WO2023179011A1

WO2023179011A1 - Image generation method and device

Info

Publication number: WO2023179011A1
Application number: PCT/CN2022/127355
Authority: WO
Inventors: 杜林鹏
Original assignee: 杭州睿影科技有限公司
Priority date: 2022-03-25
Filing date: 2022-10-25
Publication date: 2023-09-28
Also published as: CN114677454A; CN114677454B

Abstract

An image generation method and device, relating to the technical field of image processing. The method comprises: obtaining a three-dimensional point cloud image; determining statistical characteristics corresponding to the amplitude of a three-dimensional point in a neighborhood of a preset three-dimensional point in the three-dimensional point cloud image, and determining the image characteristics of the preset three-dimensional point according to the statistical characteristics, wherein the amplitude of the three-dimensional point is used for representing the electromagnetic scattering characteristics at a corresponding position of the three-dimensional point, and the neighborhood of the preset three-dimensional point comprises a preset point cloud area of the preset three-dimensional point; according to the image characteristics of the preset three-dimensional point in a preset linear direction, determining the projection characteristics of the preset three-dimensional point on a preset projection plane in the preset linear direction perpendicular to the preset projection plane; and determining a two-dimensional projection image of the three-dimensional point cloud image on the preset projection plane according to the projection characteristics on the preset projection plane. Therefore, a two-dimensional image corresponding to the three-dimensional point cloud image can be generated, and the generated two-dimensional image can adapt to different application scenes, so that the application effect of the image is improved.

Description

An image generation method and device

This application claims priority to the Chinese patent application with application number 202210305072.1 and the invention title is "an image generation method and device" submitted to the China Patent Office on March 25, 2022, the entire content of which is incorporated into this application by reference.

Technical field

The present application relates to the field of image processing technology, and in particular to an image generation method and device.

Background technique

In related technologies, three-dimensional scanning equipment such as lidar, millimeter-wave radar, or microwave radar can be used to obtain information on several points in space, including coordinates and corresponding amplitudes, to obtain a three-dimensional point cloud map. Among them, the amplitude of each three-dimensional point in the three-dimensional point cloud image is determined based on the electromagnetic scattering characteristics at the corresponding position of the three-dimensional point in space.

However, the data volume of the 3D point cloud image is large, and directly performing business processing (such as target detection, target recognition, etc.) on the 3D point cloud image is highly complex and costly. Therefore, a method is urgently needed to generate a two-dimensional image corresponding to the three-dimensional point cloud image, and then business processing can be performed based on the two-dimensional image.

Contents of the invention

In a first aspect, an embodiment of the present application discloses an image generation method, which method includes:

Obtain a three-dimensional point cloud image;

Determine the statistical characteristics corresponding to the amplitude of the three-dimensional point in the neighborhood of the preset three-dimensional point in the three-dimensional point cloud image, and determine the image characteristics of the preset three-dimensional point based on the statistical characteristics; wherein, the three-dimensional point The amplitude of is used to characterize the electromagnetic scattering characteristics at the corresponding position of the three-dimensional point, and the neighborhood of the preset three-dimensional point refers to the preset point cloud area including the preset three-dimensional point;

According to the image characteristics of the preset three-dimensional point in the preset straight line direction, the projection characteristics of the preset three-dimensional point in the preset straight line direction on the preset projection plane are determined; wherein the preset straight line direction is the same as the preset straight line direction. The default projection plane is vertical;

According to the projection characteristics on the preset projection plane, a two-dimensional projection image of the three-dimensional point cloud image on the preset projection plane is determined.

In a second aspect, an embodiment of the present application discloses an image generation device, which includes:

The three-dimensional point cloud image acquisition module is used to obtain the three-dimensional point cloud image;

The image feature acquisition module is used to determine the statistical features corresponding to the amplitudes of the three-dimensional points in the neighborhood of the preset three-dimensional points in the three-dimensional point cloud image, and determine the image features of the preset three-dimensional points based on the statistical features. ; Wherein, the amplitude of the three-dimensional point is used to characterize the electromagnetic scattering characteristics at the corresponding position of the three-dimensional point, and the neighborhood of the preset three-dimensional point refers to the preset point cloud area including the preset three-dimensional point;

The projection feature acquisition module is used to determine the projection features of the preset three-dimensional point in the preset straight line direction on the preset projection plane according to the image features of the preset three-dimensional point in the preset straight line direction; wherein, the The preset straight line direction is perpendicular to the preset projection plane;

A two-dimensional projection image acquisition module is used to determine the two-dimensional projection image of the three-dimensional point cloud image on the preset projection plane according to the projection characteristics on the preset projection plane.

In another aspect of the application, an electronic device is also disclosed. The electronic device includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory pass through The above communication bus completes mutual communication;

The memory is used to store computer programs;

The processor is configured to implement any of the above image generation methods when executing a program stored on the memory.

In yet another aspect of the implementation of the present application, a non-transitory computer-readable storage medium is also provided. A computer program is stored in the non-transitory computer-readable storage medium. When the computer program is executed by a processor, the following is implemented: Any of the image generation methods described above.

In yet another aspect of the implementation of this application, a computer program product containing instructions is also provided, which, when run on a computer, causes the computer to execute any one of the above image generation methods.

Beneficial effects of the embodiments of this application:

The image generation method provided by the embodiment of the present application obtains a three-dimensional point cloud image; determines the statistical characteristics corresponding to the amplitudes of the three-dimensional points in the neighborhood of the preset three-dimensional points in the three-dimensional point cloud image, and determines the amplitude of the preset three-dimensional point based on the statistical characteristics. Image features; among them, the amplitude of the three-dimensional point is used to characterize the electromagnetic scattering characteristics at the corresponding position of the three-dimensional point, and the neighborhood of the preset three-dimensional point refers to the preset point cloud area including the preset three-dimensional point; according to the preset straight line direction The image characteristics of the preset three-dimensional point determine the projection characteristics of the preset three-dimensional point in the preset straight line direction on the preset projection plane; the preset straight line direction is perpendicular to the preset projection plane; according to the projection characteristics on the preset projection plane , determine the two-dimensional projection of the three-dimensional point cloud image on the preset projection plane.

Based on the above processing, the statistical characteristics corresponding to the amplitudes of the three-dimensional points in the neighborhood of the preset three-dimensional points in the three-dimensional point cloud image can be determined, and based on the determined statistical characteristics, the image characteristics of the preset three-dimensional points can be determined, and based on the preset three-dimensional points. Assume the image characteristics of the preset three-dimensional points in the straight line direction and determine the two-dimensional projection image of the three-dimensional point cloud image on the preset projection plane, that is, the two-dimensional image corresponding to the three-dimensional point cloud image can be generated. Moreover, different statistical features can reflect different image features. Correspondingly, for different application scenarios, different statistical features can be used based on actual needs, and the generated two-dimensional image can reflect the corresponding image features to adapt to the current situation. Application scenarios, that is, embodiments of the present application can improve the quality of the projected two-dimensional image, thereby improving the application effect of the image.

Of course, implementing any product or method of the present application does not necessarily require achieving all the above-mentioned advantages simultaneously.

Description of the drawings

The drawings described here are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation of the present application.

Figure 1 is a flow chart of an image generation method provided by an embodiment of the present application;

Figure 2 is a schematic diagram of a three-dimensional space provided by an embodiment of the present application;

Figure 3 is a schematic diagram of another three-dimensional space provided by an embodiment of the present application;

Figure 4 is a flow chart of another image generation method provided by an embodiment of the present application;

Figure 5 is a flow chart of another image generation method provided by an embodiment of the present application;

Figure 6 is a flow chart of another image generation method provided by an embodiment of the present application;

Figure 7a is a first projection sub-image provided by an embodiment of the present application;

Figure 7b is an enlarged view of the partial image area of the first projection sub-image shown in Figure 7a;

Figure 8a is a second projection sub-image provided by the embodiment of the present application;

Figure 8b is an enlarged view of the partial image area of the second projection sub-image shown in Figure 8a;

Figure 9a is a third projection sub-image provided by the embodiment of the present application;

Figure 9b is an enlarged view of the partial image area of the third projection sub-image shown in Figure 9a;

Figure 10 is a flow chart of another image generation method provided by an embodiment of the present application;

Figure 11 is a schematic diagram of the principle of an image generation method provided by an embodiment of the present application;

Figure 12 is a flow chart of another image generation method provided by an embodiment of the present application;

Figure 13 is a flow chart of another image generation method provided by an embodiment of the present application;

Figure 14 is a comparison diagram of a two-dimensional projection image provided by an embodiment of the present application;

Figure 15 is a comparison of enlarged views of a partial image area corresponding to the same position in each of the two-dimensional projections shown in Figure 14;

Figure 16 is a comparison diagram of another enlarged view of the local image area corresponding to the same position in each of the two-dimensional projections shown in Figure 14;

Figure 17 is a comparison diagram of another enlarged view of the local image area corresponding to the same position in each of the two-dimensional projections shown in Figure 14;

Figure 18 is a comparison diagram of another two-dimensional projection image provided by the embodiment of the present application;

Figure 19 is a comparison diagram of another two-dimensional projection image provided by the embodiment of the present application;

Figure 20 is a comparison diagram of another two-dimensional projection image provided by the embodiment of the present application;

Figure 21 is a comparison diagram of another two-dimensional projection image provided by the embodiment of the present application;

Figure 22 is a comparison diagram of another two-dimensional projection image provided by the embodiment of the present application;

Figure 23 is a structural diagram of an image generation device provided by an embodiment of the present application;

Figure 24 is a structural diagram of an electronic device provided by an embodiment of the present application.

Detailed ways

In order to make the purpose, technical solutions, and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and examples. Obviously, the described embodiments are only some of the embodiments of the present application, but not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of this application.

Embodiments of the present application provide an image generation method, which can be applied to electronic devices. The electronic device can process the obtained three-dimensional point cloud image based on the method provided by the embodiments of the present application to obtain a corresponding two-dimensional image. Correspondingly, after obtaining the two-dimensional image, image recognition can be performed on the two-dimensional image. For example, the category of the object (for example, a person, an object, etc.) in the two-dimensional image can be identified, or the category to which the object in the two-dimensional image belongs can also be identified. The image area, or it can also identify whether the specified object exists in the two-dimensional image, but is not limited to this.

Referring to Figure 1, Figure 1 is a flow chart of an image generation method provided by an embodiment of the present application. The method may include the following steps:

S101: Obtain a three-dimensional point cloud image.

S102: Determine the statistical characteristics corresponding to the amplitudes of the three-dimensional points in the neighborhood of the preset three-dimensional points in the three-dimensional point cloud image, and determine the image characteristics of the preset three-dimensional points based on the statistical characteristics.

Among them, the amplitude of the three-dimensional point is used to characterize the electromagnetic scattering characteristics at the corresponding position of the three-dimensional point, and the neighborhood of the preset three-dimensional point refers to the preset point cloud area including the preset three-dimensional point. The size and shape of the preset point cloud area can be determined according to actual needs. The statistical characteristics corresponding to the amplitudes of the three-dimensional points in the above neighborhood refer to the characteristic quantities obtained by statistically processing each amplitude using statistical rules. Compared with the situation where only the amplitude of the preset three-dimensional point itself is considered, the use of statistical features helps to better characterize the corresponding amplitude characteristics of the preset three-dimensional point as a whole, thereby obtaining high-quality two-dimensional projections in the future. Images lay the foundation. For example, the obtained statistical features can be directly determined as image features of preset three-dimensional points, or the obtained statistical features can be processed, for example, according to preset operation rules (which can be determined according to different processing requirements). The features are operated or transformed, and the processing results are determined as image features of preset three-dimensional points. In the embodiment of the present application, image features may refer to features used to characterize the display characteristics or display effects of the display image corresponding to the three-dimensional point.

S103: Determine the projection characteristics of the preset three-dimensional point in the preset straight line direction on the preset projection plane according to the image characteristics of the preset three-dimensional point in the preset straight line direction.

Wherein, the preset straight line direction is perpendicular to the preset projection plane. If there are multiple preset three-dimensional points in each preset straight line direction, then there are multiple corresponding image features. Then, based on these multiple image features, the projection features corresponding to the preset straight line direction on the preset projection plane can be determined. . Projection features are the direct basis for forming a two-dimensional projection image. The selection of projection features is related to the quality of the two-dimensional projection image. In other words, in the embodiment of the present application, the projection features may refer to display features used to characterize the two-dimensional projection image. or the characteristics of the display effect. For example, the pixel value of the preset three-dimensional point corresponding to the projection feature in the three-dimensional point cloud map can be used to determine the pixel value at the position corresponding to the preset three-dimensional point on the two-dimensional projection map.

S104: Determine the two-dimensional projection image of the three-dimensional point cloud image on the preset projection plane according to the projection characteristics on the preset projection plane.

Based on the image generation method provided by the embodiment of the present application, the statistical characteristics corresponding to the amplitudes of the three-dimensional points in the neighborhood of the preset three-dimensional points in the three-dimensional point cloud image can be determined, and based on the determined statistical characteristics, the preset three-dimensional points can be determined. Image characteristics, and determine the two-dimensional projection image of the three-dimensional point cloud image on the preset projection plane based on the image characteristics of the preset three-dimensional point in the preset straight line direction, that is, the two-dimensional image corresponding to the three-dimensional point cloud image can be generated. Moreover, different statistical features can reflect different image features. Correspondingly, for different application scenarios, different statistical features can be used based on actual needs, and the generated two-dimensional image can reflect the corresponding image features to adapt to the current situation. Application scenarios, that is, embodiments of the present application can improve the quality of the projected two-dimensional image, thereby improving the application effect of the image.

For step S101, the electronic device can obtain the three-dimensional point cloud image through the radar device. The three-dimensional point cloud image in the embodiment of this application may be a SAR (Synthetic Aperture Radar) image. Radar equipment can be lidar, millimeter wave radar, microwave radar, etc.

In one embodiment, step S101 may include the following steps: obtaining a three-dimensional point cloud image of the scanned object.

For example, in security inspection scenarios such as airport security inspection and subway security inspection, the modulated signal can be emitted through the millimeter wave radar in the active millimeter wave security inspection instrument (the modulated signal is a millimeter wave signal), and the modulated signal is scanned according to the imaging algorithm (for example, The echo signals reflected by people, objects, etc.) are processed to obtain a three-dimensional point cloud image of the scanned object. Then, the electronic device can generate a corresponding two-dimensional image based on the obtained three-dimensional point cloud image based on the method provided by the embodiment of the present application. Subsequently, image recognition can be performed on the two-dimensional images to detect, identify and locate dangerous items.

When obtaining a three-dimensional point cloud image through an active millimeter wave security detector, based on the imaging mechanism and actual conditions, the echo signal reflected by the millimeter wave signal at the scanning object is evenly sampled to generate a corresponding three-dimensional point cloud image. Therefore, the obtained three-dimensional point cloud image contains multiple three-dimensional points evenly distributed in the three-dimensional space. The interval between each three-dimensional point is determined based on the sampling interval when the three-dimensional point cloud image is generated. The amplitude of a three-dimensional point in the three-dimensional point cloud image represents The electromagnetic scattering characteristics at the corresponding position of the three-dimensional point in space. The respective amplitudes of each three-dimensional point in the three-dimensional point cloud image can be represented by a three-dimensional matrix.

For the multiple 3D points in the 3D point cloud obtained by other means (for example, lidar), they are sparsely distributed in the 3D space.

When generating a two-dimensional projection image based on a three-dimensional point cloud image, the projection plane where the two-dimensional projection image to be generated is located (ie, the preset projection plane in the embodiment of the present application) can be determined, and the three-dimensional point cloud image can be projected to the preset projection plane, The corresponding two-dimensional projection map can be obtained.

The preset projection plane may include: a projection plane parallel to the front and/or back of the scanned object. In a security inspection scenario, the scanning object is a person, and the preset projection plane can be a projection plane parallel to the front and/or back of the person, and the front of the person is the plane toward which the person's face faces. Correspondingly, the two-dimensional projection of the three-dimensional point cloud image on the preset projection plane contains the complete image information of the person and the image information of the items carried by the person, which can facilitate the security inspection of the person. For example, the character carries a wrench. According to the preset projection plane parallel to the front of the character, the generated two-dimensional projection image contains complete image information of the wrench. Based on the two-dimensional projection image, the wrench can be detected, identified and positioned. .

When the preset projection plane includes projection planes parallel to the front and back of the character, multiple two-dimensional projection images in different directions can be obtained, and the image information of the character can be enriched from different directions and combined with the image information in different directions. Recognition can improve the accuracy of image recognition.

In the security inspection scene, a two-dimensional image under a preset projection perspective can be generated based on the three-dimensional point cloud image. For example, a two-dimensional image of the scanned object under the main perspective (that is, the main view of the scanned object) can be generated based on the three-dimensional point cloud image. A two-dimensional image of the scanned object from a side view (ie, a side view of the scanned object) can also be generated based on the three-dimensional point cloud image.

In one implementation, the electronic device can determine that the projection plane under the preset projection perspective is the preset projection plane, and determine whether the original three-dimensional point cloud image obtained by the radar device matches the preset projection perspective. If the original three-dimensional point cloud image matches the preset projection perspective, for example, the preset projection plane is parallel to a coordinate plane in the three-dimensional coordinate system of the original three-dimensional point cloud image, the electronic device can directly calculate the preset three-dimensional angle in the original three-dimensional point cloud image. The statistical characteristics corresponding to the amplitude of three-dimensional points in the neighborhood of the point.

For example, see FIG. 2 , which is a schematic diagram of a three-dimensional space provided by an embodiment of the present application. The scanning object is a person. The three-dimensional range shown by the cuboid in Figure 2 represents the original three-dimensional point cloud image of the person. The coordinate planes of the three-dimensional coordinate system shown in Figure 2 include: XOY plane, XOZ plane and YOZ plane. When it is necessary to generate the front view of the character as shown in Figure 2, the preset projection angle is the main view angle, and the preset projection plane under the main view angle is a projection plane parallel to the front of the character, then the preset projection plane and the three-dimensional coordinates The XOY plane in the system is parallel, that is, the original three-dimensional point cloud image matches the preset projection perspective. Correspondingly, after the electronic device obtains the original three-dimensional point cloud image shown in Figure 2, it can directly calculate the statistical characteristics corresponding to the amplitudes of the three-dimensional points in the three-dimensional point cloud image that are in the neighborhood of the preset three-dimensional points.

In another implementation, step S101 may include the following steps: rotating the original three-dimensional point cloud image based on a preset projection perspective to obtain a three-dimensional point cloud image that matches the preset projection perspective. Wherein, the preset projection plane is the projection plane under the preset projection angle.

If the original three-dimensional point cloud image obtained through the radar device does not match the preset projection perspective, for example, the preset projection plane is not parallel to all coordinate planes in the three-dimensional coordinate system of the three-dimensional point cloud image, in order to generate the three-dimensional point cloud image, in the preset projection A two-dimensional projection image under a viewing angle. After acquiring the original three-dimensional point cloud image, the electronic device can rotate the original three-dimensional point cloud image based on the positional relationship between the preset projection angle of view and the original three-dimensional point cloud image to obtain the same as the preset projection angle of view. Matching 3D point cloud images.

The electronic device can perform coordinate conversion on the original three-dimensional point cloud image based on the two-dimensional coordinate system of the preset projection plane and the coordinate mapping relationship between the two-dimensional coordinate system of the preset projection plane and the three-dimensional coordinate system of the original three-dimensional point cloud image, so as to After coordinate conversion, one coordinate plane in the three-dimensional coordinate system of the obtained three-dimensional point cloud image is parallel to the preset projection plane, and the three-dimensional point cloud image obtained by the coordinate conversion is a three-dimensional point cloud image that matches the preset projection perspective. The electronic device can then calculate the statistical characteristics corresponding to the amplitudes of the three-dimensional points in the neighborhood of the preset three-dimensional points in the three-dimensional point cloud image.

For example, see FIG. 3 , which is a schematic diagram of another three-dimensional space provided by an embodiment of the present application. The scanning object is a person. The three-dimensional range shown by the cuboid in Figure 3 represents the original three-dimensional point cloud image of the person. The coordinate planes of the three-dimensional coordinate system shown in Figure 3 include: XOY plane, XOZ plane and YOZ plane. When it is necessary to generate a two-dimensional projection of the character from the side view as shown in Figure 3, the default projection view is the side view, and the preset projection plane in the side view is the X ₁ O ₁ Y ₁ plane, X ₁ O _{The 1} Y ₁ plane is not parallel to any coordinate plane in the three-dimensional coordinate system.

Correspondingly, the electronic device can be based on the two-dimensional coordinate system of the X ₁ O ₁ Y ₁ plane and the coordinate mapping relationship between the two-dimensional coordinate system of the X ₁ O ₁ Y ₁ plane and the three-dimensional coordinate system of the original three-dimensional point cloud image. The original three-dimensional point cloud image undergoes coordinate transformation so that the XOY plane in the three-dimensional coordinate system of the three-dimensional point cloud image is parallel to the X ₁ O ₁ Y ₁ plane after the coordinate transformation. The electronic device can then calculate the statistical characteristics corresponding to the amplitudes of the three-dimensional points in the neighborhood of the preset three-dimensional points in the three-dimensional point cloud image.

Regarding step S102, the preset three-dimensional point can be any three-dimensional point in the three-dimensional point cloud image. The neighborhood of the preset three-dimensional point refers to the preset point cloud area including the preset three-dimensional point.

Preset point cloud areas can be determined based on requirements. The preset point cloud area may be a one-dimensional area, for example, in a three-dimensional point cloud image, each three-dimensional point is located on a straight line perpendicular to the preset projection plane and including the preset three-dimensional point. Alternatively, the preset point cloud area can also be a two-dimensional area. For example, each three-dimensional point in the three-dimensional point cloud image is located in a plane area parallel to the preset projection plane and containing the preset three-dimensional point. The shape of the plane area can be a rectangle, Round etc. Alternatively, the preset point cloud area may also be a three-dimensional area, for example, each three-dimensional point in a cuboid or sphere containing the preset three-dimensional points in the three-dimensional point cloud image, which is not specifically limited in this embodiment.

Different image features of three-dimensional points can be determined based on different statistical features. When determining the image features of preset three-dimensional points, different feature statistical methods can be selected based on different needs.

In one embodiment, the image features of the preset three-dimensional point may include at least one of the following: a first image feature used to participate in characterizing image details, a second image feature used to participate in determining smooth areas of the image, and a second image feature used to participate in characterizing image details. The third image feature of image background noise.

For example, based on the first image feature, the second image feature, and the third image feature, the electronic device generates a two-dimensional projection image that can reflect image details and image smooth areas, and can reduce image background noise, and can improve the generated two-dimensional projection image. The quality of the projection image to improve the accuracy of identifying scanned objects in the image. For example, when the two-dimensional projection image is applied to a security inspection scene, the two-dimensional projection image can reflect the details of different scanned objects. In addition, since the surface of the human body is relatively smooth, the obtained two-dimensional projection image can reflect the image area where the person is located, and then security inspection based on the two-dimensional projection image can accurately distinguish the persons in the two-dimensional projection image and the items carried by the person. , to determine whether the items carried by the person are dangerous goods, which can improve the accuracy of dangerous goods identification.

In one embodiment, based on Figure 1, step S102 may include at least one of the following operations, where Figure 4 is an example, showing all the operations included in step S102:

S1021: Determine the first value corresponding to the amplitude of the three-dimensional point in the first neighborhood of the preset three-dimensional point in the three-dimensional point cloud image through at least one characteristic statistical method among variance calculation, gradient calculation, and Laplacian operator. Statistical features, as the first image features of preset three-dimensional points.

The plane where the first neighborhood is located is parallel to the preset projection plane.

S1022: Determine the amplitude correspondence of the three-dimensional point in the first neighborhood of the preset three-dimensional point in the three-dimensional point cloud image through at least one characteristic statistical method among entropy value calculation, integrated side-lobe ratio calculation, and peak side-lobe ratio calculation. The second statistical feature is used as the second image feature of the preset three-dimensional point.

S1023: Determine the third value corresponding to the amplitude of the three-dimensional point in the second neighborhood of the preset three-dimensional point in the three-dimensional point cloud image through variance calculation, gradient calculation, or at least one characteristic statistical method of the Laplacian operator. Statistical features are used as third image features of preset three-dimensional points.

Wherein, the plane of the second neighborhood is perpendicular to the preset projection plane.

The statistical characteristics of the preset three-dimensional points in different dimensions can be obtained based on different feature statistics methods, and the image characteristics of the preset three-dimensional points in different dimensions can be determined based on the statistical characteristics of different dimensions. Correspondingly, in order to make the generated two-dimensional projection map reflect image features of different dimensions and facilitate subsequent image recognition of the two-dimensional projection map, the electronic device can use a variety of feature statistics methods to calculate the number of points in the neighborhood of the preset three-dimensional point. Statistical characteristics corresponding to the amplitude of three-dimensional points. Then, the image characteristics of the preset three-dimensional points can be determined based on the obtained statistical characteristics.

Using at least one characteristic statistical method such as variance calculation, gradient calculation, and Laplacian operator to determine the statistical characteristics corresponding to the amplitudes of the three-dimensional points in the neighborhood of the preset three-dimensional points in the three-dimensional point cloud image (i.e., the first statistical characteristics ). For example, gradient calculation and Laplacian operator can be used to extract the edge features of the scanned object, and the variance value obtained by variance calculation can reflect the dispersion of the amplitudes of the three-dimensional points in the neighborhood of the preset three-dimensional points. The dispersion represents The difference in amplitude of the three-dimensional points in the neighborhood of the preset three-dimensional point. The amplitude of the three-dimensional point corresponding to the edge of the scanned object in the three-dimensional point cloud image is greatly different from the amplitude of other three-dimensional points. Therefore, the first statistical feature can reflect the edge features of the scanned object corresponding to the three-dimensional point cloud image and the image details of different areas in the three-dimensional point cloud image. Using the first image feature that participates in characterizing image details as an evaluation index in the process of determining the two-dimensional projection map can avoid the loss of details in the image caused by background clutter covering dark areas in strongly scattering targets.

For security inspection scenarios, three-dimensional point cloud images can be obtained through active millimeter wave security inspection devices. According to the focusing characteristics of the three-dimensional point cloud image and the electromagnetic scattering characteristics of the scanning object, it can be known that in the three-dimensional point cloud image, the target area with strong electromagnetic scattering contains fewer three-dimensional points in the distance direction, and the distance direction is the direction perpendicular to the preset projection plane. That is, only the three-dimensional points with a smaller distance upward contain the focus information of the target area. The remaining three-dimensional points with a distance upward correspond to other areas, for example, background areas or transmission areas. The amplitudes of the corresponding three-dimensional points in other areas in the three-dimensional point cloud map are lower than the amplitudes of the corresponding three-dimensional points in the three-dimensional point cloud map of the target area. That is, the target area with strong electromagnetic scattering has a larger amplitude of the corresponding three-dimensional point in the three-dimensional point cloud diagram. Therefore, the dispersion of the three-dimensional points in the target area in the three-dimensional point cloud diagram is also larger.

Therefore, in order to avoid the loss of image details in the generated two-dimensional projection map due to the background area covering up the darker areas in the target area, at least one characteristic statistical method among variance calculation, gradient calculation and Laplacian operator can be used. , calculate the first statistical feature corresponding to the amplitude of the three-dimensional point in the first neighborhood of the preset three-dimensional point in the three-dimensional point cloud image, as the first image feature of the preset three-dimensional point cloud image. Since the first image feature is used to participate in representing image details, the two-dimensional projection map obtained based on the first image feature can reflect the details of the scanned object, that is, the quality of the generated two-dimensional projection map can be improved, and subsequent processing of the two-dimensional projection map can be facilitated. Image Identification.

Using at least one characteristic statistical method such as entropy value calculation, peak side lobe ratio calculation, and integral side lobe ratio calculation, the statistical characteristics corresponding to the amplitude of the three-dimensional point in the first neighborhood of the preset three-dimensional point in the three-dimensional point cloud image are determined ( That is, the second statistical characteristic), which can reflect the degree of order and focus of the amplitudes of the three-dimensional points in the first neighborhood of the preset three-dimensional points. The degree of order represents the amplitude of the three-dimensional points in the neighborhood of the preset three-dimensional points. The changing trend, for example, smooth increase or smooth decrease, etc. The degree of focus represents the degree of concentration of the amplitudes of the three-dimensional points in the neighborhood of the preset three-dimensional point. In the three-dimensional point cloud image, the amplitudes of the three-dimensional points corresponding to the smooth areas in the scanned object have a high degree of concentration and a smooth changing trend. Therefore, the second statistical feature can reflect the smoothness of different areas in the scanned object corresponding to the three-dimensional point cloud image. .

For security inspection scenarios, the scanning object can be a person. Since the surface of the human body is relatively smooth, the amplitude of the corresponding three-dimensional points on the human body surface in the three-dimensional point cloud image has a high degree of order. Therefore, in order to make the generated two-dimensional projection image reflect the smooth area of the scanned object, at least one of the characteristic statistical methods of entropy value calculation, integral side-lobe ratio calculation, and peak side-lobe ratio calculation can be used to calculate the three-dimensional point cloud image. The second statistical feature corresponding to the amplitude of the three-dimensional point located in the first neighborhood of the preset three-dimensional point is used as the second image feature of the preset three-dimensional point. Since the second image feature is used to determine the smooth area of the image, the two-dimensional projection map obtained based on the second image feature can reflect the smooth area of the scanned object, that is, it can improve the quality of the generated two-dimensional projection map and facilitate subsequent two-dimensional Projection images for image recognition.

Using at least one characteristic statistical method such as variance calculation, gradient calculation, and Laplacian operator, the statistical characteristics corresponding to the amplitudes of the three-dimensional points in the neighborhood of the preset three-dimensional points in the three-dimensional point cloud image (i.e., the third statistical characteristics ), can reflect the dispersion of the amplitudes of the three-dimensional points in the second neighborhood of the preset three-dimensional point, and the dispersion represents the difference in the amplitudes of the three-dimensional points in the neighborhood of the preset three-dimensional points. For example, the amplitude of each three-dimensional point in the background area in the three-dimensional point cloud image is greatly different from the amplitude of each three-dimensional point in the target area containing the scanned object in the three-dimensional point cloud image. Therefore, the third statistical feature can be used to distinguish the background area and the target area containing the scanned object.

For security inspection scenarios, the target area with strong electromagnetic scattering in the 3D point cloud image obtained by the active millimeter wave security inspection instrument contains fewer 3D points in the distance direction, and the amplitude of the contained 3D points changes greatly. The background region has smaller changes in the magnitude of the three-dimensional points it contains in the distance direction. Therefore, the difference in amplitude of the three-dimensional points in the target area and the background area in the three-dimensional point cloud image is large.

Therefore, in order to make the generated two-dimensional projection map reflect the background area and the target area containing the scanning object, and reduce the influence of the background area on the target area containing the scanning object, variance calculation, gradient calculation, or Laplacian operator can be used At least one feature statistics method in the method calculates the third statistical feature corresponding to the amplitude of the three-dimensional point in the second neighborhood of the preset three-dimensional point in the three-dimensional point cloud image as the third image feature of the preset three-dimensional point. Since the third image feature is used to participate in characterizing the image background noise, the two-dimensional projection map obtained based on the third image feature can reduce the image background noise, and can also distinguish the background area from the target area containing the scanned object, that is, the generated image can be improved. The quality of the two-dimensional projection image facilitates image recognition of the two-dimensional projection image.

When determining the statistical characteristics corresponding to the amplitudes of the three-dimensional points in the neighborhood of the preset three-dimensional points in the three-dimensional point cloud image, in addition to the characteristic statistical methods listed above, it can also be implemented through custom functions to retain image details and determine Smooth areas of the image, and determine effects such as image background noise. The specific form of the custom function can be determined according to specific processing requirements. For example, if the custom function is a power function, the electronic device can calculate the power sum of the amplitudes of the three-dimensional points in the neighborhood of the preset three-dimensional point in the three-dimensional point cloud image based on the power function, as the image feature of the preset three-dimensional point. . Correspondingly, the obtained image features can characterize the difference between the amplitude of the three-dimensional points in the neighborhood of the preset three-dimensional point and the average amplitude of the image background noise, thereby retaining image details in the generated two-dimensional projection map.

In one embodiment, based on Figure 1, step S103 may include at least one of the following operations, where Figure 5 is an example, showing all the operations included in step S103:

S1031: Determine the feature with the largest value among the first image features of the preset three-dimensional point in the preset straight line direction as the first projection feature of the preset three-dimensional point in the preset straight line direction on the preset projection plane.

S1032: Determine the feature with the largest value among the second image features of the preset three-dimensional point in the preset straight line direction as the second projection feature of the preset three-dimensional point in the preset straight line direction on the preset projection plane.

S1033: Determine the third image feature of the preset three-dimensional point in the preset straight line direction as the third projection feature of the preset three-dimensional point in the preset straight line direction on the preset projection plane.

The above-mentioned preset straight line direction is perpendicular to the preset projection plane, and the plane of the second neighborhood considered in the process of determining the third image characteristics is perpendicular to the preset projection plane. Therefore, when the second neighborhood is a one-dimensional neighborhood, for the same For a preset three-dimensional point, its second neighborhood is parallel to or coincides with the preset straight line. For a completely coincident situation (each preset three-dimensional point in the second neighborhood is each preset three-dimensional point on the preset straight line). ), the calculated statistical feature or third image feature is the same value. That is to say, at this time, the third statistical features corresponding to the amplitudes of the three-dimensional points in the second neighborhood of each preset three-dimensional point are the same, and correspondingly, the third image features of each preset three-dimensional point are also the same. Therefore, there is no need to determine the maximum value in the third image feature, and the third image feature of the preset three-dimensional point can be directly determined as the third projection feature.

Furthermore, based on the obtained projection characteristics on the preset projection plane, the two-dimensional projection image of the three-dimensional point cloud image on the preset projection plane can be determined. That is, in an optional embodiment, determining the two-dimensional projection image of the three-dimensional point cloud image on the preset projection plane based on the projection characteristics on the preset projection plane includes: based on the first projection characteristics on the preset projection plane, At least one of the second projection feature and the third projection feature determines the two-dimensional projection image of the three-dimensional point cloud image on the preset projection plane. Specifically, the projection characteristics involved in forming the two-dimensional projection image can be determined based on the generation requirements for the image characteristics or image effects of the two-dimensional projection image.

In one embodiment, the two-dimensional projection image of the three-dimensional point cloud image on the preset projection plane can be determined based on the first projection feature, the second projection feature, and the third projection feature on the preset projection plane at the same time, as shown in the figure As shown in 5, step S104 may include the following steps:

S1041: Based on the first projection feature, the second projection feature, and the third projection feature on the preset projection plane, determine the two-dimensional projection image of the three-dimensional point cloud image on the preset projection plane.

For example, the feature with the largest value among the first image features may be the feature corresponding to the maximum amplitude dispersion. The dispersion reflects the difference between the amplitudes of the three-dimensional points, and the edges of the scanned objects are adjacent to them. The difference in amplitude of the position is large. Therefore, the preset three-dimensional point corresponding to the feature with the largest value among the first image features is also the three-dimensional point corresponding to the edge of the scanned object. Furthermore, based on the first image feature, the edge of the scanned object in the three-dimensional point cloud image can be effectively determined to enrich the image details of the generated two-dimensional projection image.

Therefore, in order to make the generated two-dimensional projection map reflect the details of the scanned object, the electronic device can determine the feature with the largest value among the first image features of the preset three-dimensional point in the preset straight line direction as the feature in the preset straight line direction. The first projection feature of the preset three-dimensional point on the preset projection plane.

For example, the feature with the largest value among the second image features may be the feature corresponding to the maximum amplitude order degree. The order degree can reflect the smoothness of different areas in the scanned object corresponding to the three-dimensional point cloud image. When the order degree is the maximum The area where the preset three-dimensional point corresponding to the image feature is located is the image area with the greatest smoothness. In order to make the generated two-dimensional projection map reflect the smooth area of the scanned object, the electronic device can place the preset three-dimensional point in the preset straight line direction. The feature with the largest value among the second image features is determined as the second projection feature of the preset three-dimensional point in the preset straight line direction on the preset projection plane.

The third image feature of the preset three-dimensional point is used to participate in determining the image background noise. In order to reduce the impact of the image background noise on the target area containing the scanned object, the generated two-dimensional projection map can distinguish the background area from the target area containing the scanned object. , the electronic device can determine the third image feature of the preset three-dimensional point in the preset straight line direction as the third projection feature of the preset three-dimensional point in the preset straight line direction on the preset projection plane.

Furthermore, the electronic device can use different methods to generate a two-dimensional projection image of the three-dimensional point cloud image on the preset projection plane based on the first projection feature, the second projection feature, and the third projection feature on the preset projection plane.

In another implementation, the two-dimensional projection image of the three-dimensional point cloud image on the preset projection plane is determined based on at least one of the first projection feature, the second projection feature, and the third projection feature on the preset projection plane. ,include:

Determine the first pixel value of the preset three-dimensional point corresponding to the first projection feature on the preset projection plane in the three-dimensional point cloud image, and generate the first projection sub-image based on the first pixel value;

Determine the second pixel value of the preset three-dimensional point corresponding to the second projection feature on the preset projection plane in the three-dimensional point cloud image, and generate a second projection sub-image based on the second pixel value;

Use the third projection feature on the preset projection plane as the third pixel value, and generate a third projection sub-image based on the third pixel value;

Fusion processing is performed on at least two of the first projection sub-image, the second projection sub-image and the third projection sub-image to obtain a two-dimensional projection image corresponding to the three-dimensional point cloud image. Specifically, according to the image characteristics or image effect generation requirements of the two-dimensional projection image, the projection characteristics participating in the formation of the two-dimensional projection image can be determined, and then based on the determined projection characteristics, it is determined that the projection sub-image needs to be generated, and finally through image fusion processing Generate a two-dimensional projection of the requirements.

As an example of the embodiment of the present application, see Figure 6, which shows the fusion process of the first projection sub-image, the second projection sub-image and the third projection sub-image to obtain a two-dimensional projection image corresponding to the three-dimensional point cloud image. That is, step S1041 may include the following steps:

S10411: Determine the first pixel value of the preset three-dimensional point corresponding to the first projection feature on the preset projection plane in the three-dimensional point cloud image, and generate the first projection sub-image based on the first pixel value.

S10412: Determine the second pixel value of the preset three-dimensional point corresponding to the second projection feature on the preset projection plane in the three-dimensional point cloud image, and generate a second projection sub-image based on the second pixel value.

S10413: Use the third projection feature on the preset projection plane as the third pixel value, and generate the third projection sub-image based on the third pixel value.

S10414: Fusion process the first projection sub-image, the second projection sub-image and the third projection sub-image to obtain a two-dimensional projection image corresponding to the three-dimensional point cloud image.

A preset straight line corresponds to a pixel point in the preset projection plane. The preset three-dimensional point on the preset straight line is reversely determined according to the projection characteristics. The pixel value corresponding to the preset three-dimensional point is also the pixel value in the preset projection plane. The pixel value of this pixel. Correspondingly, by determining the pixel values corresponding to the preset three-dimensional points in each preset straight line direction, that is, determining the pixel values of each pixel point in the preset projection plane, the corresponding projection sub-image can be obtained.

The pixel value corresponding to a preset three-dimensional point is the amplitude of the preset three-dimensional point in the three-dimensional point cloud image.

Exemplarily, for the embodiment shown in Figure 2, the three-dimensional range shown by the cuboid in Figure 2 represents the three-dimensional point cloud image of the character, and the preset projection plane is the XOY plane. The default straight line direction is perpendicular to the XOY plane. It is preset that the plane where the first neighborhood of the three-dimensional point is located is parallel to the XOY plane.

For each preset three-dimensional point in the preset straight line direction, the position of the preset three-dimensional point in the three-dimensional coordinate system can be expressed as (x, y, z). Since each three-dimensional point is located in the preset straight line direction perpendicular to the XOY plane, the x and y of each three-dimensional point are the same.

Correspondingly, the preset three-dimensional point corresponding to the first projection feature can be recorded as (x, y, z ₁ ). The first pixel value of the preset three-dimensional point corresponding to the first projection feature in the three-dimensional point cloud image is: the amplitude of the preset three-dimensional point in the three-dimensional point cloud image, then the first pixel value corresponding to the preset three-dimensional point corresponding to the first projection feature A pixel value σ ₁ can be expressed as:

σ ₁ =σ(x,y,z ₁ ) (1)

σ(x,y,z ₁ ) represents the amplitude of the preset three-dimensional point corresponding to the first projection feature in the three-dimensional point cloud image.

Referring to Figures 7a and 7b, Figure 7a is a first projection sub-image provided by an embodiment of the present application. The electronic device may generate a first projection sub-image as shown in Figure 7a based on the first projection feature.

The image on the left side in Figure 7b is an enlarged view of the background area (areas other than the image area where the character is located) in the first projection sub-image shown in Figure 7a. The middle image of Figure 7b is an enlarged view of the area where the character's back is located in the first projection sub-image shown in Figure 7a. The image on the right side in Figure 7b is an enlarged view of the area where the character's legs are located in the first projection sub-image shown in Figure 7a.

It can be obtained by combining Figure 7a and Figure 7b that both Figure 7a and Figure 7b contain more image details. For example, the image in the middle of Figure 7b can reflect the different positions of the character's back area, and the image on the right side of Figure 7b can reflect that a wrench is placed on the character's leg.

Correspondingly, the preset three-dimensional point corresponding to the second projection feature can be recorded as (x, y, z ₂ ). The second pixel value of the preset three-dimensional point corresponding to the second projection feature in the three-dimensional point cloud image is: the amplitude of the preset three-dimensional point in the three-dimensional point cloud image, then the second pixel value corresponding to the preset three-dimensional point corresponding to the second projection feature The two-pixel value σ ₂ can be expressed as:

σ ₂ =σ(x,y,z ₂ ) (2)

σ(x,y,z ₂ ) represents the amplitude of the preset three-dimensional point corresponding to the second projection feature in the three-dimensional point cloud image.

Referring to Figures 8a and 8b, Figure 8a is a second projection sub-image provided by an embodiment of the present application. The electronic device may generate the second projection sub-image shown in FIG. 8a based on the second projection feature.

The image on the left in Figure 8b is an enlarged view of the background area in the second projected sub-image shown in Figure 8a. The middle image of Figure 8b is an enlarged view of the area where the character's back is located in the second projection sub-image shown in Figure 8a. The image on the right side in Figure 8b is an enlarged view of the area where the character's legs are located in the second projection sub-image shown in Figure 8a.

It can be obtained by combining Figure 8a and Figure 8b that both Figure 8a and Figure 8b can reflect the smooth area of the image, that is, they can reflect the smoothness of the scanned object. For example, since the smoothness of the human body surface is different from that of other items (e.g., a wrench carried by the character), the area containing the scanned object in Figure 8a has a different brightness than the background area. In the middle image of Figure 8b, the brightness difference at each position is small. In the image on the right side of Figure 8b, the brightness of the area where the character's legs are located and the area where the wrench carried by the character is located are different.

Exemplarily, for the embodiment shown in Figure 2, the three-dimensional range shown by the cuboid in Figure 2 represents the three-dimensional point cloud image of the character, and the preset projection plane is the XOY plane. The default straight line direction is perpendicular to the XOY plane. The plane where the second neighborhood of the preset three-dimensional point is located is perpendicular to the XOY plane.

For each preset three-dimensional point in the preset straight line direction, the position of the preset three-dimensional point in the three-dimensional coordinate system can be expressed as (x, y, z). Correspondingly, the third pixel value σ ₃ corresponding to the preset three-dimensional point corresponding to the third projection feature can be expressed as:

σ ₃ =f(σ(x,y,z)) (3)

f() represents the objective function, which is used to calculate the third image feature of the preset three-dimensional point. σ(x,y,z) represents the amplitude of the three-dimensional point in the neighborhood of the preset three-dimensional point in the three-dimensional point cloud image.

Referring to Figures 9a and 9b, Figure 9a is a third projection sub-image provided by an embodiment of the present application. The electronic device may generate the third projection sub-image shown in FIG. 9a based on the third projection feature.

The image on the left in Figure 9b is an enlarged view of the background area in the third projection sub-image shown in Figure 9a. The middle image of Figure 9b is an enlarged view of the area where the character's back is located in the third projection sub-image shown in Figure 9a. The image on the right side in Figure 9b is an enlarged view of the area where the character's legs are located in the third projection sub-image shown in Figure 9a.

Figures 9a and 9b can clearly show the boundary between the target area where the scanned object is located and the background area, as well as the complete outline of the scanned object, and there is less noise in the background area.

Furthermore, after obtaining the first projection sub-image, the second projection sub-image and the third projection sub-image, the electronic device can fuse at least two of the first projection sub-image, the second projection sub-image and the third projection sub-image. After processing, the two-dimensional projection image corresponding to the three-dimensional point cloud image is obtained.

For example, in one implementation, if the data magnitudes of the first projection sub-image, the second projection sub-image and the third projection sub-image are the same, for example, the electronic device will be in a preset three-dimensional direction in the preset straight line direction. The feature with the largest value among the first image features of the point is determined as the first projection feature, and the amplitude of the preset three-dimensional point corresponding to the first projection feature in the three-dimensional point cloud image is used as the pixel value to obtain the first projection sub-image. The feature with the largest value in the second image feature of the preset three-dimensional point in the preset straight line direction is determined as the second projection feature, and the amplitude of the preset three-dimensional point corresponding to the second projection feature in the three-dimensional point cloud image is determined As pixel values, the second projected sub-image is obtained. The feature with the largest median value of the third image feature of the preset three-dimensional point in the preset straight line direction is determined as the third projection feature, and the amplitude of the preset three-dimensional point corresponding to the third projection feature in the three-dimensional point cloud image is determined As pixel values, the third projected sub-image is obtained. Correspondingly, the first projection sub-image, the second projection sub-image and the third projection sub-image are all data magnitudes corresponding to the amplitudes of the three-dimensional points in the three-dimensional point cloud image.

The electronic device can directly perform fusion processing on the first projection sub-image, the second projection sub-image and the third projection sub-image based on the preset fusion algorithm to obtain a two-dimensional projection image corresponding to the three-dimensional point cloud image.

For example, the electronic device can perform fusion processing on each projection sub-image according to the following formula.

H(x,y)＝g(h ₁ (x, y), h ₂ (x, y), h ₃ (x, y)) (4)

H(x,y) represents the pixel value of the pixel point with coordinates (x,y) in the two-dimensional projection image. g() represents the fusion function. h ₁ (x, y) represents the pixel value of the pixel point with coordinates (x, y) in the first projection sub-image. h ₂ (x, y) represents the pixel value of the pixel point with coordinates (x, y) in the second projection sub-image. h ₃ (x, y) represents the pixel value of the pixel point with coordinates (x, y) in the third projection sub-image.

In one embodiment, the electronic device can perform fusion processing on two projection sub-images with the same data magnitude among the first projection sub-image, the second projection sub-image and the third projection sub-image to obtain an intermediate image. In addition, the electronic device can pre-process (for example, normalize) another projection sub-image, and fuse the intermediate image with the pre-processing result to obtain a corresponding two-dimensional projection image.

For example, the electronic device determines the feature with the largest value among the first image features of the preset three-dimensional point in the preset straight line direction as the first projection feature, and adds the preset three-dimensional point corresponding to the first projection feature in the three-dimensional point cloud image. The amplitude in is used as the pixel value to obtain the first projected sub-image. The feature with the largest value in the second image feature of the preset three-dimensional point in the preset straight line direction is determined as the second projection feature, and the amplitude of the preset three-dimensional point corresponding to the second projection feature in the three-dimensional point cloud image is determined As pixel values, the second projected sub-image is obtained. The first projection sub-image and the second projection sub-image are both data magnitudes corresponding to the amplitudes of the three-dimensional points in the three-dimensional point cloud image. The electronic device determines the third image feature of the preset three-dimensional point in the preset straight line direction as the third projection feature, and uses the third projection feature as the pixel value to obtain the third projection sub-image. The third projection sub-image is the data magnitude corresponding to the statistical characteristics of the amplitude of the three-dimensional points in the three-dimensional point cloud image.

Since the data magnitudes of the first projection sub-image and the second projection sub-image are the same, the electronic device can perform fusion processing on the first projection sub-image and the second projection sub-image to obtain an intermediate image. In addition, the electronic device can pre-process the third projection sub-image, and fuse the intermediate image with the pre-processing result to obtain a corresponding two-dimensional projection image.

Correspondingly, based on Figure 6, referring to Figure 10, step S10414 may include the following steps:

S104141: Use the preset fusion algorithm to process the first projection sub-image and the second projection sub-image to obtain the first intermediate image.

S104142: Perform normalization processing and grayscale transformation processing on the third projection sub-image to obtain the second intermediate image.

Among them, the grayscale transformation process is used to adjust the difference in pixel values between the background area and the target area on the third projection sub-image.

S104143: Determine the product of the pixel values corresponding to the same position on the first intermediate image and the second intermediate image, and generate a two-dimensional projection map corresponding to the three-dimensional point cloud image based on the product.

In the embodiment of this application, the fusion algorithm involved may include but is not limited to any one of the weighted fusion algorithm, the pyramid fusion algorithm, and the maximum mapping algorithm.

For example, the above-mentioned preset fusion algorithm is a weighted fusion algorithm. The pixel values corresponding to the same position in the first projection sub-image and the second projection sub-image can correspond to different weights. The electronic device can calculate the third projection according to the corresponding weights. The first intermediate image can be obtained by taking the weighted sum of the pixel values corresponding to the same position in the first projection sub-image and the second projection sub-image, and using the calculated weighted sum as the pixel value. Alternatively, the preset fusion algorithm is a maximum mapping algorithm, and the electronic device can select the largest pixel value from the pixel values corresponding to the same position in the first projection sub-image and the second projection sub-image to obtain the first intermediate image.

The electronic device can also normalize the third projection sub-image. For example, the pixel value of the third projection sub-image can be normalized to [0, 1] to adjust the data amount of the third projection sub-image. level, so that the data magnitude of the third projection sub-image is smaller than the data magnitude of the first projection sub-image and the second projection sub-image. Then, a grayscale transformation is performed on the normalized third projection subimage. For example, the electronic device can perform a linear grayscale transformation on the normalized third projection subimage, or it can also be based on a nonlinear transformation. The function (for example, exponential function, logarithmic function, power function, gamma function, etc.) performs nonlinear grayscale transformation on the normalized third projection sub-image to adjust the normalized third projection sub-image. The contrast of different areas in the image can avoid the loss of image detail information in the subsequent image fusion process, or ensure a better noise reduction effect.

Specifically, through grayscale transformation processing, the difference in pixel values between the background area and the target area on the third projection sub-image can be reduced (that is, the contrast of different image areas is reduced), or the difference between the background area and the target area on the third projection sub-image can be improved. The difference in pixel values between areas (i.e., improving the contrast of different image areas) is related to the image processing requirements and the functional form used in the grayscale transformation process. For example, in order to achieve a better noise reduction effect during the image fusion process, an exponential function can be used in the grayscale transformation process to improve the contrast of different areas in the projected sub-image; or, for example, in order to retain more noise during the image fusion process For image detail information, a logarithmic function can be used to reduce the contrast of different areas in the projected sub-image.

Taking the millimeter wave security inspection scene as an example, the echo at the edge of the human body is weak due to the scattering angle. If the contrast of the projected sub-image is too high, part of the human body edge information may be lost in the subsequent image fusion process. Therefore, through grayscale Transformation reduces image contrast, ensuring that as much useful detail information is retained as possible while reducing background noise. If the contrast of the projected sub-image is low, the noise reduction effect during the image fusion process will be average. Therefore, the contrast can be improved through grayscale transformation, thereby achieving better noise reduction effect during the image fusion process.

Performing grayscale transformation processing on the normalized third projection sub-image can adjust the difference in pixel values between the background area on the normalized third projection sub-image and the target area containing the scanned object, so that the final result is The two-dimensional projection image can clearly distinguish the background area from the target area containing the scanned object.

Furthermore, the electronic device can calculate a product of pixel values corresponding to the same position on the first intermediate image and the second intermediate image, and generate a two-dimensional projection image corresponding to the three-dimensional point cloud image based on the calculated product. For example, the electronic device can normalize the calculated product according to the preset normalization interval (for example, 0 to 255), and use the normalized result as a pixel value to obtain a two-dimensional projection image corresponding to the three-dimensional point cloud image. .

Based on the above processing, by presetting different image features of three-dimensional points, image details can be retained, image smooth areas can be determined, and image background noise can be determined to reduce the impact of image background noise on the target area containing the scanned object, that is, to achieve noise reduction. . Correspondingly, the obtained two-dimensional projection image can reflect the details and smooth areas of the scanned object, and can distinguish the background area from the target area containing the scanned object. The background noise is low, which can improve the quality of the generated two-dimensional projection image. Subsequent When the two-dimensional projection map is applied to different scenes, the application effect of the image is improved.

Refer to Figure 11, which is a schematic diagram of the principle of an image generation method provided by an embodiment of the present application. The larger cube in Figure 11 represents the 3D point cloud, which is the three-dimensional point cloud image in the embodiment of this application. Different statistical features represent different image features of the preset three-dimensional points in the embodiment of the present application. The three smaller cubes in Figure 11 respectively represent the respective neighborhoods of the three preset three-dimensional points in the preset straight line direction. The preset projection plane is the front of the larger cube in Figure 11, and the line connecting the three preset three-dimensional points (ie, the preset straight line) is perpendicular to the preset projection plane.

The electronic device can obtain a three-dimensional point cloud image, and based on different feature statistics methods, calculate the statistical characteristics corresponding to the amplitudes of the three-dimensional points in the neighborhood of the preset three-dimensional points. Based on the calculated statistical characteristics, it is determined that the preset three-dimensional points are different. The image features of the preset three-dimensional points can be obtained in different dimensions. Then, the electronic device can determine the projection characteristics of the preset three-dimensional point corresponding to the image feature on the preset projection plane based on each image feature of the preset three-dimensional point. Furthermore, the electronic device can obtain the projection sub-image corresponding to the projection feature based on each projection feature of the preset three-dimensional point on the preset projection plane. Furthermore, the electronic device can fuse each projection sub-image to obtain a two-dimensional projection image of the three-dimensional point cloud image on the preset projection plane.

Based on the above processing, different image features of three-dimensional points can be preset to retain image details, determine image smooth areas, and determine image background noise, thereby reducing the impact of image background noise on the target area containing the scanned object. Correspondingly, the obtained two-dimensional projection image can reflect the details and smooth areas of the scanned object, and can distinguish the background area from the target area containing the scanned object. The background noise is low, which can improve the quality of the generated two-dimensional projection image. Subsequent When the two-dimensional projection map is applied to different scenes, the application effect of the image is improved.

Referring to Figure 12, Figure 12 is a flow chart of an image generation method provided by an embodiment of the present application.

The electronic device can obtain the 3D point cloud, which is the three-dimensional point cloud image in the embodiment of the present application, and determine the plane under the preset projection perspective as the preset projection plane. If the three-dimensional point cloud image matches the preset projection perspective, for example, the preset projection plane is parallel to a coordinate plane in the three-dimensional coordinate system of the three-dimensional point cloud image, the electronic device can directly calculate the neighborhood statistical features, that is, according to different features Statistical method, calculate different image features of preset three-dimensional points respectively. If the three-dimensional point cloud image does not match the preset projection perspective, for example, the preset projection plane is not parallel to all coordinate planes in the three-dimensional coordinate system of the three-dimensional point cloud image, the electronic device can perform three-dimensional rotation, that is, the electronic device is based on the preset projection perspective Coordinate conversion is performed on the three-dimensional point cloud image so that after the coordinate conversion, a coordinate plane in the three-dimensional coordinate system of the three-dimensional point cloud image is parallel to the preset projection plane, thereby obtaining a three-dimensional point cloud image that matches the preset projection perspective. The electronic device can then perform neighborhood statistical feature calculations, that is, calculate different image features of preset three-dimensional points according to different feature statistical methods.

Furthermore, the electronic device can perform projection, that is, based on the different image characteristics of the preset three-dimensional point, determine the different projection characteristics of the preset three-dimensional point on the preset projection plane, and based on the different projection characteristics, generate the corresponding projection characteristics. Projected subgraph. Furthermore, the electronic device can perform fusion processing on each projection sub-image to obtain a two-dimensional projection image of the three-dimensional point cloud image on the preset projection plane.

Referring to Figure 13, Figure 13 is a flow chart of an image generation method provided by an embodiment of the present application.

The electronic device can obtain a 3D point cloud, which is a three-dimensional point cloud image in the embodiment of the present application. Then, the electronic device can calculate different image features of the preset three-dimensional points in the three-dimensional point cloud image according to different feature statistics methods. The image characteristics of a preset three-dimensional point are: According to different feature statistics methods, the calculated image characteristics of the preset three-dimensional point are: Assume that the three-dimensional point is the statistical characteristic of the amplitude of each three-dimensional point in the neighborhood of the central point. The statistical characteristics of the preset three-dimensional point include: statistical characteristics 1, statistical characteristics 2,..., statistical characteristics n. Each statistical characteristic corresponds to a characteristic statistical method. n represents the number of characteristic statistical methods, and n is greater than or equal to 1. . Feature statistical methods can include: mode calculation, maximum value calculation, variance calculation, signal to noise ratio calculation, entropy value calculation, convolution kernel operator and custom function, etc.

Then, for each image feature of the preset three-dimensional point, the electronic device determines the projection feature of the preset three-dimensional point corresponding to the image feature on the preset projection plane based on the image feature, and generates the projection feature based on each projection feature. The projected subgraph corresponding to the projected feature. The generated projection sub-images include: projection image 1, projection image 2,..., projection image n. The electronic device can fuse each projection sub-image (i.e., projection image 1, projection image 2,..., projection image n) to obtain a front view of the scanned object. The front view of the scanned object is also a three-dimensional point cloud image in the main viewing direction. 2D projection diagram.

In order to more clearly illustrate the technical effects of the image generation method provided by the embodiments of the present application, through the comparison of the two-dimensional projection diagrams shown in Figures 14 to 22, the image generation method provided by the embodiments of the present application and the image generation of related technologies are compared. methods for comparison. The image generation method in the related art is the maximum projection method. The maximum projection method uses the following formula to determine the projection characteristics of a preset three-dimensional point in the preset straight line direction projected on the preset projection plane:

σ ₄ =max(σ(x,y,x)) (5)

σ ₄ represents the projection characteristics of the preset three-dimensional point in the preset straight line direction determined based on the maximum projection method on the preset projection plane, and σ (x, y, x) represents the preset three-dimensional point in the preset straight line direction. In the three-dimensional point cloud map, max() represents the maximum value function.

The image on the left in Figures 14 to 22 is a two-dimensional projection image obtained based on the image generation method in the related art, and the image on the right is a two-dimensional projection image obtained based on the image generation method provided by the embodiment of the present application. Figures 15, 16 and 17 are multiple sets of enlarged views of local image areas. Each set of enlarged views is an enlarged view of the local image area corresponding to the same position in the two two-dimensional projections in Figure 14.

The two-dimensional projection on the right in Figure 14 has clearer edges in different areas than the two-dimensional projection on the left, and the two-dimensional projection on the right in Figure 14 contains more edges than the two-dimensional projection on the left. Image details.

The image in Figure 15 is an enlarged view of the background area in the two-dimensional projection image. It can be seen that the background area in the image on the right in Figure 15 is darker than the background area in the image on the left, that is, the background area in the image on the right has a lower amplitude than the background area in the image on the left.

The image in Figure 16 is an enlarged view of the area where the character's knees are located in the two-dimensional projection. It can be seen that the edge of the knee pad in the image on the right side of Figure 16 is clearer than that in the image on the left side.

The scanning object in Figures 17 and 18 is a wrench. The image on the right in Figures 17 and 18 has a clearer edge of the wrench than the image on the left. Moreover, the image on the right side of Figure 17 contains more image details than the image on the left side. For example, in the image on the right side of Figure 17, the details of the head of the wrench, the details of the handle, etc. can be clearly observed.

The scanning objects in Figures 19 and 20 are mobile phones and headphone boxes. In Figures 19 and 20, the edges of the mobile phone and headphone box are clearer in the images on the right than in the images on the left. Moreover, the image on the right side of Figure 20 contains more image details than the image on the left side. For example, it can be clearly observed in the image on the right side of Figure 20 that the scanned object includes a mobile phone.

The scanned object in Figure 21 is plasticine. The echo signal of the millimeter wave signal reflected from the plasticine is weaker than the echo signal of the millimeter wave signal reflected from the human body. The image on the right in Figure 21 has a clearer outline of the scanned object than the image on the left, and the contrast between light and dark is more obvious.

The scanning object in Figure 22 is the tool. The edge of the tool in the image on the right side of Figure 22 is clearer than the image on the left side, and the image on the right side of Figure 22 contains more image details than the image on the left side.

It can be seen that compared with the two-dimensional projection image generated by the image generation method based on related technologies, the edges of different areas in the two-dimensional projection image generated based on the image generation method provided by the embodiment of the present application are clearer, that is, the outline of the scanned object is clearer. And it contains more image details, and can clearly distinguish different areas, and the background noise is low, which can improve the quality of the generated two-dimensional projection map. When the subsequent two-dimensional projection map is applied to different scenes, the application effect of the image is improved. .

Corresponding to the method embodiment of Figure 1, see Figure 23, which is a structural diagram of an image generation device provided by an embodiment of the present application. Contents not explained in detail in the following embodiments can refer to the description of the above embodiments. As shown in Figure 23, the image generation device includes:

The three-dimensional point cloud image acquisition module 2301 is used to obtain the three-dimensional point cloud image;

The image feature acquisition module 2302 is used to determine the statistical features corresponding to the amplitudes of the three-dimensional points in the neighborhood of the preset three-dimensional points in the three-dimensional point cloud image, and determine the image of the preset three-dimensional points based on the statistical features. Features; wherein, the amplitude of the three-dimensional point is used to characterize the electromagnetic scattering characteristics at the corresponding position of the three-dimensional point, and the neighborhood of the preset three-dimensional point refers to the preset point cloud area including the preset three-dimensional point;

The projection feature acquisition module 2303 is used to determine the projection features of the preset three-dimensional point in the preset straight line direction on the preset projection plane according to the image features of the preset three-dimensional point in the preset straight line direction; wherein, the The preset straight line direction is perpendicular to the preset projection plane;

The two-dimensional projection image acquisition module 2304 is used to determine the two-dimensional projection image of the three-dimensional point cloud image on the preset projection plane according to the projection characteristics on the preset projection plane.

In an optional embodiment, the image features of the preset three-dimensional point include at least one of the following: a first image feature used to participate in characterizing image details, a second image feature used to participate in determining smooth areas of the image, and The third image feature participates in characterizing the image background noise.

In an optional embodiment, the image feature acquisition module 2302 is specifically used for at least one of the following:

Through at least one characteristic statistical method among variance calculation, gradient calculation, and Laplacian operator, the first value corresponding to the amplitude of the three-dimensional point in the first neighborhood of the preset three-dimensional point in the three-dimensional point cloud image is determined. Statistical features, as the first image features of the preset three-dimensional point; wherein the plane where the first neighborhood is located is parallel to the preset projection plane;

Determine the amplitude of the three-dimensional point in the first neighborhood of the preset three-dimensional point in the three-dimensional point cloud image through at least one characteristic statistical method among entropy value calculation, integrated side-lobe ratio calculation, and peak side-lobe ratio calculation. The second statistical feature corresponding to the value is used as the second image feature of the preset three-dimensional point;

Determine the third value corresponding to the amplitude of the three-dimensional point in the second neighborhood of the preset three-dimensional point in the three-dimensional point cloud image through at least one characteristic statistical method among variance calculation, gradient calculation, or Laplacian operator. Statistical features, as the third image features of the preset three-dimensional point; wherein the plane where the second neighborhood is located is perpendicular to the preset projection plane.

In an optional embodiment, the projection feature acquisition module 2303 is specifically used for at least one of the following:

Determine the feature with the largest median value among the first image features of the preset three-dimensional point in the preset straight line direction as the first projection feature of the preset three-dimensional point in the preset straight line direction on the preset projection plane; The feature with the largest median value of the second image features of the preset three-dimensional point in the preset straight line direction is determined as the second projection feature of the preset three-dimensional point in the preset straight line direction on the preset projection plane;

Determine the third image feature of the preset three-dimensional point in the preset straight line direction as the third projection feature of the preset three-dimensional point in the preset straight line direction on the preset projection plane;

The two-dimensional projection image acquisition module 2304 is specifically configured to determine where the three-dimensional point cloud image is based on at least one of the first projection feature, the second projection feature, and the third projection feature on the preset projection plane. The two-dimensional projection image on the preset projection plane.

In an optional embodiment, the two-dimensional projection image acquisition module 2304 is specifically used to:

Determine the first pixel value of the preset three-dimensional point corresponding to the first projection feature on the preset projection plane in the three-dimensional point cloud image, and generate a first projection sub-image based on the first pixel value;

Use the third projection feature on the preset projection plane as a third pixel value, and generate a third projection sub-image based on the third pixel value;

Fusion processing is performed on at least two of the first projection sub-image, the second projection sub-image and the third projection sub-image to obtain a two-dimensional projection image corresponding to the three-dimensional point cloud image.

Using a preset fusion algorithm, process the first projection sub-image and the second projection sub-image to obtain a first intermediate image;

Perform normalization processing and grayscale transformation processing on the third projection subimage to obtain a second intermediate image; wherein the grayscale transformation process is used to adjust the relationship between the background area and the target area on the third projection subimage. The difference in pixel values between

A product of pixel values corresponding to the same position on the first intermediate image and the second intermediate image is determined, and a two-dimensional projection image corresponding to the three-dimensional point cloud image is generated based on the product.

In an optional embodiment, the three-dimensional point cloud image acquisition module 2301 is specifically used to acquire a three-dimensional point cloud image of the scanned object;

The preset projection plane includes: a projection plane parallel to the front and/or back of the scanned object.

In an optional embodiment, the three-dimensional point cloud image acquisition module 2301 is specifically configured to rotate the original three-dimensional point cloud image based on a preset projection perspective to obtain a three-dimensional point cloud image that matches the preset projection perspective;

The preset projection plane is the projection plane under the preset projection viewing angle.

Based on the image generation device provided by the embodiment of the present application, the statistical characteristics corresponding to the amplitudes of the three-dimensional points in the neighborhood of the preset three-dimensional points in the three-dimensional point cloud image can be determined, and based on the determined statistical characteristics, the preset three-dimensional points can be determined. Image characteristics, and determine the two-dimensional projection image of the three-dimensional point cloud image on the preset projection plane based on the image characteristics of the preset three-dimensional point in the preset straight line direction, that is, the two-dimensional image corresponding to the three-dimensional point cloud image can be generated. Moreover, different statistical features can reflect different image features. Correspondingly, for different application scenarios, different statistical features can be used based on actual needs, and the generated two-dimensional image can reflect the corresponding image features to adapt to the current situation. Application scenarios, that is, embodiments of the present application can improve the quality of the projected two-dimensional image, thereby improving the application effect of the image.

The embodiment of the present application also provides an electronic device, as shown in Figure 24, including a processor 2401, a communication interface 2402, a memory 2403, and a communication bus 2404. The processor 2401, the communication interface 2402, and the memory 2403 communicate through the communication bus 2404. complete mutual communication,

Memory 2403, used to store computer programs;

The processor 2401 is used to implement the following steps when executing the program stored in the memory 2403:

Obtain a three-dimensional point cloud image;

The communication bus mentioned in the above-mentioned electronic equipment can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The communication bus can be divided into address bus, data bus, control bus, etc. For ease of presentation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.

The communication interface is used for communication between the above-mentioned electronic devices and other devices.

The memory may include random access memory (Random Access Memory, RAM) or non-volatile memory (Non-Volatile Memory, NVM), such as at least one disk memory. In an optional embodiment, the memory may also be at least one storage device located remotely from the aforementioned processor.

The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (Network Processor, NP), etc.; it can also be a digital signal processor (Digital Signal Processor, DSP), special integrated Circuit (Application Specific Integrated Circuit, ASIC), Field-Programmable Gate Array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

In yet another embodiment provided by this application, a non-transitory computer-readable storage medium is also provided. The non-transitory computer-readable storage medium stores a computer program, and the computer program is implemented when executed by a processor. The steps for any of the above image generation methods.

In yet another embodiment provided by this application, a computer program product containing instructions is also provided, which, when run on a computer, causes the computer to execute any of the image generation methods in the above embodiments.

In the above embodiments, it may be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, it may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a non-transitory computer-readable storage medium or transmitted from one non-transitory computer-readable storage medium to another. For example, the computer instructions may be transferred from a non-transitory computer-readable storage medium to another. A website, computer, server or data center transmits data to another website, computer, server or data center through wired (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, microwave, etc.) Make the transfer. The non-transitory computer-readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server or data center integrated with one or more available media. The available media may be magnetic media (eg, floppy disk, hard disk, magnetic tape), optical media (eg, DVD), or semiconductor media (eg, Solid State Disk (SSD)), etc.

It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that these entities or operations are mutually exclusive. any such actual relationship or sequence exists between them. Furthermore, the terms "comprises," "comprises," or any other variations thereof are intended to cover a non-exclusive inclusion such that a process, method, article, or apparatus that includes a list of elements includes not only those elements, but also those not expressly listed other elements, or elements inherent to the process, method, article or equipment. Without further limitation, an element defined by the statement "comprises a..." does not exclude the presence of additional identical elements in a process, method, article, or apparatus that includes the stated element.

Each embodiment in this specification is described in a related manner. The same and similar parts between the various embodiments can be referred to each other. Each embodiment focuses on its differences from other embodiments. In particular, the apparatus, electronic equipment, non-transitory computer-readable storage medium and computer program product embodiments are described simply because they are basically similar to the method embodiments. For relevant details, please refer to the partial description of the method embodiments. That’s it.

The above are only preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection of the present application. within the range.

Claims

An image generation method, the method includes:

Obtain a three-dimensional point cloud image;

Determine the statistical characteristics corresponding to the amplitude of the three-dimensional point in the neighborhood of the preset three-dimensional point in the three-dimensional point cloud image, and determine the image characteristics of the preset three-dimensional point based on the statistical characteristics; wherein, the three-dimensional point The amplitude of is used to characterize the electromagnetic scattering characteristics at the corresponding position of the three-dimensional point, and the neighborhood of the preset three-dimensional point refers to the preset point cloud area including the preset three-dimensional point;

According to the image characteristics of the preset three-dimensional point in the preset straight line direction, the projection characteristics of the preset three-dimensional point in the preset straight line direction on the preset projection plane are determined; wherein the preset straight line direction is the same as the preset straight line direction. The default projection plane is vertical;

According to the projection characteristics on the preset projection plane, a two-dimensional projection image of the three-dimensional point cloud image on the preset projection plane is determined.
The method according to claim 1, wherein the image features of the preset three-dimensional point include at least one of the following: a first image feature used to participate in characterizing image details, a second image feature used to participate in determining a smooth area of the image , and the third image feature used to participate in characterizing the image background noise.
The method according to claim 2, wherein the statistical characteristics corresponding to the amplitudes of the three-dimensional points in the neighborhood of the preset three-dimensional points in the three-dimensional point cloud image are determined, and the predetermined three-dimensional points are determined based on the statistical characteristics. Assume that the image characteristics of a three-dimensional point include at least one of the following:

Through at least one characteristic statistical method among variance calculation, gradient calculation, and Laplacian operator, the first value corresponding to the amplitude of the three-dimensional point in the first neighborhood of the preset three-dimensional point in the three-dimensional point cloud image is determined. Statistical features, as the first image features of the preset three-dimensional point; wherein the plane where the first neighborhood is located is parallel to the preset projection plane;

Determine the amplitude of the three-dimensional point in the first neighborhood of the preset three-dimensional point in the three-dimensional point cloud image through at least one characteristic statistical method among entropy value calculation, integrated side-lobe ratio calculation, and peak side-lobe ratio calculation. The second statistical feature corresponding to the value is used as the second image feature of the preset three-dimensional point;

Determine the third value corresponding to the amplitude of the three-dimensional point in the second neighborhood of the preset three-dimensional point in the three-dimensional point cloud image through at least one characteristic statistical method among variance calculation, gradient calculation, or Laplacian operator. Statistical features, as the third image features of the preset three-dimensional point; wherein the plane where the second neighborhood is located is perpendicular to the preset projection plane.
The method according to claim 2, wherein the projection of the preset three-dimensional point in the preset straight line direction on the preset projection plane is determined based on the image characteristics of the preset three-dimensional point in the preset straight line direction. Characteristics, including at least one of the following:

Determine the feature with the largest median value among the first image features of the preset three-dimensional point in the preset straight line direction as the first projection feature of the preset three-dimensional point in the preset straight line direction on the preset projection plane;

Determine the feature with the largest median value among the second image features of the preset three-dimensional point in the preset straight line direction as the second projection feature of the preset three-dimensional point in the preset straight line direction on the preset projection plane;

Determine the third image feature of the preset three-dimensional point in the preset straight line direction as the third projection feature of the preset three-dimensional point in the preset straight line direction on the preset projection plane;

Determining the two-dimensional projection image of the three-dimensional point cloud image on the preset projection plane according to the projection characteristics on the preset projection plane includes:

Based on at least one of the first projection feature, the second projection feature and the third projection feature on the preset projection plane, a two-dimensional projection image of the three-dimensional point cloud image on the preset projection plane is determined.
The method according to claim 4, wherein the location of the three-dimensional point cloud image is determined based on at least one of the first projection feature, the second projection feature and the third projection feature on the preset projection plane. The two-dimensional projection image on the above-mentioned preset projection plane includes:

Determine the first pixel value of the preset three-dimensional point corresponding to the first projection feature on the preset projection plane in the three-dimensional point cloud image, and generate a first projection sub-image based on the first pixel value;

Determine the second pixel value of the preset three-dimensional point corresponding to the second projection feature on the preset projection plane in the three-dimensional point cloud image, and generate a second projection sub-image based on the second pixel value;

Use the third projection feature on the preset projection plane as a third pixel value, and generate a third projection sub-image based on the third pixel value;

Fusion processing is performed on at least two of the first projection sub-image, the second projection sub-image and the third projection sub-image to obtain a two-dimensional projection image corresponding to the three-dimensional point cloud image.
The method according to claim 5, wherein at least two of the first projection sub-image, the second projection sub-image and the third projection sub-image are fused to obtain the three-dimensional point cloud image. The corresponding two-dimensional projection images include:

Using a preset fusion algorithm, process the first projection sub-image and the second projection sub-image to obtain a first intermediate image;

Perform normalization processing and grayscale transformation processing on the third projection subimage to obtain a second intermediate image; wherein the grayscale transformation process is used to adjust the relationship between the background area and the target area on the third projection subimage. The difference in pixel values between

A product of pixel values corresponding to the same position on the first intermediate image and the second intermediate image is determined, and a two-dimensional projection image corresponding to the three-dimensional point cloud image is generated based on the product.
The method according to claim 1, wherein said obtaining a three-dimensional point cloud image includes:

Obtain a three-dimensional point cloud image of the scanned object;

The preset projection plane includes: a projection plane parallel to the front and/or back of the scanned object.
The method according to claim 1, wherein said obtaining a three-dimensional point cloud image includes:

Rotate the original three-dimensional point cloud image based on the preset projection perspective to obtain a three-dimensional point cloud image that matches the preset projection perspective;

The preset projection plane is the projection plane under the preset projection viewing angle.
An image generating device, the device includes:

The three-dimensional point cloud image acquisition module is used to obtain the three-dimensional point cloud image;

The image feature acquisition module is used to determine the statistical features corresponding to the amplitudes of the three-dimensional points in the neighborhood of the preset three-dimensional points in the three-dimensional point cloud image, and determine the image features of the preset three-dimensional points based on the statistical features. ; Wherein, the amplitude of the three-dimensional point is used to characterize the electromagnetic scattering characteristics at the corresponding position of the three-dimensional point, and the neighborhood of the preset three-dimensional point refers to the preset point cloud area including the preset three-dimensional point;

The projection feature acquisition module is used to determine the projection features of the preset three-dimensional point in the preset straight line direction on the preset projection plane according to the image features of the preset three-dimensional point in the preset straight line direction; wherein, the The preset straight line direction is perpendicular to the preset projection plane;

A two-dimensional projection image acquisition module is used to determine the two-dimensional projection image of the three-dimensional point cloud image on the preset projection plane according to the projection characteristics on the preset projection plane.
The device according to claim 9, wherein the image characteristics of the preset three-dimensional point include at least one of the following:

A first image feature is used to participate in characterizing image details, a second image feature is used to participate in determining smooth areas of the image, and a third image feature is used to participate in characterizing image background noise.
The device according to claim 10, wherein the image feature acquisition module is specifically used for at least one of the following:

Through at least one characteristic statistical method among variance calculation, gradient calculation, and Laplacian operator, the first value corresponding to the amplitude of the three-dimensional point in the first neighborhood of the preset three-dimensional point in the three-dimensional point cloud image is determined. Statistical features, as the first image features of the preset three-dimensional point; wherein the plane where the first neighborhood is located is parallel to the preset projection plane;

Determine the amplitude of the three-dimensional point in the first neighborhood of the preset three-dimensional point in the three-dimensional point cloud image through at least one characteristic statistical method among entropy value calculation, integrated side-lobe ratio calculation, and peak side-lobe ratio calculation. The second statistical feature corresponding to the value is used as the second image feature of the preset three-dimensional point;

Determine the third value corresponding to the amplitude of the three-dimensional point in the second neighborhood of the preset three-dimensional point in the three-dimensional point cloud image through at least one characteristic statistical method among variance calculation, gradient calculation, or Laplacian operator. Statistical features, as the third image features of the preset three-dimensional point; wherein the plane where the second neighborhood is located is perpendicular to the preset projection plane.
The device according to claim 10, wherein the projection feature acquisition module is specifically used for at least one of the following:

Determine the feature with the largest median value among the first image features of the preset three-dimensional point in the preset straight line direction as the first projection feature of the preset three-dimensional point in the preset straight line direction on the preset projection plane;

Determine the feature with the largest median value among the second image features of the preset three-dimensional point in the preset straight line direction as the second projection feature of the preset three-dimensional point in the preset straight line direction on the preset projection plane;

Determine the third image feature of the preset three-dimensional point in the preset straight line direction as the third projection feature of the preset three-dimensional point in the preset straight line direction on the preset projection plane;

The two-dimensional projection map acquisition module is specifically used for:

Based on at least one of the first projection feature, the second projection feature and the third projection feature on the preset projection plane, a two-dimensional projection image of the three-dimensional point cloud image on the preset projection plane is determined.
The device according to claim 12, wherein the two-dimensional projection image acquisition module is specifically used for:

Determine the first pixel value of the preset three-dimensional point corresponding to the first projection feature on the preset projection plane in the three-dimensional point cloud image, and generate a first projection sub-image based on the first pixel value;

Determine the second pixel value of the preset three-dimensional point corresponding to the second projection feature on the preset projection plane in the three-dimensional point cloud image, and generate a second projection sub-image based on the second pixel value;

Use the third projection feature on the preset projection plane as a third pixel value, and generate a third projection sub-image based on the third pixel value;

Fusion processing is performed on at least two of the first projection sub-image, the second projection sub-image and the third projection sub-image to obtain a two-dimensional projection image corresponding to the three-dimensional point cloud image.
The device according to claim 13, wherein the two-dimensional projection image acquisition module is specifically configured to use a preset fusion algorithm to process the first projection sub-image and the second projection sub-image to obtain first intermediate image;

Perform normalization processing and grayscale transformation processing on the third projection subimage to obtain a second intermediate image; wherein the grayscale transformation process is used to adjust the relationship between the background area and the target area on the third projection subimage. The difference in pixel values between

A product of pixel values corresponding to the same position on the first intermediate image and the second intermediate image is determined, and a two-dimensional projection image corresponding to the three-dimensional point cloud image is generated based on the product.
The device according to claim 9, wherein the three-dimensional point cloud image acquisition module is specifically used to acquire a three-dimensional point cloud image of the scanned object; the preset projection plane includes: parallel to the front and/or back of the scanned object. the projection plane.
The device according to claim 9, wherein the three-dimensional point cloud image acquisition module is specifically configured to rotate the original three-dimensional point cloud image based on a preset projection perspective to obtain a three-dimensional point cloud image that matches the preset projection perspective; The preset projection plane is the projection plane under the preset projection angle.
An electronic device, including a processor, a communication interface, a memory and a communication bus, wherein the processor, communication interface and memory complete communication with each other through the communication bus; the memory is used to store computer programs; the processor is used to execute the memory When the stored program is loaded on the computer, the method steps described in any one of claims 1-8 are implemented.
A non-transitory computer-readable storage medium. A computer program is stored in the non-transitory computer-readable storage medium. When the computer program is executed by a processor, the method steps described in any one of claims 1-8 are implemented.
A computer program product containing instructions that, when run on a computer, causes the computer to perform the steps of the method described in any one of claims 1-8.