A method, device and equipment for determining asphalt adhesion index

Abstract

The present application provides a method, device and equipment for determining asphalt adhesion index, the method comprising: obtaining target spectral data corresponding to a mixture, the mixture being composed of asphalt and stone; the target spectral data comprising K target spectral images, different target spectral images corresponding to different bands; for each pixel point of the target spectral image, determining the category of the pixel point based on the K spectral intensity values ​​corresponding to the pixel point in the K target spectral images, the category being stone category or non-stone category; determining the index parameter corresponding to the mixture based on the category of each pixel point; determining the asphalt adhesion index corresponding to the mixture based on the index parameter. Through the scheme of the present application, an accurate asphalt adhesion index can be obtained.

Description

Technical Field

[0001] The present application relates to the field of spectral imaging technology, and in particular to a method, device and equipment for determining asphalt adhesion index. Background Art

[0002] Asphalt is a waterproof, moisture-proof and anti-corrosion organic gelling material, composed of hydrocarbons of different molecular weights and their non-metallic derivatives. Asphalt is a complex mixture of black-brown and high-viscosity organic liquid. Asphalt is mainly used in industries such as coatings, plastics, rubber, and road paving.

[0003] Asphalt adhesion refers to the bonding performance of asphalt and stone under different temperature and humidity conditions. Different test methods are used due to different particle sizes of mineral aggregates. The bonding performance of asphalt and coarse aggregate is called adhesion, the bonding performance of asphalt and fine aggregate is called water stability, and the bonding performance of asphalt and filler is called hydrophilicity. The adhesion between asphalt and stone (referred to as asphalt adhesion) has an important impact on pavement performance. If the asphalt adhesion is poor, the road surface will be damaged, cracked, and stone splashed, affecting road safety and vehicle driving.

[0004] In order to detect the adhesion of asphalt, the boiling method can be used. The boiling method is to boil the mixture of asphalt and stone in water and judge the adhesion of asphalt by observing the shedding of asphalt. However, in the above method, it is necessary to manually observe the shedding of asphalt. The test results are affected by human factors and may not accurately give the asphalt adhesion, resulting in erroneous asphalt adhesion test results. Summary of the invention

[0005] The present application provides a method for determining an asphalt adhesion index, the method comprising:

[0006] Acquire target spectral data corresponding to a mixture, wherein the mixture is composed of asphalt and stone; wherein the target spectral data includes spectral data of all regions of the mixture, and the target spectral data includes K target spectral images, and different target spectral images correspond to different wavebands;

[0007] For each pixel point of the target spectrum image, determining the category of the pixel point based on K spectral intensity values ​​corresponding to the pixel point in the K target spectrum images, wherein the category is a stone category or a non-stone category;

[0008] Determine the index parameter corresponding to the mixed material based on the category of each pixel point; wherein the index parameter includes the number of pixel points of the non-stone category or the stone category, or the number of multiple three-dimensional points corresponding to the pixel points of the non-stone category or the stone category in the three-dimensional space, or the surface area of ​​the space composed of the multiple three-dimensional points;

[0009] The asphalt adhesion index corresponding to the mixture is determined based on the index parameters.

[0010] The present application provides a device for determining an asphalt adhesion index, the device comprising:

[0011] An acquisition module is used to acquire target spectral data corresponding to a mixed material, the mixed material being composed of asphalt and stone; wherein the target spectral data includes spectral data of all regions of the mixed material, and the target spectral data includes K target spectral images, and different target spectral images correspond to different wavebands;

[0012] A processing module, for determining, for each pixel point of the target spectrum image, a category of the pixel point based on K spectral intensity values ​​corresponding to the pixel point in the K target spectrum images, wherein the category is a stone category or a non-stone category;

[0013] A determination module is used to determine the index parameters corresponding to the mixed material based on the category of each pixel point; wherein the index parameters include the number of pixel points of non-stone category or stone category, or the number of multiple three-dimensional points corresponding to the pixel points of non-stone category or stone category in three-dimensional space, or the surface area of ​​the space composed of the multiple three-dimensional points; based on the index parameters, the asphalt adhesion index corresponding to the mixed material is determined.

[0014] The present application provides an electronic device, comprising: a processor and a machine-readable storage medium, wherein the machine-readable storage medium stores machine-executable instructions that can be executed by the processor; the processor is used to execute the machine-executable instructions to implement the above-mentioned example method for determining the asphalt adhesion index.

[0015] The present application provides a computer program product, which may include a computer program. When the computer program is executed by a processor, it implements the above-mentioned method for determining the asphalt adhesion index.

[0016] The present application provides a machine-readable storage medium, which stores machine-executable instructions that can be executed by a processor; wherein the processor is used to execute the machine-executable instructions to implement the above-mentioned example method for determining the asphalt adhesion index.

[0017] It can be seen from the above technical solutions that in the embodiment of the present application, the spectral data corresponding to the mixture (composed of asphalt and stone) can be used to determine the asphalt adhesion index corresponding to the mixture, rather than manually observing the shedding of asphalt to judge the asphalt adhesion, so as to realize automatic detection of the asphalt adhesion index based on spectral imaging technology, and obtain accurate asphalt adhesion index. Considering that the spectral data of stone has obvious differentiation, while the spectral data of asphalt has poor differentiation, the spectral data can be used to determine whether the pixel point belongs to the stone category, and the index parameter corresponding to the mixture is determined based on the category of each pixel point, and then the asphalt adhesion index corresponding to the mixture is determined, so as to further improve the detection accuracy of the asphalt adhesion index. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a flow chart of a method for determining an asphalt adhesion index in one embodiment of the present application;

[0019] Figure 2 is a physical schematic diagram of asphalt particles in one embodiment of the present application;

[0020] Figure 3 It is a structural schematic diagram of an asphalt adhesion index detection system in one embodiment of the present application;

[0021] Figure 4A It is a schematic diagram of the structure of a clamping tool in one embodiment of the present application;

[0022] Figure 4B is a schematic structural diagram of a clamp in one embodiment of the present application;

[0023] Figure 5 It is a structural schematic diagram of an asphalt adhesion index detection system in one embodiment of the present application;

[0024] Fig. 6A is a schematic diagram of a spectral image in one embodiment of the present application;

[0025] Figure 6B is a schematic diagram of K coordinate points of a spectrum curve in one embodiment of the present application;

[0026] Figure 6C is a schematic diagram of a spectrum curve in one embodiment of the present application;

[0027] Figure 7 is a schematic diagram of a target color image in one embodiment of the present application;

[0028] Figure 8 It is a structural schematic diagram of a device for determining asphalt adhesion index in one embodiment of the present application;

[0029] Fig. 9 It is a hardware structure diagram of an electronic device in one embodiment of the present application. DETAILED DESCRIPTION

[0030] In the embodiment of the present application, a method for determining the asphalt adhesion index is proposed, which can be applied to electronic equipment, see Figure 1 FIG. 4 is a flow chart of the method, which may include:

[0031] Step 101, obtaining target spectral data corresponding to a mixture, where the mixture is composed of asphalt and stone; wherein the target spectral data includes spectral data of all regions of the mixture, and the target spectral data includes K target spectral images, where K is a positive integer, and different target spectral images correspond to different bands.

[0032] Step 102: for each pixel of the target spectrum image, determine the category of the pixel based on K spectral intensity values ​​corresponding to the pixel in the K target spectrum images. The category may be a stone category or a non-stone category, and the non-stone category may be an asphalt category.

[0033] Step 103, determine the index parameter corresponding to the mixed material based on the category of each pixel point; wherein the index parameter may include but is not limited to: the number of pixel points of the non-stone category or the stone category, or, the number of multiple three-dimensional points corresponding to the pixel points of the non-stone category or the stone category in the three-dimensional space, or, the surface area of ​​the space composed of multiple three-dimensional points corresponding to the pixel points of the non-stone category or the stone category in the three-dimensional space.

[0034] Step 104: Determine the asphalt adhesion index corresponding to the mixture based on the index parameter.

[0035] Exemplarily, obtaining target spectral data corresponding to the mixed material may include, but is not limited to: for each region of the mixed material, when the region of the mixed material is within the field of view of the spectral camera, obtaining spectral data of the region collected by the spectral camera, the spectral data may include K spectral images; wherein all regions of the mixed material may be within the field of view of the spectral camera in sequence. Then, the spectral data of all regions of the mixed material may be spliced ​​to obtain the target spectral data; wherein, if there is an overlapping region between the first region and the second region of the mixed material, when splicing the spectral data of the first region and the second region, the spectral data of the overlapping region in the second region may be deleted.

[0036] Exemplarily, the field of view of the spectral camera includes a first clamping fixture and a second clamping fixture, and the first clamping fixture and the second clamping fixture are perpendicular to each other in the same plane. The two clamps of the first clamping fixture move toward each other, and the two clamps of the second clamping fixture move toward each other. Based on this, the first clamping fixture is used to fix the mixed material through two clamps, and the first clamping fixture is used to rotate the mixed material in the first direction at a configured first angle, so that all areas in the first direction are in the field of view of the spectral camera in sequence. The second clamping fixture is used to fix the mixed material through two clamps, and the second clamping fixture is used to rotate the mixed material in the second direction at a configured second angle, so that all areas in the second direction are in the field of view of the spectral camera in sequence. The second direction is perpendicular to the first direction in the same plane.

[0037] Exemplarily, determining the category of the pixel point based on the K spectral intensity values ​​corresponding to the pixel point in the K target spectral images may include, but is not limited to: determining the mapping relationship between the band and the spectral intensity value of the pixel point based on the K spectral intensity values ​​corresponding to the pixel point in the K target spectral images and the bands corresponding to the K target spectral images. Based on the mapping relationship between the band and the spectral intensity value, determining the target spectral feature corresponding to the pixel point; wherein the target spectral feature is used to reflect the characteristics of the light of the pixel point in different bands. Determine the category of the pixel point based on the target spectral feature.

[0038] For example, when determining the mapping relationship between the band and the spectral intensity value of the pixel point, a spectral curve corresponding to the pixel point can be generated based on the K spectral intensity values ​​corresponding to the pixel point in the K target spectral images and the bands corresponding to the K target spectral images. The horizontal axis of the spectral curve represents the K bands, and the vertical axis of the spectral curve represents the K spectral intensity values. On this basis, the target spectral feature corresponding to the pixel point can be determined based on the spectral curve, and there is no restriction on the method for determining the target spectral feature.

[0039] Exemplarily, for each pixel point of the target spectral image, the target spectral feature corresponding to the pixel point can be determined based on the K spectral intensity values ​​corresponding to the pixel point in the K target spectral images, that is, the target spectral feature is directly determined based on the K spectral intensity values, that is, the target spectral feature is used to reflect the spectral information of the K spectral intensity values, and there is no restriction on the method of determining this target spectral feature.

[0040] Exemplarily, determining the category of the pixel point based on the target spectral feature may include, but is not limited to: if the mixed material includes a component type of stone, then obtaining the target component type of the stone. Based on the stored sample spectral features of each component type, obtain the sample spectral features corresponding to the target component type; if the similarity between the target spectral feature and the sample spectral feature is greater than a threshold, determine the category of the pixel point as a stone category; if the similarity between the target spectral feature and the sample spectral feature is not greater than a threshold, determine the category of the pixel point as a non-stone category. Alternatively, based on the stored spectral classification model of each component type, obtain the spectral classification model corresponding to the target component type, the spectral classification model is used to identify the stone of the target component type; input the target spectral feature to the spectral classification model, and the spectral classification model determines the category of the pixel point as a stone category or a non-stone category based on the target spectral feature.

[0041] Exemplarily, determining the asphalt adhesion index corresponding to the mixture based on the index parameter may include but is not limited to: if the index parameter includes the number of pixels of the non-stone category or the stone category, then calculating the first ratio value between the number of pixels of the non-stone category or the stone category and the total number of pixels of the target spectral image; determining the asphalt coating rate corresponding to the mixture based on the first ratio value; determining the asphalt adhesion index corresponding to the mixture based on the asphalt coating rate. For example, if the asphalt adhesion index can include the asphalt coating rate, then the asphalt coating rate is directly used as the asphalt adhesion index. Alternatively, if the asphalt adhesion index can include the asphalt adhesion grade, then the asphalt adhesion grade corresponding to the mixture is determined based on the asphalt coating rate; wherein, the greater the asphalt coating rate, the greater the asphalt adhesion grade, and the better the stability of the mixture.

[0042] Exemplarily, determining the index parameter corresponding to the mixed material based on the category of each pixel point may include, but is not limited to: selecting a frame of target spectrum image from K target spectrum images of target spectrum data. Performing three-dimensional reconstruction based on the target spectrum image to obtain a target three-dimensional image, the target three-dimensional image may include multiple three-dimensional points; wherein, for each three-dimensional point of the target three-dimensional image, the category of the three-dimensional point is stone category or non-stone category; wherein, the category of the three-dimensional point corresponding to the pixel point of the non-stone category in the target three-dimensional image is non-stone category, and the category of the three-dimensional point corresponding to the pixel point of the stone category in the target three-dimensional image is stone category. Determine the index parameter corresponding to the mixed material based on the category of each three-dimensional point.

[0043] Exemplarily, the index parameters corresponding to the mixed material are determined based on the category of each 3D point, which may include but is not limited to: based on the category of each 3D point, a first category of 3D points of non-stone category or stone category is selected from the target 3D image, and a second category of 3D points in the 3D space is determined based on the first category of 3D points; a third category of 3D points in the 3D space is determined based on all 3D points of the target 3D image. For example, the first category of 3D points may be used as the second category of 3D points, and all 3D points of the target 3D image may be used as the third category of 3D points. Alternatively, a spatial interpolation algorithm may be used to interpolate the first category of 3D points to obtain the second category of 3D points, that is, the second category of 3D points includes the first category of 3D points and the interpolated 3D points. A spatial interpolation algorithm may be used to interpolate all 3D points of the target 3D image to obtain the third category of 3D points.

[0044] If the index parameter includes the number of multiple 3D points corresponding to non-stone or stone pixel points in the 3D space, the number of the second-category 3D points and the number of the third-category 3D points can be counted, on this basis, the number of the second-category 3D points can be the index parameter. Alternatively, if the index parameter includes the surface area of ​​the space composed of multiple 3D points corresponding to non-stone or stone pixel points in the 3D space, the first surface area of ​​the three-dimensional space composed of all the second-category 3D points is calculated, and the second surface area of ​​the three-dimensional space composed of all the third-category 3D points is calculated, on this basis, the first surface area can be the index parameter.

[0045] Exemplarily, determining the asphalt adhesion index corresponding to the mixture based on the index parameters may include but is not limited to: determining the asphalt coating rate corresponding to the mixture based on a second ratio value between the number of second-category three-dimensional points and the number of third-category three-dimensional points; determining the asphalt adhesion index corresponding to the mixture based on the asphalt coating rate; or determining the asphalt coating rate corresponding to the mixture based on a third ratio value between the first surface area and the second surface area; determining the asphalt adhesion index corresponding to the mixture based on the asphalt coating rate.

[0046] Exemplarily, performing three-dimensional reconstruction based on the target spectral image to obtain a target three-dimensional image may include, but is not limited to: obtaining a target RGB image corresponding to the mixed material; wherein, when each region of the mixed material is in the field of view of the RGB camera in turn, obtaining an RGB image of the region captured by the RGB camera, and splicing the RGB images of all regions of the mixed material to obtain the target RGB image. Performing three-dimensional reconstruction based on the target spectral image and the target RGB image to obtain a target three-dimensional image; wherein, the target three-dimensional image includes a first three-dimensional image corresponding to the target spectral image and / or a second three-dimensional image corresponding to the target RGB image. Alternatively, when each region of the mixed material is in the field of view of the RGB camera in turn, an RGB image of the region collected by the RGB camera is obtained; and, when each region of the mixed material is in the field of view of the spectral camera in turn, spectral data of the region collected by the spectral camera is obtained; for each region of the mixed material, a frame of spectral image is selected from K spectral images of the spectral data of the region, and the spectral image is used for splicing to obtain the target spectral image; three-dimensional reconstruction is performed based on the spectral image and the RGB image of the region to obtain a three-dimensional sub-image corresponding to the region; wherein the three-dimensional sub-image includes a first three-dimensional sub-image of the region corresponding to the spectral image and / or a second three-dimensional sub-image of the region corresponding to the RGB image; the three-dimensional sub-images corresponding to all regions are spliced ​​to obtain the target three-dimensional image.

[0047] Exemplarily, a target color image can also be generated based on the category of each pixel point and the target color image can be displayed; wherein the size of the target color image is consistent with the size of the target spectral image, or the size of the target color image is consistent with the size of the target three-dimensional image; wherein the target color image includes a first image area corresponding to the mixed material and a second image area outside the mixed material, and the pixel value of each pixel point in the second image area is a first color value; for each pixel point in the first image area, if the category of the pixel point is a stone category, the pixel value of the pixel point is a second color value; if the category of the pixel point is a non-stone category, the pixel value of the pixel point is a third color value.

[0048] It can be seen from the above technical solutions that in the embodiment of the present application, the spectral data corresponding to the mixed material (composed of asphalt and stone) can be used to determine the asphalt adhesion index corresponding to the mixed material, rather than manually observing the shedding of asphalt to judge the asphalt adhesion, so as to realize automatic detection of the asphalt adhesion index based on spectral imaging technology, and obtain accurate asphalt adhesion index. Considering that the spectral data of stone has obvious discrimination, while the spectral data of asphalt has poor discrimination, the spectral data can be used to determine whether the pixel point belongs to the stone category, so as to find all the pixel points belonging to the stone category, and then determine the asphalt adhesion index corresponding to the mixed material, and further improve the detection accuracy of the asphalt adhesion index.

[0049] The above technical solutions of the embodiments of the present application are described below in combination with specific application scenarios.

[0050] The adhesion of asphalt and stone (referred to as asphalt adhesion) has an important impact on road performance. If the asphalt adhesion is poor, the road surface will be damaged, cracked, and stone splashed, affecting road safety and vehicle driving. Ensuring that the adhesion of asphalt and stone meets the requirements is an important measure to ensure road durability and safety. In order to ensure the adhesion of asphalt and stone, it is necessary to be able to detect the adhesion of asphalt and stone.

[0051] For example, the water boiling method is to boil a mixture of asphalt and stone in water and observe the asphalt shedding to determine the asphalt adhesion. However, in the above method, the asphalt shedding needs to be observed manually, and the test results are affected by human factors, which may result in erroneous asphalt adhesion test results.

[0052] For example, you can collect RGB images of asphalt and analyze asphalt adhesion using RGB images. Figure 2 The figure below shows a real picture of asphalt particles. The asphalt surface is black, the shape of the asphalt particles is very irregular, and the asphalt particles are prone to highlight under the illumination of the light source. It is difficult to clearly image the asphalt, and the highlight area is prone to overexposure, making it difficult to distinguish the object category. Based on this, it may not be possible to accurately analyze the asphalt adhesion through RGB images, that is, incorrect asphalt adhesion detection results may be obtained.

[0053] In response to the above findings, a method for determining the asphalt adhesion index is proposed in an embodiment of the present application, which can use the spectral data corresponding to the mixed material (composed of asphalt and stone) to determine the asphalt adhesion index corresponding to the mixed material, thereby realizing automatic detection of the asphalt adhesion index based on spectral imaging technology.

[0054] In the embodiment of the present application, a system for detecting asphalt adhesion index is proposed. The system for detecting asphalt adhesion index may include an asphalt adhesion index detecting device (i.e., an electronic device), a carrier (i.e., a loading platform), and a clamping tool. Figure 3 As shown, it is a structural schematic diagram of the asphalt adhesion index detection system.

[0055] The asphalt adhesion index detection equipment may include an RGB camera, a spectral camera and a light source (taking two light sources as an example, the number of light sources may be more, or only one light source may be included).

[0056] RGB cameras are optional, that is, you can deploy RGB cameras and collect RGB images through RGB cameras, or you can not deploy RGB cameras. Based on the RGB images collected by RGB cameras, you can use RGB images to implement the preview function, and you can use RGB images and spectral images to form binocular stereo vision.

[0057] The light source can be a full-spectrum LED light source or other types of light sources. The light source ensures sufficient illumination in the field of view, such as illumination>1000lx. The light source ensures uniform illumination in the field of view, such as RI>95%.

[0058] The spectral camera can be a snapshot area array spectral camera, or a single-point spectral camera or a linear array spectral camera. There is no restriction on the type of spectral camera. A snapshot area array spectral camera is taken as an example. The imaging band range of the spectral camera can be 380nm-980nm, or it can be an imaging band range of any configuration. There is no restriction on this imaging band range. The spectral camera can use metasurfaces, filter arrays, single exposure compressed imaging, and FPI to achieve spectral detection. There is no restriction on this, as long as the imaging band range is met.

[0059] The stage (i.e., the stage) is used to place the clamping tooling. The clamping tooling is placed in the field of view of the spectral camera, and the clamping tooling is placed in the field of view of the RGB camera. Figure 3 As shown, the clamping fixture can be placed on the lower side of the spectral camera, and the clamping fixture can be placed on the lower side of the RGB camera. The clamping fixture is used to fix the asphalt and rotate the asphalt so that each area of ​​the asphalt is within the field of view of the spectral camera (RGB camera), and then the image of the area is collected by the spectral camera (RGB camera).

[0060] For example, the distance between the clamping fixture and the spectral camera (RGB camera) can be set based on experience, such as the distance between the clamping fixture and the spectral camera (RGB camera) is between 15cm and 20cm, and there is no limit on this distance. The distance between the clamping fixture and the spectral camera can be the distance between the center of the asphalt and the lens of the spectral camera, and the distance between the clamping fixture and the RGB camera can be the distance between the center of the asphalt and the lens of the RGB camera.

[0061] See also Figure 4AAs shown, it is a schematic diagram of the structure of the clamping fixture, and the clamping fixture may include two groups (four) of mechanical structural components (referred to as fixtures) and corresponding guide rails. For example, a first clamping fixture and a second clamping fixture may be deployed (i.e., the first clamping fixture and the second clamping fixture exist in the field of view of the spectral camera). The first clamping fixture may be composed of fixture A1 and fixture A2, and the second clamping fixture may be composed of fixture B1 and fixture B2. Alternatively, the first clamping fixture may be composed of fixture B1 and fixture B2, and the second clamping fixture may be composed of fixture A1 and fixture A2. In the subsequent process, the first clamping fixture is composed of fixture A1 and fixture A2, and the second clamping fixture is composed of fixture B1 and fixture B2.

[0062] The first clamping fixture and the second clamping fixture are perpendicular to each other in the same plane. Figure 4A It shows the top view of the first clamping fixture and the second clamping fixture, and the top view of the spectral camera (RGB camera) perpendicular to the paper surface.

[0063] The two fixtures (fixture A1 and fixture A2) of the first clamping fixture move toward each other, are installed on the same track, and have the same stroke. For example, when fixture A1 moves to the right by a stroke s1, fixture A2 moves to the left by a stroke s1 synchronously, that is, fixture A1 and fixture A2 move toward each other, and have the same stroke.

[0064] The two clamps (clamp B1 and clamp B2) of the second clamping fixture move toward each other, are installed on the same track, and have the same stroke. For example, when clamp B1 moves downward by a stroke s2, clamp B2 moves upward by a stroke s2 synchronously, that is, clamp B1 and clamp B2 move toward each other, and have the same stroke.

[0065] from Figure 4A It can be seen that the first clamping fixture can rotate the mixed material (composed of asphalt and stone) in a first direction, and the first direction can be a direction from bottom to top or from top to bottom, so that each area in the first direction is in the field of view of the spectral camera (RGB camera) in turn, and then the spectral camera (RGB camera) is used to collect images of each area in the first direction of the mixed material.

[0066] from Figure 4A It can be seen that the second clamping fixture can rotate the mixed material (composed of asphalt and stone) in the second direction, and the second direction can be from left to right or from right to left, so that each area in the second direction is in the field of view of the spectral camera (RGB camera) in turn, and then the spectral camera (RGB camera) is used to collect images of each area in the second direction of the mixed material.

[0067] See also Figure 4BAs shown, it is a schematic diagram of the structure of the clamp. For each clamp (such as clamp A1, clamp A2, clamp B1 and clamp B2), the clamp may include a motor, a rotating shaft and silicone. The motor is used to control the rotation of the rotating shaft, and the rotating shaft is a movable telescopic mechanism. When the rotating shaft is subjected to force, there is a buffer, which is more convenient for clamping asphalt particles. For example, clamp A1 and clamp A2 cooperate to control the rotating shaft to rotate the mixed material in the first direction, and clamp B1 and clamp B2 cooperate to control the rotating shaft to rotate the mixed material in the second direction.

[0068] The head of the clamp is milky white silicone, which does not stick to asphalt. In this way, when the mixed material (asphalt) is fixed by clamp A1 and clamp A2 and the asphalt is rotated, even if the asphalt is a high-viscosity organic liquid, the asphalt will not stick to clamp A1 and clamp A2, thus preventing the asphalt from sticking to the clamps.

[0069] In the above application scenario, a method for determining the asphalt adhesion index is proposed in the embodiment of the present application, see Figure 5 FIG. 1 is a flow chart of the method, which may include the following steps:

[0070] First, the data collection process: During the data collection process, the mixed material needs to be controlled to rotate, the spectral data is collected through the spectral camera, and the RGB image is collected through the RGB camera.

[0071] For example, before the data collection process, after the asphalt adhesion index detection device is powered on and the light source is stable, a whiteboard (i.e., a standard reflector) can be used for spectral calibration. For example, a spectral camera is used to image the standard reflector to collect spectral data, and based on the spectral data, it is confirmed whether the spectral response is normal. If normal, the data collection process can be executed. If abnormal, the data collection process is not executed, but the abnormality is repaired. After each device is turned on, the spectral calibration can be performed once.

[0072] Exemplarily, during the data collection process, the mixed material (which may be composed of asphalt and stone) may be placed on a stage between two clamps (such as clamp A1 and clamp A2) of the first clamping fixture, and then the mixed material may be fixed by the two clamps of the first clamping fixture (such as clamp A1 and clamp A2 move toward each other until clamp A1 and clamp A2 clamp the mixed material and the mixed material is fixed). After the mixed material is stabilized, the spectral camera may be used to collect spectral data 1 of area r11 of the mixed material (i.e., the area currently in the field of view of the spectral camera), and the RGB camera may be used to collect RGB images 1 of area r12 of the mixed material (i.e., the area currently in the field of view of the RGB camera). When the field of view of the spectral camera is the same as that of the RGB camera, area r11 and area r12 are the same.

[0073] Then, based on the configured first angle, the mixed material is rotated in the first direction by two clamps of the first clamping fixture. The first angle can be configured based on experience, such as 30 degrees, 60 degrees, 90 degrees, etc. In addition, the number of rotations in the first direction can be determined based on the first angle. For example, after rounding up 360 degrees divided by the first angle, the rounded value is subtracted by 1 to obtain the number of rotations in the first direction. Taking the first angle of 90 degrees as an example, the number of rotations in the first direction is 3 times. The reason for subtracting 1 here is that the spectral data and RGB image have been collected once before the mixed material is rotated.

[0074] For example, the mixed material is rotated in a first direction by two clamps of the first clamping tool, and the rotation angle of the mixed material is 90 degrees, such as rotating the mixed material 90 degrees in a direction from bottom to top.

[0075] After the mixed material is stabilized, spectral data 2 for region r21 of the mixed material may be collected by a spectral camera, and an RGB image 2 for region r22 of the mixed material may be collected by an RGB camera.

[0076] Then, the mixed material is rotated in the first direction by the two clamps of the first clamping fixture, and the rotation angle of the mixed material is 90 degrees. After the mixed material is stable, the spectral data 3 of the area r31 of the mixed material can be collected by the spectral camera, and the RGB image 3 of the area r32 of the mixed material can be collected by the RGB camera. And so on, until all the areas in the first direction are in the field of view of the spectral camera (RGB camera) in turn. For example, after the mixed material is rotated 3 times (the number of rotations is 3 times), all the areas in the first direction are in the field of view of the spectral camera (RGB camera) in turn.

[0077] So far, we can obtain spectral data 1, spectral data 2, spectral data 3 and spectral data 4, RGB image 1, RGB image 2, RGB image 3 and RGB image 4. Since the spectral data and RGB image are collected after the mixed material is stabilized, the spectral data and RGB image are both for static mixed materials.

[0078] After the rotation process of the first clamping tool is completed, the mixed material is rotated to the initial position, the first clamping tool is withdrawn, and the second clamping tool continues to rotate the mixed material, so as to image the previously blocked area. For example, the mixed material is fixed by two clamps of the second clamping tool (such as clamp B1 and clamp B2 move towards each other until clamp B1 and clamp B2 clamp the mixed material and the mixed material is fixed). Then, the mixed material is rotated to the previously blocked area by the second clamping tool.

[0079] After the mixed material is stabilized, spectral data 5 for region r51 of the mixed material may be collected by a spectral camera, and an RGB image 5 for region r52 of the mixed material may be collected by an RGB camera.

[0080] Then, based on the configured second angle, the mixed material is rotated in the second direction by the two clamps of the second clamping fixture. The second angle can be configured based on experience, such as 90 degrees, 180 degrees, etc. In addition, the number of rotations in the second direction can be determined based on the second angle. For example, 360 degrees is divided by the second angle and rounded up to get the number of rotations in the second direction. Taking the second angle of 180 degrees as an example, the number of rotations in the second direction is 2 times. For example, when the mixed material is rotated to the previously blocked area by the second clamping fixture, it has been rotated once, and only one rotation is needed to complete the rotation process.

[0081] For example, the mixed material is rotated in the second direction by two clamps of the second clamping tool, and the rotation angle of the mixed material is 180 degrees, such as rotating the mixed material 180 degrees from left to right.

[0082] After the mixed material is stabilized, spectral data 6 for region r61 of the mixed material may be collected by a spectral camera, and an RGB image 6 for region r62 of the mixed material may be collected by an RGB camera.

[0083] This is repeated until all regions in the second direction are in the field of view of the spectral camera (RGB camera) in sequence. For example, after the mixed material is rotated twice (i.e., the number of rotations is 2 times), all regions in the second direction are in the field of view of the spectral camera (RGB camera) in sequence.

[0084] At this point, spectral data 5 and spectral data 6, RGB image 5 and RGB image 6 can be obtained.

[0085] In summary, during the data collection process, multiple regions of the mixed material (such as all regions in the first direction and all regions in the second direction, and the second direction and the first direction may be perpendicular to each other) may be controlled to be in the field of view of the spectral camera (RGB camera) in sequence. For each region of the mixed material, when the region of the mixed material is in the field of view of the spectral camera (RGB camera), the spectral data of the region is collected by the spectral camera, and the RGB image of the region is collected by the RGB camera.

[0086] Second, the image stitching process. In the image stitching process, the spectral data of all regions of the mixed material can be stitched to obtain the target spectral data corresponding to the mixed material. For the convenience of distinction, the data before stitching is called spectral data, and the data after stitching is called target spectral data. The RGB images of all regions of the mixed material can be stitched to obtain the target RGB image corresponding to the mixed material. For the convenience of distinction, the data before stitching is called RGB image, and the data after stitching is called target RGB image.

[0087] For example, spectral data 1, spectral data 2, spectral data 3, spectral data 4, spectral data 5 and spectral data 6 can be spliced ​​to obtain target spectral data, and the target spectral data can include spectral data of all regions of the mixed material. RGB image 1, RGB image 2, RGB image 3, RGB image 4, RGB image 5 and RGB image 6 can be spliced ​​to obtain a target RGB image, and the target RGB image can include RGB images of all regions of the mixed material.

[0088] Exemplarily, when the spectral data of all regions of the mixed material are spliced ​​to obtain the target spectral data corresponding to the mixed material, if there is an overlapping area between the first region and the second region of the mixed material, when the spectral data of the first region and the second region are spliced, the spectral data of the overlapping area in the second region can be deleted, or the spectral data of the overlapping area in the first region can be deleted.

[0089] For example, when there is an overlapping area between two regions (there may be multiple groups), the two regions are referred to as the first region and the second region, the spectral data of the first region may be any spectral data, and the spectral data of the second region may be any spectral data. For example, if there is an overlapping area between spectral data 1 and spectral data 2, such as spectral data 1 includes spectral data of region S1 and spectral data of region S2, and spectral data 2 includes spectral data of region S2 and spectral data of region S3, then region S2 is an overlapping area, and the spectral data of region S2 in spectral data 2 is deleted, so that the spliced ​​data includes spectral data of region S1 and spectral data of region S2 of spectral data 1, and spectral data of region S3 of spectral data 2.

[0090] The present embodiment does not limit how to obtain the overlapping area of ​​the first area and the second area. For example, since the spectral data of the overlapping area are consistent, if the spectral data of a sub-area of ​​the first area is consistent with the spectral data of a sub-area of ​​the second area, then this sub-area is the overlapping area of ​​the first area and the second area, and when splicing the spectral data, the duplicate spectral data can be deleted.

[0091] It should be noted that since the spectral data of different regions are not completely consistent, but the spectral data of the same region are completely consistent, when the spectral data of a sub-region of the first region is consistent with the spectral data of a sub-region of the second region, it means that this sub-region is the overlapping region of the first region and the second region.

[0092] Exemplarily, when the RGB images of all areas of the mixed material are spliced ​​to obtain a target RGB image corresponding to the mixed material, if there is an overlapping area between the third area and the fourth area of ​​the mixed material, when the RGB images of the third area and the fourth area are spliced, the RGB image of the overlapping area in the fourth area can be deleted, or the RGB image of the overlapping area in the third area can be deleted.

[0093] For example, when there is an overlapping area between two regions (there may be multiple groups), the two regions are referred to as the third region and the fourth region, the RGB image of the third region may be any RGB image, and the RGB image of the fourth region may be any RGB image. For example, if there is an overlapping area between RGB image 1 and RGB image 2, such as RGB image 1 includes a sub-image of region S1 and a sub-image of region S2, and RGB image 2 includes a sub-image of region S2 and a sub-image of region S3, then region S2 is an overlapping area, and the sub-image of region S2 in RGB image 2 is deleted. In this way, the spliced ​​data may include the sub-image of region S1 and the sub-image of region S2 of RGB image 1, and the sub-image of region S3 of RGB image 2.

[0094] This embodiment does not limit how to obtain the overlapping area of ​​the third area and the fourth area. For example, since the pixel values ​​of the overlapping area are consistent, if the pixel value of a sub-area of ​​the third area is consistent with the pixel value of a sub-area of ​​the fourth area, then this sub-area is the overlapping area.

[0095] Third, three-dimensional reconstruction process. In the three-dimensional reconstruction process, the target spectral data may include K target spectral images (for details, see the subsequent embodiments), and a frame of the target spectral image may be selected from the K target spectral images of the target spectral data to reconstruct a first three-dimensional image corresponding to the target spectral image, and reconstruct a second three-dimensional image corresponding to the target RGB image.

[0096] In a possible implementation, all RGB images may be first stitched together to obtain a target RGB image (image stitching process), and then three-dimensional reconstruction may be performed based on the target RGB image and the target spectral image to obtain a first three-dimensional image corresponding to the target spectral image and a second three-dimensional image corresponding to the target RGB image.

[0097] Exemplarily, in the three-dimensional reconstruction process, the target spectrum data may include K target spectrum images, and a frame of the target spectrum image may be selected from the K target spectrum images, such as randomly selecting a frame of the target spectrum image, which is recorded as the target spectrum image X. On this basis, three-dimensional reconstruction may be performed based on the target spectrum image X and the target RGB image to obtain a first three-dimensional image corresponding to the target spectrum image X and a second three-dimensional image corresponding to the target RGB image. For example, based on the target spectrum image X and the target RGB image, a binocular vision algorithm may be used for three-dimensional reconstruction, and no restriction is imposed on the binocular vision algorithm. For example, the pixel point 1 corresponding to the position point in the three-dimensional space in the target spectrum image X is found, and the pixel point 2 corresponding to the position point in the target RGB image is found. Based on the pixel point 1 and the pixel point 2, a triangulation algorithm is used to obtain the depth information corresponding to the position point, and the depth information is combined with the pixel point 1 to obtain the three-dimensional point 1 corresponding to the target spectrum image X, and the depth information is combined with the pixel point 2 to obtain the three-dimensional point 2 corresponding to the target RGB image. The three-dimensional points corresponding to the target spectrum image X constitute a first three-dimensional image, that is, the first three-dimensional image includes a plurality of three-dimensional points, and the plurality of three-dimensional points of the first three-dimensional image correspond one-to-one to the plurality of pixel points of the target spectrum image X. The three-dimensional points corresponding to the target RGB image constitute a second three-dimensional image, that is, the second three-dimensional image includes a plurality of three-dimensional points, and the plurality of three-dimensional points of the second three-dimensional image correspond one-to-one to the plurality of pixel points of the target RGB image.

[0098] Exemplarily, in order to achieve three-dimensional reconstruction of the target spectral image X and obtain a first three-dimensional image corresponding to the target spectral image X, the RGB camera can be replaced by the spectral camera 1 (i.e., two spectral cameras are deployed), and three-dimensional reconstruction can be performed based on the spectral image collected by the spectral camera 1 and the target spectral image X to obtain a first three-dimensional image, such as using a binocular vision algorithm for three-dimensional reconstruction. Alternatively, the RGB camera can be replaced by a laser for laser line scanning (i.e., a laser and a spectral camera are deployed), and the laser line is scanned by the laser, and three-dimensional reconstruction is performed in combination with the target spectral image X to obtain a first three-dimensional image. Alternatively, the RGB camera can be replaced by a laser for structured light scanning (i.e., a laser and a spectral camera are deployed), and the structured light is scanned by the laser, and three-dimensional reconstruction is performed in combination with the target spectral image X to obtain a first three-dimensional image. Of course, the above are just a few examples of three-dimensional reconstruction, and there is no limitation to this.

[0099] In a possible implementation, three-dimensional reconstruction may be performed based on the RGB image and the spectral image to obtain a first three-dimensional sub-image corresponding to the spectral image and a second three-dimensional sub-image corresponding to the RGB image. Then, image stitching may be performed on all first three-dimensional sub-images corresponding to the spectral images to obtain a first three-dimensional image corresponding to the target spectral image, and image stitching may be performed on all second three-dimensional sub-images corresponding to the RGB images to obtain a second three-dimensional image corresponding to the target RGB image. This process is described below.

[0100] After collecting spectral data 1 for a certain area of ​​the mixed material through a spectral camera and collecting an RGB image 1 of the area through an RGB camera, the spectral data 1 includes K spectral images. A frame of spectral image can be selected from the K spectral images, such as randomly selecting a frame of spectral image, such as selecting a spectral image with a wavelength band of 380nm from the spectral data 1, and three-dimensional reconstruction is performed based on the spectral image and the RGB image 1 to obtain a first three-dimensional sub-image 1 corresponding to the spectral image and a second three-dimensional sub-image 1 corresponding to the RGB image 1.

[0101] After collecting spectral data 2 for a certain area of ​​the mixed material through a spectral camera and collecting an RGB image 2 of the area through an RGB camera, a spectral image with a wavelength band of 380nm is selected from the spectral data 2, and then three-dimensional reconstruction is performed based on the spectral image and the RGB image 2 to obtain a first three-dimensional sub-image 2 corresponding to the spectral image and a second three-dimensional sub-image 2 corresponding to the RGB image 2.

[0102] By analogy, the first three-dimensional sub-image and the second three-dimensional sub-image of each area of ​​the mixed material can be obtained, and then the first three-dimensional sub-images of all areas can be spliced ​​to obtain the first three-dimensional image corresponding to the target spectral image. During the splicing process, duplicate data can be removed. The splicing process is similar to the splicing process of the spectral data and will not be repeated here. In addition, the second three-dimensional sub-images of all areas can be spliced ​​to obtain the second three-dimensional image corresponding to the target RGB image. During the splicing process, duplicate data can be removed. The splicing process is similar to the splicing process of the RGB image and will not be repeated here.

[0103] Fourth, three-dimensional preview process. In the three-dimensional preview process, after obtaining the second three-dimensional image corresponding to the target RGB image, the second three-dimensional image can be displayed, so that the user can preview the second three-dimensional image of the mixed material, and the second three-dimensional image can fully display the three-dimensional surface data of the mixed material, and the second three-dimensional image can be an RGB image, which can clearly display the three-dimensional morphology of the mixed material.

[0104] Fifth, the classification process of the mixed material. In the classification process of the mixed material, the category of each pixel point of the target spectral image can be determined. Since the mixed material is composed of asphalt and stone, the category of the pixel point can be the stone category or the non-stone category (the non-stone category can be the asphalt category).

[0105] Exemplarily, during the classification process of the mixed material, target spectral data corresponding to the mixed material may be obtained. Referring to the image stitching process, the spectral data has been stitched to obtain the target spectral data.

[0106] The target spectrum data may include K target spectrum images (the target spectrum image is obtained by splicing the spectrum images of all regions), where K is a positive integer, and different target spectrum images correspond to different bands. For example, assuming that the imaging band range of the spectral camera is 380nm-980nm, and a target spectrum image is collected every 10nm band, then the target spectrum data includes target spectrum image 1 of band 380nm, target spectrum image 2 of band 390nm, target spectrum image 3 of band 400nm, and so on.

[0107] For each target spectrum image, the target spectrum image may include multiple pixels, and the pixel value of each pixel may represent a spectrum intensity value, also referred to as spectrum intensity, where spectrum intensity represents the physical quantity of spectrum line energy. Different target spectrum images have the same size, such as the same width and the same height.

[0108] For each pixel point (x, y) of the target spectrum image 1, the pixel value of the pixel point (x, y) represents the spectral intensity value at the band 380nm. For each pixel point (x, y) of the target spectrum image 2, the pixel value of the pixel point (x, y) represents the spectral intensity value at the band 390nm, and so on.

[0109] After obtaining K target spectrum images of the target spectrum data, for each pixel point of the target spectrum image, determine the K spectral intensity values ​​corresponding to the pixel point in the K target spectrum images. For example, for the pixel point (x, y), determine the spectral intensity value of the pixel point (x, y) in the target spectrum image 1 at the 380nm band, determine the spectral intensity value of the pixel point (x, y) in the target spectrum image 2 at the 390nm band, determine the spectral intensity value of the pixel point (x, y) in the target spectrum image 3 at the 400nm band, and so on, to obtain the K spectral intensity values ​​corresponding to the pixel point (x, y).

[0110] Exemplarily, for each pixel of the target spectrum image, based on the K spectral intensity values ​​corresponding to the pixel in the K target spectrum images, the category of the pixel can be determined, which can be a stone category or a non-stone category (i.e., an asphalt category). For example, based on the K spectral intensity values ​​corresponding to the pixel in the K target spectrum images and the bands corresponding to the K target spectrum images, the mapping relationship between the band and the spectral intensity value of the pixel can be determined. Then, based on the mapping relationship between the band and the spectral intensity value, the target spectrum feature corresponding to the pixel is determined, and the target spectrum feature is used to reflect the characteristics of the light of the pixel in different bands. Alternatively, based on the K spectral intensity values ​​corresponding to the pixel in the K target spectrum images, the target spectrum feature corresponding to the pixel can be directly determined, and the target spectrum feature is used to reflect the spectral information of the K spectral intensity values. Of course, the above two examples of determining the target spectrum feature are given, and the method of determining the target spectrum feature is not limited, as long as it is related to the K spectral intensity values.

[0111] For example, for each pixel of the target spectrum image, when determining the mapping relationship between the band and the spectral intensity value of the pixel, a spectral curve corresponding to the pixel can be generated based on the K spectral intensity values ​​corresponding to the pixel in the K target spectrum images and the bands corresponding to the K target spectrum images. The horizontal axis of the spectral curve represents the K bands, and the vertical axis of the spectral curve represents the K spectral intensity values. On this basis, the target spectrum feature corresponding to the pixel can be determined based on the spectral curve.

[0112] For example, see Fig. 6A As shown, it is a schematic diagram of a target spectrum image, and the target spectrum image may include a plurality of pixel points. Fig. 6A 11 pixels are shown in the figure (i.e., pixels numbered 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, and 11, where pixel 1, pixel 2, and pixel 3 represent pixels of gravel / stone, and other pixels are not limited). It is necessary to obtain the spectral curves corresponding to these 11 pixels. Here, taking 11 pixels as an example, it is necessary to obtain the spectral curve corresponding to each pixel of the target spectral image.

[0113] For example, taking the pixel point (x, y) as an example, a coordinate system can be constructed. The horizontal coordinate is horizontal to the right and the vertical coordinate is vertical upward. The horizontal coordinate represents the band, and the coordinate system represents the spectral intensity value. The horizontal coordinate of the first coordinate point is 380nm, and the vertical coordinate of the first coordinate point is the spectral intensity value corresponding to the pixel point (x, y) in the target spectrum image 1. The horizontal coordinate of the second coordinate point is 390nm, and the vertical coordinate of the second coordinate point is the spectral intensity value corresponding to the pixel point (x, y) in the target spectrum image 2. The horizontal coordinate of the third coordinate point is 400nm, and the vertical coordinate of the third coordinate point is the spectral intensity value corresponding to the pixel point (x, y) in the target spectrum image 3. And so on, the horizontal coordinate of the Kth coordinate point is 980nm, and the vertical coordinate of the Kth coordinate point is the spectral intensity value corresponding to the pixel point (x, y) in the target spectrum image K. At this point, K coordinate points can be created in the coordinate system, see Figure 6B As shown in the figure, it is a schematic diagram of the K coordinate points of the spectrum curve. On this basis, the curve composed of these K coordinate points is the spectrum curve corresponding to the pixel point (x, y), see Figure 6C The diagram shown is a schematic diagram of the spectral curve.

[0114] In summary, for each pixel point of the target spectral image, a spectral curve corresponding to the pixel point can be generated, the abscissa of the spectral curve represents K bands, and the ordinate represents K spectral intensity values.

[0115] Exemplarily, for each pixel of the target spectrum image, after obtaining the spectrum curve corresponding to the pixel, the target spectrum feature corresponding to the pixel can be determined based on the spectrum curve.

[0116] For example, see Figure 6B As shown, the spectrum curve may include K arrays, that is, it exists in the form of arrays. For each array, the array may include a band value and a spectral intensity value, such as the first array corresponding to the first coordinate point, including a band value of 380nm and a spectral intensity value corresponding to the band value, and so on.

[0117] On this basis, the spectral curve (K arrays) can be feature extracted to obtain the target spectral features corresponding to the pixel point. For example, the spectral curve (K arrays) can be input into a feature extractor, and the feature extractor can extract the features of the spectral curve to obtain the target spectral features. For another example, an array analysis algorithm can be used to analyze the spectral curve (K arrays) to obtain the target spectral features. Of course, the above is just an example, as long as the target spectral features of the K arrays can be extracted.

[0118] For example, see Figure 6CAs shown, the spectrum curve can also be a spectrum curve graph, that is, it exists in the form of a graph, and the spectrum curve graph includes a spectrum curve composed of K coordinate points. On this basis, the spectrum curve graph can be feature extracted to obtain the target spectrum feature corresponding to the pixel point. For example, the spectrum curve graph can be input into a feature extractor, and the feature extractor can extract the feature of the spectrum curve graph to obtain the target spectrum feature. For another example, the spectrum curve graph can be analyzed by an image analysis algorithm to obtain the target spectrum feature. Of course, the above is just an example, and it is sufficient to be able to extract the target spectrum feature of the spectrum curve graph.

[0119] Exemplarily, for each pixel point of the target spectral image, after obtaining the target spectral feature corresponding to the pixel point, the category of the pixel point can be determined based on the target spectral feature. The category can be a stone category (used to indicate that the pixel point belongs to stone) or a non-stone category (used to indicate that the pixel point does not belong to stone. Since the mixed material is composed of asphalt and stone, the pixel point that does not belong to stone belongs to asphalt).

[0120] For example, the category of each pixel point of the target spectral image may be determined in the following manner.

[0121] Method 1: It is necessary to store the sample spectral features of each component type in advance. On this basis, the category of the pixel point can be determined based on the sample spectral features and the target spectral features corresponding to the pixel point.

[0122] For example, considering that only a few types of stones (such as basalt, tuff, limestone, etc.) are mixed with asphalt to obtain materials, it is necessary to obtain the sample spectral features of these component types in advance and store the sample spectral features of each component type.

[0123] For example, a sample material is obtained by mixing a basalt-type stone with asphalt, and the spectral characteristics corresponding to the sample material are obtained (see the method for obtaining the target spectral characteristics), that is, the spectral characteristics of each pixel of the spectral image. It is known which pixels of the spectral image belong to the basalt-type stone (which can be calibrated by the user), so the spectral characteristics of the pixel can be selected from the spectral characteristics, and the spectral characteristics of the pixel are used as the sample spectral characteristics t1 corresponding to the basalt component type. Similarly, the sample spectral characteristics t2 corresponding to the tuff component type and the sample spectral characteristics t3 corresponding to the limestone component type can be obtained, as shown in Table 1, and the sample spectral characteristics of each component type can be pre-stored.

[0124] Table 1

[0125]

[0126] In summary, the spectral data of stones of different composition types (such as sample spectral characteristics) can be pre-stored. Each time the sample spectral characteristics are tested, corresponding options can be provided through the operation interface so that the user can select the composition type of the stone. After obtaining the sample spectral characteristics, the correspondence between the composition type and the sample spectral characteristics can be stored, thereby improving the efficiency and accuracy of algorithm recognition.

[0127] When the component types of the stone are increased, corresponding options can be provided through the operation interface for the newly added component types so that the user can select the newly added component types. Then, the sample spectral characteristics of the component type can be obtained by the above method and added to Table 1.

[0128] Exemplarily, the category of each pixel point of the target spectral image can be determined in the following manner: if the mixed material includes only one type of stone, the target component type of the stone is obtained. For example, when mixing asphalt with stone, usually only one type of stone is mixed with asphalt, and the target component type of the stone can be obtained, such as basalt component type, or tuff component type, or limestone component type. On this basis, based on the stored sample spectral features of each component type, the sample spectral features corresponding to the target component type can be obtained. For example, if the target component type is a basalt component type, by querying Table 1, the sample spectral feature corresponding to the target component type is the sample spectral feature t1.

[0129] For each pixel of the target spectral image, the target spectral feature corresponding to the pixel is known, and the similarity between the target spectral feature and the sample spectral feature is calculated, such as Euclidean distance, cosine similarity, etc.

[0130] If the similarity between the target spectral feature and the sample spectral feature is greater than a threshold value (which can be configured based on experience, such as 80%, 90%, etc.), it means that the component at the location of the pixel is close to the target component type, so the category of the pixel is determined to be stone. If the similarity between the target spectral feature and the sample spectral feature is not greater than the threshold value, it means that the component at the location of the pixel is significantly different from the target component type, so the category of the pixel is determined to be non-stone, that is, the pixel corresponds to asphalt.

[0131] In method 1, the sample spectral features corresponding to several types of stone components (such as basalt, tuff, limestone, etc.) can be pre-stored, and then the target spectral features can be compared with the sample spectral features to distinguish whether the pixel point is a stone category, instead of storing the sample spectral features corresponding to asphalt to distinguish whether the pixel point is an asphalt category. The reason is that asphalt particles usually refer to stones wrapped with asphalt, and stones have different component types, such as basalt, tuff, limestone, etc. It is found in actual measurements that the spectral data of stone and asphalt have obvious distinction, and the spectral data of stones of different component types also have obvious distinction, while the spectral data of asphalt has poor distinction, and it is impossible to accurately distinguish whether the pixel point is an asphalt category based on the spectral data of asphalt. Therefore, the spectral data can be used to determine whether the pixel point belongs to the stone category, so as to find all the pixels belonging to the stone category.

[0132] Method 2: It is necessary to pre-train and store the spectral classification model of each component type. On this basis, the category of the pixel point can be determined based on the spectral classification model and the target spectral features corresponding to the pixel point.

[0133] For example, considering that only a few types of stone components are used to mix with asphalt to obtain materials, it is necessary to pre-train spectral classification models of these component types and store the spectral classification models.

[0134] For example, a sample material is obtained by mixing a basalt-type stone with asphalt, and the spectral features corresponding to the sample material are obtained, that is, the spectral features of each pixel of the spectral image. It is known which pixels of the spectral image (recorded as pixel w1) belong to the basalt-type stone, and it is known which pixels of the spectral image (recorded as pixel w2) do not belong to the basalt-type stone. Therefore, the spectral features of pixel w1 can be selected from the spectral features as a positive sample, and the spectral features of pixel w2 can be selected from the spectral features as a negative sample. Then, the spectral features of pixel w1 are input into the initial classification model to obtain the predicted category of pixel w1, and the spectral features of pixel w2 are input into the initial classification model to obtain the predicted category of pixel w2. Based on the difference between the predicted category of pixel w1 and the true category of pixel w1 (such as 1, indicating a pixel belonging to stone), and the difference between the predicted category of pixel w2 and the true category of pixel w2 (such as 0, indicating a pixel not belonging to stone), the loss value can be calculated, and then the initial classification model is trained based on the loss value to obtain a trained spectral classification model, and the spectral classification model is stored. This spectral classification model corresponds to the basalt component type. Similarly, the spectral classification model corresponding to the tuff component type and the spectral classification model corresponding to the limestone component type can be obtained.

[0135] When the component types of the stone are increased, for the newly added component types, the spectral classification model corresponding to the component types can be obtained by the above method, and the spectral classification model can be stored.

[0136] Exemplarily, the category of each pixel point of the target spectral image can be determined in the following manner: if the mixed material includes only one type of stone, the target component type of the stone is obtained. On this basis, based on the stored spectral classification model of each component type, the spectral classification model corresponding to the target component type is obtained, and the spectral classification model is used to identify the stone of the target component type. For example, the spectral classification model corresponding to the basalt component type can identify whether the pixel point belongs to a stone of the basalt component type.

[0137] For each pixel point of the target spectral image, the target spectral feature corresponding to the pixel point is known, and the target spectral feature can be input into the spectral classification model, and the spectral classification model determines the category of the pixel point as a stone category or a non-stone category based on the target spectral feature. For example, the spectral classification model performs a classification operation based on the target spectral feature to obtain a predicted category, and outputs the predicted category. If the predicted category is the first value (such as 1), it means that the category of the pixel point is a stone category. If the predicted category is the second value (such as 0), it means that the category of the pixel point is a non-stone category (such as asphalt).

[0138] With respect to the spectral classification model, the spectral classification model may be a SVM (Support Vector Machine) classifier or a model based on a deep learning algorithm, and there is no restriction on this.

[0139] In method 2, the spectral classification models corresponding to several component types of stone (such as basalt, tuff, limestone, etc.) can be pre-stored, and then the target spectral features can be input into the spectral classification model corresponding to the target component type to distinguish whether the category of the pixel point is the stone category, instead of storing the spectral classification model corresponding to asphalt to distinguish whether the category of the pixel point is the asphalt category.

[0140] Sixth, the process of determining the asphalt adhesion index. In the process of determining the asphalt adhesion index, the index parameter corresponding to the mixture can be determined based on the category of each pixel point. Then, the asphalt adhesion index corresponding to the mixture is determined based on the index parameter. For example, the process can be carried out in the following manner:

[0141] Method 1: Based on the category of each pixel point, the number of pixels of the non-stone category or the stone category is determined, that is, the index parameter includes the number of pixels of the non-stone category or the stone category. Then, the first ratio value between the number of pixels of the non-stone category or the stone category and the total number of pixels of the target spectral image is calculated; the asphalt coating rate corresponding to the mixture is determined based on the first ratio value; and the asphalt adhesion index corresponding to the mixture is determined based on the asphalt coating rate.

[0142] For example, if the first ratio value between the number of pixels in the non-stone category and the total number of pixels is calculated, the first ratio value is used as the asphalt wrapping rate. If the first ratio value between the number of pixels in the stone category and the total number of pixels is calculated, the difference between the fixed value (such as 1) and the first ratio value is used as the asphalt wrapping rate. Obviously, the asphalt wrapping rate can be determined based on the pixels in the non-stone category, and the asphalt wrapping rate can be determined based on the pixels in the stone category. When determining the asphalt wrapping rate based on the pixels in the non-stone category, there is no need to perform the difference operation between the fixed value and the first ratio value, thereby saving the amount of calculation.

[0143] For example, the asphalt adhesion index may include an asphalt coating rate, which indicates how much of the surface of the mixture is asphalt. Obviously, the greater the asphalt coating rate, the better the asphalt coating effect of the mixture. Based on this, the first ratio value between the number of pixels of the non-stone category and the total number of pixels of the target spectral image can be used as the asphalt coating rate corresponding to the mixture, and the asphalt coating rate can be used as an asphalt adhesion index. For example, assuming that the total number of pixels of the target spectral image is 10,000, the number of pixels of the stone category is 500, and the number of pixels of the non-stone category is 9,500, the first ratio value can be 95%, and therefore, the asphalt coating rate corresponding to the mixture can be 95%.

[0144] For example, the asphalt adhesion index may include an asphalt adhesion grade (asphalt adhesion level). Based on this, after obtaining the asphalt coating rate corresponding to the mixture, the asphalt adhesion grade corresponding to the mixture may also be determined based on the asphalt coating rate, and the asphalt adhesion grade may be used as an asphalt adhesion index. The greater the asphalt coating rate, the greater the asphalt adhesion grade, and the better the stability of the mixture, and the better the asphalt coating effect of the mixture. For example, if the asphalt coating rate is greater than 99%, that is, the asphalt coating rate is close to 100%, the asphalt adhesion grade may be level 5, which is the highest level of asphalt adhesion grade. If the asphalt coating rate is greater than 90% and not greater than 99%, the asphalt adhesion grade may be level 4. If the asphalt coating rate is greater than 70% and not greater than 90%, the asphalt adhesion grade may be level 3. If the asphalt coating rate is greater than 30% and not greater than 70%, the asphalt adhesion grade may be level 2. If the asphalt coating rate is not greater than 30%, the asphalt adhesion grade may be level 1, which is the lowest level of asphalt adhesion grade.

[0145] Method 2: Based on the category of each pixel point, the asphalt adhesion index can be determined using the following steps.

[0146] Step S11: selecting a frame of target spectrum image from the K target spectrum images of the target spectrum data, such as randomly selecting a frame of target spectrum image, which is subsequently recorded as target spectrum image X.

[0147] Step S12: Perform three-dimensional reconstruction based on the target spectral image X to obtain a target three-dimensional image. Referring to the three-dimensional reconstruction process, the target three-dimensional image may be a first three-dimensional image corresponding to the target spectral image X, may be a second three-dimensional image corresponding to the target RGB image, or may be a first three-dimensional image and a second three-dimensional image.

[0148] Exemplarily, the target 3D image may include a plurality of 3D points, and for each 3D point of the target 3D image, the category of the 3D point may be a stone category or a non-stone category. For example, the category of the 3D point corresponding to the pixel point of the non-stone category in the target 3D image may be a non-stone category, and the category of the 3D point corresponding to the pixel point of the stone category in the target 3D image may be a stone category.

[0149] For example, if the target three-dimensional image is the first three-dimensional image corresponding to the target spectrum image X, the first three-dimensional image includes multiple three-dimensional points, the target spectrum image X includes multiple pixel points, and the multiple three-dimensional points of the first three-dimensional image correspond one to one with the multiple pixel points of the target spectrum image X. Referring to the classification process of the mixed material, the category of each pixel point of the target spectrum image X has been known. For each three-dimensional point of the first three-dimensional image, the pixel point corresponding to the three-dimensional point is found from the target spectrum image X, and the category of this pixel point is the category of the three-dimensional point. If the category of the pixel point is the stone category, the category of the three-dimensional point is the stone category, and if the category of the pixel point is the non-stone category, the category of the three-dimensional point is the non-stone category.

[0150] For example, if the target 3D image is the second 3D image corresponding to the target RGB image, the second 3D image includes multiple 3D points, the target RGB image includes multiple pixel points, and the multiple 3D points of the second 3D image correspond to the multiple pixel points of the target RGB image one by one. For each 3D point of the second 3D image, find the pixel point corresponding to the 3D point from the target RGB image, and the category of this pixel point is the category of the 3D point. If the category of the pixel point is the stone category, the category of the 3D point is the stone category, and if the category of the pixel point is the non-stone category, the category of the 3D point is the non-stone category.

[0151] Referring to the classification process of the mixed material, the category of each pixel point of the target spectral image X has been known. In addition, multiple pixels of the target RGB image correspond one-to-one to multiple pixels of the target spectral image X. Therefore, for each pixel point of the target RGB image (denoted as pixel point P1), the pixel point corresponding to pixel point P1 (denoted as pixel point P2) can be found from the target spectral image X, and the category of pixel point P2 is the category of pixel point P1. If the category of pixel point P2 is the stone category, the category of pixel point P1 is the stone category. If the category of pixel point P2 is the non-stone category, the category of pixel point P1 is the non-stone category.

[0152] In summary, a target three-dimensional image can be obtained, and the target three-dimensional image can include multiple three-dimensional points. For each three-dimensional point of the target three-dimensional image, the category of the three-dimensional point can be known.

[0153] Step S13, based on the category of each three-dimensional point in the target three-dimensional image, determine the number of multiple three-dimensional points corresponding to the pixel points of the non-stone category or the stone category in the three-dimensional space, that is, determine the number of multiple three-dimensional points of the non-stone category or the stone category, that is, the index parameters corresponding to the mixed material include the number of three-dimensional points of the non-stone category or the stone category.

[0154] Exemplarily, based on the category of each 3D point in the target 3D image, first-category 3D points of non-stone category or stone category can be selected from the target 3D image, that is, 3D points of non-stone category or stone category can be selected as first-category 3D points.

[0155] Determine the second type of three-dimensional points in the three-dimensional space based on the first type of three-dimensional points. For example, the first type of three-dimensional points can be used as the second type of three-dimensional points. Alternatively, the first type of three-dimensional points can be interpolated using a spatial interpolation algorithm to obtain the second type of three-dimensional points, that is, the second type of three-dimensional points include the first type of three-dimensional points and the three-dimensional points obtained by interpolation. Regarding the spatial interpolation algorithm, it is used to convert the measurement data of discrete points into a continuous data surface for comparison with the distribution pattern of other spatial phenomena. It can include a spatial interpolation algorithm and a spatial extrapolation algorithm. The spatial interpolation algorithm is to infer the unknown point data in the same area through the data of known points. The spatial extrapolation algorithm is to infer the data of other areas through the data of known areas. In step S13, all the first type of three-dimensional points are used as input data of the spatial interpolation algorithm, and then the spatial interpolation algorithm is used to interpolate these first type of three-dimensional points, and the second type of three-dimensional points are used as output data of the spatial interpolation algorithm. There is no restriction on this process.

[0156] After determining the second type of three-dimensional points in the three-dimensional space, the number of the second type of three-dimensional points may be counted.

[0157] Step S14: determining a third type of three-dimensional points in the three-dimensional space based on all three-dimensional points of the target three-dimensional image.

[0158] For example, all 3D points of the target 3D image can be used as the third type of 3D points. Alternatively, a spatial interpolation algorithm can be used to interpolate all 3D points of the target 3D image to obtain the third type of 3D points, that is, the third type of 3D points can include all 3D points of the target 3D image and the interpolated 3D points.

[0159] After determining the third type of 3D points in the 3D space, the number of the third type of 3D points may be counted.

[0160] Step S15, calculate the second ratio value between the number of second-category three-dimensional points and the number of third-category three-dimensional points, determine the asphalt coating ratio corresponding to the mixture based on the second ratio value, and determine the asphalt adhesion index corresponding to the mixture based on the asphalt coating ratio. Step S15 can refer to method 1 and will not be repeated here.

[0161] For example, if a 3D point of a non-stone type is regarded as a first-class 3D point, the second ratio value is regarded as the asphalt coating ratio corresponding to the mixed material. If a 3D point of a stone type is regarded as a first-class 3D point, the difference between a fixed value (such as 1) and the second ratio value is regarded as the asphalt coating ratio.

[0162] Method 3: Based on the category of each pixel point, the asphalt adhesion index can be determined using the following steps.

[0163] Step S21 : selecting a frame of target spectrum image from the K target spectrum images of the target spectrum data, such as randomly selecting a frame of target spectrum image, which is subsequently recorded as target spectrum image X.

[0164] Step S22: Perform three-dimensional reconstruction based on the target spectral image X to obtain a target three-dimensional image.

[0165] Step S23: based on the category of each 3D point in the target 3D image, select a first category of 3D points of non-stone category or stone category from the target 3D image, and determine a second category of 3D points in the 3D space based on the first category of 3D points.

[0166] Step S24: Determine a third type of three-dimensional point in the three-dimensional space based on all three-dimensional points of the target three-dimensional image.

[0167] Step S25, calculate the first surface area of ​​the three-dimensional space composed of all the second-category three-dimensional points, and calculate the second surface area of ​​the three-dimensional space composed of all the third-category three-dimensional points. For example, the index parameter corresponding to the mixed material includes the surface area of ​​the space composed of multiple three-dimensional points corresponding to the pixel points of the non-stone category or the stone category in the three-dimensional space, that is, the first surface area of ​​the three-dimensional space composed of all the second-category three-dimensional points is used as the index parameter.

[0168] For example, after obtaining the second type of three-dimensional points, these second type of three-dimensional points constitute a three-dimensional space, and the first surface area of ​​the three-dimensional space composed of all second type three-dimensional points can be calculated, that is, the three-dimensional space is divided into small faces based on these second type three-dimensional points, and the sum of the areas of these faces is taken as the first surface area.

[0169] For example, after obtaining the third type of three-dimensional points, these third type of three-dimensional points form a three-dimensional space, and the second surface area of ​​the three-dimensional space composed of all the third type of three-dimensional points can be calculated.

[0170] Step S26, calculate the third ratio value between the first surface area and the second surface area, determine the asphalt coating rate corresponding to the mixture based on the third ratio value, and determine the asphalt adhesion index corresponding to the mixture based on the asphalt coating rate. Step S26 can refer to method 1 and will not be repeated here.

[0171] For example, if a 3D point of non-stone type is regarded as the first type of 3D point, the third ratio value is regarded as the asphalt coating rate corresponding to the mixed material. If a 3D point of stone type is regarded as the first type of 3D point, the difference between the fixed value (such as 1) and the third ratio value is regarded as the asphalt coating rate.

[0172] Seventh, the display process of the asphalt adhesion index. In the display process of the asphalt adhesion index, a target color image can be generated based on the category of each pixel point, and the target color image can be displayed.

[0173] Exemplarily, the size of the target color image is consistent with the size of the target spectral image (i.e., any spectral image of the target spectral data), and the size of the target color image is consistent with the size of the target RGB image, such as the width size is consistent and the height size is consistent, that is, the target color image can be a two-dimensional image.

[0174] Alternatively, the size of the target color image is consistent with the size of the target three-dimensional image, such as the size of the target color image is consistent with the size of the first three-dimensional image corresponding to the target spectral image, or the size of the target color image is consistent with the size of the second three-dimensional image corresponding to the target RGB image, such as the width size is consistent, the height size is consistent, and the depth value is consistent, that is, the target color image can be a three-dimensional image.

[0175] Exemplarily, the target color image includes a first image area corresponding to the mixed material and a second image area outside the mixed material. For example, the first image area may include pixel points corresponding to the mixed material, and the second image area may include pixel points corresponding to the non-mixed material.

[0176] Exemplarily, the pixel value of each pixel in the second image area may be a first color value, and the first color value may be any color value. For each pixel in the first image area, if the category of the pixel is a stone category, the pixel value of the pixel may be a second color value, and the second color value may be any color value, as long as it is different from the first color value. For each pixel in the first image area, if the category of the pixel is a non-stone category, the pixel value of the pixel may be a third color value, and the third color value may be any color value, as long as it is different from the first color value and the second color value.

[0177] See also Figure 7 As shown, it is a schematic diagram of the target color image. Here, a two-dimensional image is taken as an example. The target color image of a three-dimensional image is similar and will not be described here. The target color image represents the categories of stone and asphalt. Black represents the first color value, that is, the black pixel points belong to the second image area. Green represents the second color value, that is, the green pixel points belong to the pixel points of the stone category in the first image area. Red represents the third color value, that is, the red pixel points belong to the pixel points of the non-stone category in the first image area.

[0178] It can be seen from the above technical scheme that in the embodiment of the present application, the spectral data corresponding to the mixture (composed of asphalt and stone) can be used to determine the asphalt adhesion index corresponding to the mixture, instead of manually observing the asphalt shedding to judge the asphalt adhesion, so as to realize the automatic detection of the asphalt adhesion index based on the spectral imaging technology, and obtain an accurate asphalt adhesion index. Considering that the spectral data of the stone has obvious discrimination, while the spectral data of the asphalt has poor discrimination, the spectral data can be used to determine whether the pixel point belongs to the stone category, so as to find all the pixels belonging to the stone category, and then determine the asphalt adhesion index corresponding to the mixture, and further improve the detection accuracy of the asphalt adhesion index. By adopting a snapshot hyperspectral camera, the spatial image information and spectral information of the mixture can be obtained at the same time, and the information of different substances can be effectively extracted by spectral imaging, and the algorithm can be automatically classified, and the asphalt and stone can be distinguished by spectral imaging, and the asphalt wrapping rate can be automatically counted, so as to obtain the asphalt adhesion index. By adopting the above method, the asphalt adhesion index of a single asphalt particle can be obtained within 50 seconds on average.

[0179] Based on the same application concept as the above method, a device for determining asphalt adhesion index is proposed in the embodiment of the present application, see Figure 8 FIG. 1 is a schematic diagram of the structure of the device, wherein the device comprises:

[0180] An acquisition module 81 is used to acquire target spectral data corresponding to a mixture, where the mixture consists of asphalt and stone; the target spectral data includes spectral data of all regions of the mixture, and the target spectral data includes K target spectral images, where different target spectral images correspond to different bands; a processing module 82 is used to determine the category of each pixel point in the target spectral image based on K spectral intensity values ​​corresponding to the pixel point in the K target spectral images, where the category is a stone category or a non-stone category; a determination module 83 is used to determine an index parameter corresponding to the mixture based on the category of each pixel point; the index parameter includes the number of pixels of the non-stone category or the stone category, or the number of multiple three-dimensional points corresponding to pixels of the non-stone category or the stone category in a three-dimensional space, or the surface area of ​​a space composed of the multiple three-dimensional points; an asphalt adhesion index corresponding to the mixture is determined based on the index parameter.

[0181] Exemplarily, when the acquisition module 81 acquires the target spectrum data corresponding to the mixed material, it is specifically used to:

[0182] For each region of the mixed material, when the region of the mixed material is within the field of view of the spectral camera, spectral data of the region collected by the spectral camera is acquired, the spectral data including K spectral images; wherein all regions of the mixed material are sequentially within the field of view of the spectral camera;

[0183] The spectral data of all regions of the mixed material are spliced ​​to obtain the target spectral data; wherein, if there is an overlapping area between the first region and the second region of the mixed material, when the spectral data of the first region and the second region are spliced, the spectral data of the overlapping area in the second region is deleted.

[0184] Exemplarily, a first clamping fixture and a second clamping fixture exist within the field of view of the spectral camera, the first clamping fixture and the second clamping fixture are perpendicular to each other in the same plane, the two clamps of the first clamping fixture move toward each other, and the two clamps of the second clamping fixture move toward each other;

[0185] The first clamping fixture is used to fix the mixed material through two clamps, and the first clamping fixture is used to rotate the mixed material in a first direction at a configured first angle so that all areas in the first direction are in the field of view of the spectral camera in sequence;

[0186] The second clamping fixture is used to fix the mixed material through two clamps, and the second clamping fixture is used to rotate the mixed material in the second direction at a configured second angle so that all areas in the second direction are in the field of view of the spectral camera in sequence;

[0187] The second direction is perpendicular to the first direction.

[0188] Exemplarily, when determining the category of the pixel point based on the K spectral intensity values ​​corresponding to the pixel point in K target spectral images, the processing module 82 is specifically used to: determine the mapping relationship between the band and the spectral intensity value of the pixel point based on the K spectral intensity values ​​corresponding to the pixel point in the K target spectral images and the bands corresponding to the K target spectral images; determine the target spectral feature corresponding to the pixel point based on the mapping relationship between the band and the spectral intensity value; wherein the target spectral feature is used to reflect the characteristics of light of the pixel point in different bands; and determine the category of the pixel point based on the target spectral feature.

[0189] Exemplarily, when the processing module 82 determines the category of the pixel point based on the target spectral feature, it is specifically used to: if the mixed material includes a stone of a component type, obtain the target component type of the stone; based on the stored sample spectral features of each component type, obtain the sample spectral features corresponding to the target component type; if the similarity between the target spectral feature and the sample spectral feature is greater than a threshold, determine the category of the pixel point as a stone category; if the similarity between the target spectral feature and the sample spectral feature is not greater than the threshold, determine the category of the pixel point as a non-stone category;

[0190] Alternatively, based on the stored spectral classification model of each component type, a spectral classification model corresponding to the target component type is obtained, and the spectral classification model is used to identify the stone of the target component type;

[0191] The target spectral feature is input into the spectral classification model, and the spectral classification model determines the category of the pixel point as a stone category or a non-stone category based on the target spectral feature.

[0192] Exemplarily, when the determination module 83 determines the asphalt adhesion index corresponding to the mixture based on the index parameter, it is specifically used to: if the index parameter includes the number of pixel points of the non-stone category or the stone category, then calculate the first ratio value between the number of pixel points of the non-stone category or the stone category and the total number of pixel points of the target spectral image; determine the asphalt encapsulation rate corresponding to the mixture based on the first ratio value; determine the asphalt adhesion index corresponding to the mixture based on the asphalt encapsulation rate.

[0193] Exemplarily, when the determination module 83 determines the index parameters corresponding to the mixed material based on the category of each pixel point, it is specifically used to: select a frame of target spectral image from K target spectral images of target spectral data; perform three-dimensional reconstruction based on the target spectral image to obtain a target three-dimensional image, wherein the target three-dimensional image includes multiple three-dimensional points; for each three-dimensional point of the target three-dimensional image, the category of the three-dimensional point is stone category or non-stone category; the category of the three-dimensional point corresponding to the pixel point of the non-stone category in the target three-dimensional image is non-stone category, and the category of the three-dimensional point corresponding to the pixel point of the stone category in the target three-dimensional image is stone category; determine the index parameters corresponding to the mixed material based on the category of each three-dimensional point.

[0194] Exemplarily, when the determination module 83 determines the index parameter corresponding to the mixed material based on the category of each three-dimensional point, it is specifically used to: based on the category of each three-dimensional point, select a first category of three-dimensional points of non-stone category or stone category from the target three-dimensional image, and determine a second category of three-dimensional points in the three-dimensional space based on the first category of three-dimensional points; determine a third category of three-dimensional points in the three-dimensional space based on all three-dimensional points of the target three-dimensional image; if the index parameter includes the number of multiple three-dimensional points corresponding to pixel points of non-stone category or stone category in the three-dimensional space, then count the number of second category three-dimensional points and the number of third category three-dimensional points; or, if the index parameter includes the surface area of ​​the space composed of multiple three-dimensional points corresponding to pixel points of non-stone category or stone category in the three-dimensional space, then calculate the first surface area of ​​the three-dimensional space composed of all second category three-dimensional points, and calculate the second surface area of ​​the three-dimensional space composed of all third category three-dimensional points. When the determination module 83 determines the asphalt adhesion index corresponding to the mixture based on the index parameters, it is specifically used to: determine the asphalt wrapping rate corresponding to the mixture based on the second ratio value between the number of second-category three-dimensional points and the number of third-category three-dimensional points; determine the asphalt adhesion index corresponding to the mixture based on the asphalt wrapping rate; or, determine the asphalt wrapping rate corresponding to the mixture based on the third ratio value between the first surface area and the second surface area; determine the asphalt adhesion index corresponding to the mixture based on the asphalt wrapping rate.

[0195] Exemplarily, the acquisition module 81 is also used to acquire a target RGB image corresponding to the mixed material; wherein, when each region of the mixed material is in the field of view of the RGB camera in turn, the RGB image of the region captured by the RGB camera is acquired, and the RGB images of all regions of the mixed material are spliced ​​to obtain a target RGB image; or, when each region of the mixed material is in the field of view of the RGB camera in turn, the RGB image of the region captured by the RGB camera is acquired; and, when each region of the mixed material is in the field of view of the spectral camera in turn, the spectral data of the region captured by the spectral camera is acquired; when the determination module 83 performs three-dimensional reconstruction based on the target spectral image to obtain a target three-dimensional image, it is specifically used to: The target three-dimensional image is obtained by performing three-dimensional reconstruction based on the target spectral image and the target RGB image; wherein the target three-dimensional image includes a first three-dimensional image corresponding to the target spectral image and / or a second three-dimensional image corresponding to the target RGB image; or, for each area of ​​the mixed material, a frame of spectral image is selected from K spectral images of spectral data of the area, and the spectral image is used for splicing to obtain the target spectral image; three-dimensional reconstruction is performed based on the spectral image and the RGB image of the area to obtain a three-dimensional sub-image corresponding to the area; the three-dimensional sub-image includes a first three-dimensional sub-image of the area corresponding to the spectral image and / or a second three-dimensional sub-image of the area corresponding to the RGB image; the three-dimensional sub-images corresponding to all areas are spliced ​​to obtain the target three-dimensional image.

[0196] Exemplarily, the processing module 82 is also used to generate a target color image based on the category of each pixel point and display the target color image; the size of the target color image is consistent with the size of the target spectral image, or the size of the target color image is consistent with the size of the target three-dimensional image; the target color image includes a first image area corresponding to the mixed material and a second image area outside the mixed material, and the pixel value of each pixel point in the second image area is a first color value; for each pixel point in the first image area, if the category of the pixel point is a stone category, the pixel value of the pixel point is a second color value; if the category of the pixel point is a non-stone category, the pixel value of the pixel point is a third color value.

[0197] Based on the same application concept as the above method, an electronic device is proposed in the embodiment of the present application, see Fig. 9 As shown, the electronic device includes: a processor 91 and a machine-readable storage medium 92, the machine-readable storage medium 92 stores machine-executable instructions that can be executed by the processor 91; the processor 91 is used to execute the machine-executable instructions to implement the method for determining the asphalt adhesion index disclosed in the above example of this application.

[0198] Based on the same application concept as the above method, an embodiment of the present application also provides a machine-readable storage medium, on which a number of computer instructions are stored. When the computer instructions are executed by a processor, the method for determining the asphalt adhesion index disclosed in the above example of the present application can be implemented.

[0199] The above-mentioned machine-readable storage medium may be any electronic, magnetic, optical or other physical storage device, which may contain or store information, such as executable instructions, data, etc. For example, the machine-readable storage medium may be: RAM (Radom Access Memory), volatile memory, non-volatile memory, flash memory, storage drive (such as hard disk drive), solid state drive, any type of storage disk (such as optical disk, DVD, etc.), or similar storage medium, or a combination thereof.

[0200] Based on the same application concept as the above method, an embodiment of the present application also provides a computer program product, which may include a computer program. When the computer program is executed by a processor, it can implement the method for determining the asphalt adhesion index disclosed in the above example of the present application.

[0201] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the embodiments of the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0202] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A method for determining asphalt adhesion index, characterized in that: The method comprises: Acquire target spectral data corresponding to a mixture, wherein the mixture is composed of asphalt and stone; wherein the target spectral data includes spectral data of all regions of the mixture, and the target spectral data includes K target spectral images, and different target spectral images correspond to different wavebands; For each pixel point of the target spectral image, the category of the pixel point is determined based on the K spectral intensity values ​​corresponding to the pixel point in the K target spectral images, and the category is a stone category or a non-stone category; wherein, based on the stored sample spectral features of each stone component type, the sample spectral features corresponding to the target component type of the stone in the mixed material are obtained; if the similarity between the target spectral feature of the pixel point and the sample spectral feature is greater than a threshold value, the category of the pixel point is determined to be a stone category, and the target spectral feature is determined based on the K spectral intensity values; otherwise, the category of the pixel point is determined to be a non-stone category; or, based on the stored spectral classification model of each stone component type, a spectral classification model corresponding to the target component type is obtained; the target spectral feature is input to the spectral classification model, and the spectral classification model determines the category of the pixel point as a stone category or a non-stone category based on the target spectral feature; Determine the index parameter corresponding to the mixed material based on the category of each pixel point; wherein the index parameter includes the number of pixel points of the non-stone category or the stone category, or the number of multiple three-dimensional points corresponding to the pixel points of the non-stone category or the stone category in the three-dimensional space, or the surface area of ​​the space composed of the multiple three-dimensional points; The asphalt adhesion index corresponding to the mixture is determined based on the index parameters.

2. The method according to claim 1, characterized in that The step of obtaining target spectral data corresponding to the mixed material includes: For each region of the mixed material, when the region of the mixed material is within the field of view of the spectral camera, spectral data of the region collected by the spectral camera is acquired, the spectral data including K spectral images; wherein all regions of the mixed material are sequentially within the field of view of the spectral camera; The spectral data of all regions of the mixed material are spliced ​​to obtain the target spectral data; wherein, if there is an overlapping area between the first region and the second region of the mixed material, when the spectral data of the first region and the second region are spliced, the spectral data of the overlapping area in the second region is deleted.

3. The method according to claim 2, characterized in that A first clamping fixture and a second clamping fixture exist within the field of view of the spectral camera, the first clamping fixture and the second clamping fixture are perpendicular to each other in the same plane, the two clamps of the first clamping fixture move toward each other, and the two clamps of the second clamping fixture move toward each other; The first clamping fixture is used to fix the mixed material through two clamps, and the first clamping fixture is used to rotate the mixed material in a first direction at a configured first angle so that all areas in the first direction are in the field of view of the spectral camera in sequence; The second clamping fixture is used to fix the mixed material through two clamps, and the second clamping fixture is used to rotate the mixed material in the second direction at a configured second angle so that all areas in the second direction are in the field of view of the spectral camera in sequence; The second direction is perpendicular to the first direction.

4. The method according to claim 1, characterized in that: The process of determining the target spectral feature of the pixel point based on the K spectral intensity values ​​includes: Based on K spectral intensity values ​​corresponding to the pixel point in the K target spectral images and the wavebands corresponding to the K target spectral images, determining a mapping relationship between the waveband of the pixel point and the spectral intensity value; Based on the mapping relationship between the waveband and the spectral intensity value, the target spectral feature corresponding to the pixel point is determined; wherein the target spectral feature is used to reflect the characteristics of the light of the pixel point in different wavebands.

5. The method according to claim 1, characterized in that The determining of the asphalt adhesion index corresponding to the mixture based on the index parameter includes: If the index parameter includes the number of pixels of the non-stone category or the stone category, then a first ratio value between the number of pixels of the non-stone category or the stone category and the total number of pixels of the target spectral image is calculated; Determining an asphalt coating ratio corresponding to the mixture based on the first ratio value; The asphalt adhesion index corresponding to the mixture is determined based on the asphalt coating rate.

6. The method according to claim 1, characterized in that The determining of the index parameter corresponding to the mixed material based on the category of each pixel point includes: Selecting a frame of target spectrum image from the K target spectrum images of the target spectrum data; A target three-dimensional image is obtained by performing three-dimensional reconstruction based on the target spectral image, wherein the target three-dimensional image includes a plurality of three-dimensional points; wherein, for each three-dimensional point of the target three-dimensional image, the category of the three-dimensional point is a stone category or a non-stone category; wherein, the category of the three-dimensional point corresponding to the pixel point of the non-stone category in the target three-dimensional image is a non-stone category, and the category of the three-dimensional point corresponding to the pixel point of the stone category in the target three-dimensional image is a stone category; The index parameter corresponding to the mixed material is determined based on the category of each three-dimensional point.

7. The method according to claim 6, characterized in that The step of determining the index parameter corresponding to the mixed material based on the category of each three-dimensional point includes: Based on the category of each 3D point, a first category of 3D points of non-stone category or stone category are selected from the target 3D image, and a second category of 3D points in the 3D space are determined based on the first category of 3D points; and a third category of 3D points in the 3D space are determined based on all 3D points of the target 3D image; If the index parameter includes the number of multiple three-dimensional points corresponding to the pixel points of the non-stone category or the stone category in the three-dimensional space, then the number of the second category of three-dimensional points and the number of the third category of three-dimensional points are counted; Alternatively, if the index parameter includes the surface area of ​​a space composed of a plurality of three-dimensional points corresponding to the pixel points of the non-stone category or the stone category in the three-dimensional space, then the first surface area of ​​the three-dimensional space composed of all the second-category three-dimensional points is calculated, and the second surface area of ​​the three-dimensional space composed of all the third-category three-dimensional points is calculated; The determining of the asphalt adhesion index corresponding to the mixture based on the index parameter includes: Based on the second ratio value between the number of second-category three-dimensional points and the number of third-category three-dimensional points, determine the asphalt coating ratio corresponding to the mixture; based on the asphalt coating ratio, determine the asphalt adhesion index corresponding to the mixture; or, based on the third ratio value between the first surface area and the second surface area, determine the asphalt coating ratio corresponding to the mixture; based on the asphalt coating ratio, determine the asphalt adhesion index corresponding to the mixture.

8. The method according to claim 6, characterized in that The three-dimensional reconstruction based on the target spectral image to obtain the target three-dimensional image includes: Acquire a target RGB image corresponding to the mixed material; wherein, when each area of ​​the mixed material is in the field of view of the RGB camera in turn, acquire an RGB image of the area captured by the RGB camera, and splice the RGB images of all areas of the mixed material to obtain a target RGB image; The target three-dimensional image is obtained by performing three-dimensional reconstruction based on the target spectrum image and the target RGB image; wherein the target three-dimensional image includes a first three-dimensional image corresponding to the target spectrum image and / or a second three-dimensional image corresponding to the target RGB image; or, When each region of the mixed material is in the field of view of the RGB camera in turn, an RGB image of the region collected by the RGB camera is obtained; and when each region of the mixed material is in the field of view of the spectral camera in turn, spectral data of the region collected by the spectral camera is obtained; For each region of the mixed material, a frame of spectral image is selected from K spectral images of spectral data of the region, and the spectral image is used for splicing to obtain the target spectral image; Performing three-dimensional reconstruction based on the spectral image and the RGB image of the region to obtain a three-dimensional sub-image corresponding to the region; wherein the three-dimensional sub-image includes a first three-dimensional sub-image of the region corresponding to the spectral image and / or a second three-dimensional sub-image of the region corresponding to the RGB image; The target three-dimensional image is obtained by splicing the three-dimensional sub-images corresponding to all the regions.

9. The method according to claim 6, characterized in that The method further comprises: Generate a target color image based on the category of each pixel, and display the target color image; Among them, the size of the target color image is consistent with the size of the target spectral image, or the size of the target color image is consistent with the size of the target three-dimensional image; the target color image includes a first image area corresponding to the mixed material and a second image area outside the mixed material, and the pixel value of each pixel in the second image area is a first color value; for each pixel in the first image area, if the category of the pixel is a stone category, the pixel value of the pixel is a second color value; if the category of the pixel is a non-stone category, the pixel value of the pixel is a third color value.

10. A device for determining asphalt adhesion index, characterized in that: The device comprises: An acquisition module is used to acquire target spectral data corresponding to a mixed material, the mixed material being composed of asphalt and stone; wherein the target spectral data includes spectral data of all regions of the mixed material, and the target spectral data includes K target spectral images, and different target spectral images correspond to different wavebands; A processing module is used to determine the category of each pixel point of the target spectral image based on K spectral intensity values ​​corresponding to the pixel point in K target spectral images, and the category is a stone category or a non-stone category; wherein, when the processing module determines the category of the pixel point based on the K spectral intensity values ​​corresponding to the pixel point in the K target spectral images, it is specifically used to: obtain the sample spectral feature corresponding to the target component type of the stone in the mixed material based on the stored sample spectral feature of each stone component type; if the similarity between the target spectral feature of the pixel point and the sample spectral feature is greater than a threshold, the category of the pixel point is determined to be a stone category, and the target spectral feature is determined based on the K spectral intensity values; otherwise, the category of the pixel point is determined to be a non-stone category; or, based on the stored spectral classification model of each stone component type, obtain the spectral classification model corresponding to the target component type; input the target spectral feature to the spectral classification model, and the spectral classification model determines the category of the pixel point as a stone category or a non-stone category based on the target spectral feature; A determination module is used to determine the index parameters corresponding to the mixed material based on the category of each pixel point; wherein the index parameters include the number of pixel points of non-stone category or stone category, or the number of multiple three-dimensional points corresponding to the pixel points of non-stone category or stone category in three-dimensional space, or the surface area of ​​the space composed of the multiple three-dimensional points; based on the index parameters, the asphalt adhesion index corresponding to the mixed material is determined.

11. An electronic device, characterized in that: include: a processor and a machine-readable storage medium storing machine-executable instructions executable by the processor; The processor is used to execute machine executable instructions to implement the method described in any one of claims 1-9.

12. A computer program product, comprising a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 9 is implemented.

Images