface is three-dimensionally reconstructed according to the matched surface feature points, external parameters of the binocular camera, and internal parameters of the binocular camera to obtain the real-time three-dimensional image of the ice surface on the ice sensor detection surface.

6. The ice sensor calibration method based on binocular vision method according to claim 5 is characterized in that: The method of performing three-dimensional reconstruction of ice on the real-time ice sensor detection surface based on the matched surface feature points, the external parameters of the binocular camera, and the internal parameters of the binocular camera includes: A stereo matching algorithm is used to construct a local point cloud of the ice image of the real-time ice sensor detection surface. The spatial relationship between adjacent image pairs is obtained through the ice image matching of the real-time ice sensor detection surface, and the global point cloud of the real-time ice sensor detection surface is established, thereby realizing three-dimensional reconstruction of the ice on the real-time ice sensor detection surface.

7. The ice sensor calibration method based on binocular vision method according to claim 1, characterized in that: Before step S1, an icing sensor is set in an icing environment; and a binocular vision system is set outside the icing environment, that is, in a non-icing environment.

8. The ice sensor calibration method based on binocular vision method according to claim 7, characterized in that: An icing component to be detected is also arranged in the icing environment, and the detection surface of the icing sensor is in the same direction and position as the icing surface to be detected of the icing component to be detected.

9. A system for executing the ice sensor calibration method based on binocular vision method as described in any one of claims 1 to 8, characterized in that: include: The data acquisition module is used for the ice sensor to continuously collect ice information on its detection surface in an icing environment to obtain a real-time photoelectric response value x of the ice sensor; at the same time, the binocular vision system in a non-icing environment is used to continuously collect ice process images of the ice sensor detection surface to obtain a real-time sequence of ice image pairs of the ice sensor detection surface; The ice surface three-dimensional image acquisition module of the ice sensor detection surface is used to combine the real-time ice image pair sequence of the ice sensor detection surface and reconstruct the ice on the real-time ice sensor detection surface in three dimensions based on the binocular vision principle to obtain the real-time three-dimensional image of the ice surface of the ice sensor detection surface; The ice thickness value acquisition module of the ice sensor detection surface is used to obtain the real-time ice thickness value y of the ice sensor detection surface based on the real-time three-dimensional image of the ice surface of the ice sensor detection surface; The fitting module is used to associate the ice thickness value y of the real-time ice sensor detection surface with the photoelectric response value x of the real-time ice sensor, and obtain the function correlation formula y=f(x) by fitting.

10. The ice sensor calibration system based on binocular vision method according to claim 9, characterized in that: The data acquisition module includes a photoelectric response value acquisition module of the icing sensor and an icing image pair sequence acquisition module of the icing sensor detection surface. The photoelectric response value acquisition module of the icing sensor is used for the icing sensor to continuously collect icing information on its detection surface in an icing environment to obtain a real-time photoelectric response value x of the icing sensor; The icing image pair sequence acquisition module of the icing sensor detection surface is used to continuously collect icing process images of the icing sensor detection surface using a binocular vision system in a non-icing environment to obtain a real-time icing image pair sequence of the icing sensor detection surface.

Images