CN109990662B - Automatic target scoring method, device, equipment and computer readable storage medium - Google Patents

Abstract

The invention discloses an automatic target reporting method, device, equipment and computer-readable storage medium. Wherein, the automatic target reporting method includes: acquiring a first captured image corresponding to a first moment and a second captured image corresponding to a second moment; wherein, the first moment is the moment before the shooting behavior occurs, and the second moment is after the shooting behavior occurs. time; using the bullet hole recognition model based on the training of the convolutional neural network, identify the number of bullet holes and the coordinates of each bullet hole in the first collected image and the second collected image respectively; If the number of bullet holes in the image is determined to have new bullet holes, then the coordinates of the new bullet holes are determined according to the coordinates of each bullet hole in the first captured image and the second captured image, and the shooting is determined according to the coordinates of the newly added bullet holes. score. According to the embodiment of the present invention, the shooting performance of the user can be quickly determined, and the statistical efficiency and security of the shooting performance can be improved.

Description

Automatic target reporting method, device, equipment and computer-readable storage medium

technical field

The invention belongs to the technology of shooting and reporting targets, and in particular relates to an automatic target reporting method, device, equipment and computer-readable storage medium.

Background technique

At present, in the training of live ammunition shooting, most of the traditional manual target reporting methods are used to count the shooting results, which results in a large workload and low efficiency for statistical shooting results, and the shooting results cannot be reported in time. In addition, due to the need for manual target reporting, the target reporting personnel often need to be at the shooting scene during the shooting training process, which makes the statistical security poor and may endanger the life safety of the target reporting personnel. Therefore, the traditional method of manual target reporting is difficult to meet the requirements of live ammunition training in modern military science and technology training.

With the development and progress of modern science and technology, automatic target reporting systems based on various sensors have appeared, such as double-layer electrode short-circuit sampling systems, acoustic and electrical positioning automatic target reporting systems, semiconductor electronic target systems and laser screen bullet point positioning systems. Although these automatic target reporting systems can improve statistical efficiency and safety compared with traditional manual target reporting systems, the hardware structure of these automatic target reporting systems is relatively complex and the cost is relatively high. Moreover, these automatic target reporting systems need to use special targets in shooting training to achieve automatic target reporting, which can only be applied to a single training subject and have poor adaptability.

SUMMARY OF THE INVENTION

Embodiments of the present invention provide an automatic target reporting method, device, device, and computer-readable storage medium, which can quickly determine a user's shooting performance based on the collected images of the target before and after the shooting behavior occurs, and improve the statistical efficiency and efficiency of shooting performance. safety.

On the one hand, the embodiment of the present invention provides a kind of automatic target reporting method, including:

acquiring the first captured image corresponding to the first moment and the second captured image corresponding to the second moment; wherein, the first moment is the moment before the shooting behavior occurs, and the second moment is the moment after the shooting behavior occurs;

Using the bullet hole recognition model trained based on the convolutional neural network, respectively identify the number of bullet holes and the coordinates of each bullet hole in the first captured image and the second captured image;

If it is determined that there are newly added bullet holes according to the number of bullet holes in the first collected image and the second collected image, the coordinates of the newly added bullet holes are determined according to the coordinates of each bullet hole in the first collected image and the second collected image, And the shooting results are determined according to the coordinates of the newly added bullet holes.

Further, before identifying the number of bullet holes and the coordinates of each bullet hole in the first captured image and the second captured image respectively, the method further includes:

Using the image segmentation model trained based on the convolutional segmentation network, identify the target surface image and the background image in the first captured image and the second captured image, respectively;

The background images in the first captured image and the second captured image are removed.

Further, it is determined that there are newly added bullet holes according to the number of bullet holes in the first collected image and the second collected image, including:

If it is determined that the number of bullet holes corresponding to the second captured image is greater than the number of bullet holes corresponding to the first captured image, it is determined that there are newly added bullet holes.

Further, it is determined that there are newly added bullet holes according to the number of bullet holes in the first collected image and the second collected image, including:

If it is determined that the number of bullet holes corresponding to the second acquisition image is greater than the number of bullet holes corresponding to the first acquisition image, acquire a third acquisition image corresponding to the third moment; wherein the third moment corresponds to the second moment;

Identify the number of bullet holes in the third captured image by using the bullet hole recognition model trained based on the convolutional neural network;

If it is determined that the number of bullet holes corresponding to the third captured image is greater than the number of bullet holes corresponding to the first captured image, it is determined that there are newly added bullet holes.

Further, according to the coordinates of each bullet hole in the first captured image and the second captured image, determine the coordinates of the newly added bullet holes, including:

The coordinates of each bullet hole corresponding to the second captured image are compared with the coordinates of each bullet hole corresponding to the first captured image, and it is determined that the coordinates of the bullet holes only appearing in the second captured image are the coordinates of the newly added bullet holes.

Further, before identifying the number of bullet holes and the coordinates of each bullet hole in the first captured image and the second captured image, the method further includes:

acquiring four corner points corresponding to the target area in the first captured image or the second captured image;

According to the four corner points corresponding to the target area and the four corner points corresponding to the standard target surface image, the image registration transformation matrix is calculated;

After determining the coordinates of the newly added bullet holes, it also includes:

According to the image registration transformation matrix, the coordinates of the newly added bullet holes are calibrated.

Further, it also includes:

If it is determined that there are no newly added bullet holes according to the number of bullet holes in the first captured image and the second captured image, the shooting score is determined to be zero.

On the other hand, an embodiment of the present invention provides an automatic target reporting device, the device comprising:

An image acquisition unit configured to acquire a first captured image corresponding to the first moment and a second captured image corresponding to the second moment; wherein the first moment is the moment before the shooting behavior occurs, and the second moment is the time after the shooting behavior occurs. time;

a bullet hole identification unit, configured to use a bullet hole identification model trained based on a convolutional neural network to identify the number of bullet holes and the coordinates of each bullet hole in the first captured image and the second captured image, respectively;

The achievement determination unit is configured to, if it is determined that there are new bullet holes according to the number of bullet holes in the first collected image and the second collected image, then according to the coordinates of each bullet hole in the first collected image and the second collected image, determine The coordinates of the new bullet holes are added, and the shooting results are determined according to the coordinates of the newly added bullet holes.

In another aspect, an embodiment of the present invention provides an automatic target reporting device, the device comprising: a processor and a memory storing computer program instructions;

The above-mentioned automatic target reporting method is realized when the processor executes the computer program instructions.

In another aspect, an embodiment of the present invention provides a computer-readable storage medium, where computer program instructions are stored on the computer-readable storage medium, and when the computer program instructions are executed by a processor, the above-mentioned automatic target reporting method is implemented.

The automatic target reporting method, device, device, and computer-readable storage medium according to the embodiments of the present invention can obtain the first and second captured images of the target before and after the shooting behavior occurs, and then use the convolutional neural network-based training to obtain the first and second captured images of the target. It identifies the number of bullet holes and the coordinates of each bullet hole in the first captured image and the second captured image respectively, and then determines the number of newly added bullet holes based on the number of identified bullet holes and the coordinates of each bullet hole. Coordinates, and finally determine the shooting results, so that the user's shooting results can be accurately and efficiently counted. At the same time, since the embodiment of the present invention does not require manual target reporting, the safety of the shooting field can be improved. In addition, since the embodiments of the present invention can collect images of targets and identify the collected images to count shooting results, there is no need to use special targets, various training subjects can be implemented, and the adaptability is strong. Therefore, the embodiments of the present invention can better meet the needs of modern military science and technology, and have wider application prospects.

Description of drawings

In order to illustrate the technical solutions of the embodiments of the present invention more clearly, the following briefly introduces the accompanying drawings that need to be used in the embodiments of the present invention. For those of ordinary skill in the art, without creative work, the Additional drawings can be obtained from these drawings.

Fig. 1 is the schematic flow sheet of the automatic target reporting method provided by an embodiment of the present invention;

2 is a schematic diagram of a moment provided by an embodiment of the present invention;

3 is a schematic flowchart of an automatic target reporting method provided by another embodiment of the present invention;

4 is a schematic flowchart of an automatic target reporting method provided by another embodiment of the present invention;

5 is a schematic structural diagram of an automatic target reporting device provided by an embodiment of the present invention;

FIG. 6 is a schematic diagram of a hardware structure of an automatic target reporting device provided by an embodiment of the present invention.

Detailed ways

The features and exemplary embodiments of various aspects of the present invention will be described in detail below. In order to make the objectives, technical solutions and advantages of the present invention more clear, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only configured to explain the present invention, and are not configured to limit the present invention. It will be apparent to those skilled in the art that the present invention may be practiced without some of these specific details. The following description of the embodiments is only intended to provide a better understanding of the present invention by illustrating examples of the invention.

It should be noted that, in this document, 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 any relationship between these entities or operations. any such actual relationship or sequence exists. Moreover, the terms "comprising", "comprising" or any other variation thereof are intended to encompass a non-exclusive inclusion such that a process, method, article or device that includes a list of elements includes not only those elements, but also includes not explicitly listed or other elements inherent to such a process, method, article or apparatus. Without further limitation, an element defined by the phrase "comprises" does not preclude the presence of additional identical elements in a process, method, article, or device that includes the element.

In order to solve the problems in the prior art, the embodiments of the present invention provide an automatic target reporting method, apparatus, device, and computer-readable storage medium. First, the automatic target reporting method provided by the embodiment of the present invention will be introduced below.

FIG. 1 shows a schematic flowchart of an automatic target reporting method provided by an embodiment of the present invention. As shown in Figure 1, the automatic target reporting method includes:

S110, acquiring the first captured image corresponding to the first moment and the second captured image corresponding to the second moment; wherein, the first moment is the moment before the shooting behavior occurs, and the second moment is the moment after the shooting behavior occurs;

S120, using the bullet hole identification model trained based on the convolutional neural network, respectively identify the number of bullet holes and the coordinates of each bullet hole in the first collected image and the second collected image;

S130. If it is determined according to the number of bullet holes in the first collected image and the second collected image that there are newly added bullet holes, then according to the coordinates of each bullet hole in the first collected image and the second collected image, determine the number of newly added bullet holes. Coordinates, and determine the shooting results according to the coordinates of the newly added bullet holes.

In the embodiment of the present invention, it is possible to obtain the first and second images of the target before and after the shooting behavior occurs, and then use the bullet hole recognition model trained based on the convolutional neural network to identify the first and second images respectively. Second, collect the number of bullet holes and the coordinates of each bullet hole in the image, and then determine the coordinates of the newly added bullet holes based on the number of bullet holes identified and the coordinates of each bullet hole, and finally determine the shooting score, so as to be accurate and efficient. Geostatistical user's shooting scores. At the same time, since the embodiment of the present invention does not require manual target reporting, the safety of the shooting field can be improved. In addition, since the embodiments of the present invention can collect images of targets and identify the collected images to count shooting results, there is no need to use special targets, various training subjects can be implemented, and the adaptability is strong. Therefore, the embodiments of the present invention can better meet the needs of modern military science and technology, and have wider application prospects.

The automatic target reporting method according to the embodiment of the present invention can be applied to a shooting score calculation device, for example, a computer device and the like.

In step S110 of the embodiment of the present invention, the first time is the time before the shooting behavior occurs, and the second time is the time after the shooting behavior occurs. It should be noted that the shooting behavior refers to the behavior of the user pulling the trigger once, that is, before and after each time the user pulls the trigger, a captured image must be collected to determine the shooting corresponding to each shooting behavior of the user. score.

In the embodiment of the present invention, an image acquisition device or a video acquisition device that communicates with the shooting score calculation device by wire or wireless may be set at a position opposite to the target to acquire the acquired image. Among them, the position opposite to the target refers to the position in front of the target face where the image of the target face can be collected. The image of the target surface can be collected.

In some embodiments of the present invention, an image acquisition device may be used to capture and acquire images according to a preset time node. The image acquisition device may be a camera. The preset time nodes include, before the user starts all shooting behaviors, take at least one shot to obtain a captured image with a blank target surface, and after the user completes one shooting behavior, take at least one shot to obtain the user completes the shot. Acquisition images after firing behavior. At this point, before the user completes the first shooting behavior, the acquired image with a blank target surface can be used as the first acquisition image corresponding to the first shooting behavior. After the user completes the first shooting behavior, the acquired image of the user The acquired image after the shooting behavior is completed can be used as the second acquired image corresponding to the first shooting behavior. Before the user completes the second shooting behavior, the acquired image after the user completes the first shooting behavior can be used as the first acquisition image corresponding to the second shooting behavior. The captured image after the shooting behavior can be used as the second captured image corresponding to the second shooting behavior. By analogy, the first captured image and the second captured image corresponding to each shooting behavior can be determined.

In other embodiments of the present invention, a video capture device may be used to capture a video image of the entire shooting process of the user, and the captured image may be acquired from the video image according to a preset time node. The image acquisition device may be a camera. The preset time nodes include, before the user starts all shooting behaviors, obtain the captured image with a blank target surface from the video image, and after the user completes a shooting behavior, obtain the user's completion of the shooting behavior from the video image from the video image. Post-acquisition images. At this point, before the user completes the first shooting behavior, the acquired image with a blank target surface can be used as the first acquisition image corresponding to the first shooting behavior. After the user completes the first shooting behavior, the acquired image of the user The acquired image after the shooting behavior is completed can be used as the second acquired image corresponding to the first shooting behavior. Before the user completes the second shooting behavior, the acquired image after the user completes the first shooting behavior can be used as the first acquisition image corresponding to the second shooting behavior. The captured image after the shooting behavior can be used as the second captured image corresponding to the second shooting behavior. By analogy, the first captured image and the second captured image corresponding to each shooting behavior can be determined.

Specifically, the preset time node can be set according to the firing frequency of the user. Taking the acquisition of captured images from video images as an example, the first captured image can be acquired at the 0th moment before the user starts shooting, and then one captured image is acquired from the 1st to the kth moment in turn according to the user's shooting frequency. Wherein, the time interval between every two moments when the captured images are acquired may be calculated according to the firing frequency of the user.

In this embodiment of the present invention, the first captured image that can be acquired at the 0th moment before the user starts shooting is an image of a blank target surface without bullet holes, which can be used as an initial reference for subsequent bullet hole identification.

Since the essence of automatic target reporting is to determine the position of the newly added bullet hole on the target surface after the user completes a shooting behavior, the target reporting work can only be performed when the position of the newly added bullet hole is accurately obtained. Therefore, in the embodiment of the present invention , bullet hole identification is the core link of the entire target reporting process.

In step S120 of the embodiment of the present invention, the bullet hole identification model is obtained by training based on a convolutional neural network. Specifically, a convolutional neural network for identifying the coordinates of each bullet hole and the number of bullet holes in the collected image can be built first. Then, collect a large number of images and videos collected at the same position relative to the target during shooting training, and collect images and videos under various environmental factors as much as possible, such as: wind and sand, rain and snow, different light environments, different The number of bullet holes and the coordinates of different bullet holes, etc. Wherein, if the captured image is captured by an image capturing device, subsequent processing can be performed directly; if it is a video image captured by a video capturing device, it needs to be converted into a static image before subsequent processing. Next, the position of each bullet hole in the image can be calibrated, and a part of the calibrated image can be used as a training image and the other part as a test image. Finally, the constructed convolutional neural network can be trained using the training images. After the training is completed, the bullet hole recognition model can be obtained, and the bullet hole recognition model can be tested through the test image. If the test is passed, the bullet hole recognition model can be used to Capture images for identification.

Therefore, in step S120 of this embodiment of the present invention, the bullet hole identification model trained by the above method can be used to identify bullet holes in the captured image, and the coordinates and the number of bullet holes in the captured image are given.

It should be noted that, in this embodiment of the present invention, the coordinates of the bullet hole refer to the coordinates of the center of the bullet hole, and the coordinates may be the x-coordinate and y-coordinate of the bullet hole in the two-dimensional plane coordinate system corresponding to the target surface. Moreover, the convolutional neural network built to train the bullet hole recognition model is a U-net network.

In the embodiment of the present invention, the bullet hole identification model based on the training of the convolutional neural network is used to identify the coordinates of the bullet holes and the number of bullet holes. Therefore, the bullet holes can be identified directly in the color collected images, without the need for Grayscale the image to retain all the valid information in the image. At the same time, in the process of bullet hole identification, various information in the collected images can be comprehensively used, instead of just using grayscale difference information to identify bullet holes, which can reduce the impact of uneven lighting in the external environment on bullet hole identification, and improve the performance of bullet holes. Adaptability of bullet hole identification methods to different environments. In addition, because the bullet hole identification model is used for bullet hole identification, there is no need to perform cumbersome image processing on the collected images, which can enhance the real-time performance of bullet hole identification, thereby improving the real-time performance of target reporting.

In some embodiments of the present invention, in step S130, the specific method for determining that there are new bullet holes according to the number of bullet holes in the first captured image and the second captured image may include: if determining the number of bullet holes corresponding to the second captured image If it is greater than the number of bullet holes corresponding to the first captured image, it is determined that there are new bullet holes.

That is, the number of bullet holes corresponding to the second captured image is compared with the number of bullet holes corresponding to the first captured image. If the number of bullet holes remains unchanged, it is determined that there are no new bullet holes. If the number of bullet holes corresponding to the second captured image is If the number is one more than the number of bullet holes corresponding to the first captured image, it can be determined that there are newly added bullet holes.

Since the bullet hole identification model may also be misjudged, that is, other objects are mistakenly judged as bullet holes, therefore, in other embodiments of the present invention, the judgment of whether there is a new bullet hole is also added to detect Confirmation mechanism, thereby reducing the misjudgment rate of bullet hole identification and improving the accuracy of new bullet hole identification in automatic target reporting.

Specifically, in these embodiments, the specific method for determining that there are newly added bullet holes according to the number of bullet holes in the first captured image and the second captured image in step S130 may include:

If it is determined that the number of bullet holes corresponding to the second acquisition image is greater than the number of bullet holes corresponding to the first acquisition image, acquire a third acquisition image corresponding to the third moment; wherein the third moment corresponds to the second moment;

Identify the number of bullet holes in the third captured image by using the bullet hole recognition model trained based on the convolutional neural network;

If it is determined that the number of bullet holes corresponding to the third captured image is greater than the number of bullet holes corresponding to the first captured image, it is determined that there are newly added bullet holes.

Specifically, take the time diagram shown in FIG. 2 as an example. First, the captured image obtained at time k is taken as the second captured image, and the captured image obtained at time k-1 is taken as the first captured image. After the coordinates of each bullet hole and the number of bullet holes, it is determined whether the number of bullet holes corresponding to the second captured image is greater than the number of bullet holes corresponding to the first captured image, so as to determine whether there are new bullet holes at time k. If the number of bullet holes corresponding to the second acquisition image is compared with the number of bullet holes corresponding to the first acquisition image, and the number of bullet holes remains unchanged, it is determined that there is no new bullet hole at time k; if the number of bullet holes is one more, then Immediately obtain the third captured image corresponding to the current moment (time t), and use the bullet hole recognition model to identify the number of bullet holes in the third captured image, and then determine whether the number of bullet holes corresponding to the third captured image is greater than that of the first captured image number of bullet holes. If the number of bullet holes corresponding to the third captured image is compared with the number of bullet holes corresponding to the first captured image, and the number of bullet holes remains unchanged, it is determined that there is no new bullet hole at time k; if the number of bullet holes is one more, it is determined that There are new bullet holes at time k.

In the embodiment of the present invention, the third time is a time after the execution of the same shooting behavior as the second time and before the completion of the next shooting behavior.

When the captured image is captured by the image capturing device, another captured image may be captured at the third moment as the third captured image. When a video image is captured by a video capture device, a frame of image corresponding to the third moment can be immediately obtained from the video data as the third captured image.

In step S130 of the embodiment of the present invention, a specific method for determining the coordinates of the newly added bullet holes according to the coordinates of the bullet holes in the first captured image and the second captured image may include:

The coordinates of each bullet hole corresponding to the second captured image are compared with the coordinates of each bullet hole corresponding to the first captured image, and it is determined that the coordinates of the bullet holes only appearing in the second captured image are the coordinates of the newly added bullet holes.

Specifically, the x-coordinates and y-coordinates of the coordinates of all the bullet holes in the first captured image can be summed up respectively, and the results are recorded as sumX1 and sumY1. Similarly, the x-coordinates of the coordinates of all the bullet holes in the second captured image can be summed up. Sum X and Y coordinates respectively, the results are recorded as sumX2 and sumY2, subtract sumX1 from sumX2, and subtract sumY1 from sumY2, then you can get the coordinates of the bullet holes that only appear in the second captured image, and add the x coordinates of the bullet holes. It is marked as Xnew, and the y coordinate is marked as Ynew. At this point, the coordinates (Xnew, Ynew) of the newly added bullet hole can be obtained.

In the embodiment of the present invention, determining the shooting result is determining the shooting ring value. The determination of the shooting ring value is the last step of automatic target reporting, the purpose is to determine which ring value range the bullet hole is in, so as to achieve the judgment of the shooting performance. The target lines in the target paper image are regular rings, and the radial distances between the target lines are equal. Therefore, the "bullet center" method can be used to determine the shooting ring value. This method is actually to calculate the distance between the bullet center and the bullseye, so as to use the relationship between the distance and the ring value to determine the shooting ring value. If the shooting ring value is 1, the shooting score is 1 point, if the shooting ring value is 2, the shooting score is 2 points, and so on, if the shooting ring value is 10, the shooting score is 10 points.

FIG. 3 shows a schematic flowchart of an automatic target reporting method provided by another embodiment of the present invention. As shown in Figure 3, the difference from the embodiment shown in Figure 1 is that before step S240, the automatic target reporting method also includes:

S220, using the image segmentation model obtained based on the training of the convolutional segmentation network, respectively identify the target surface image and the background image in the first acquisition image and the second acquisition image;

S230. Remove background images in the first captured image and the second captured image.

In step S220 of the embodiment of the present invention, the image segmentation model may be obtained by training a convolutional segmentation network. First, a convolutional segmentation network can be built to identify the target and background images in the acquired images. Then, you can collect a large number of target papers, collect images and video images during shooting training, and collect images and videos under various environmental factors as much as possible, such as: wind and sand, rain and snow, and different light environments. Wherein, if the captured image is captured by an image capturing device, subsequent processing can be performed directly; if it is a video image captured by a video capturing device, it needs to be converted into a static image before subsequent processing. The image is calibrated, the contour position information of the target surface image in the image is identified, and part of the calibrated image is used as a training image, and the other part is used as a test image. Finally, the training image can be input into the constructed convolutional segmentation network for deep training, and an image segmentation model can be obtained that can segment the target image in the collected image, and the image segmentation model can be tested through the test image. The image segmentation model can be used to identify the target surface image on the acquired image.

Therefore, in step S220 of the embodiment of the present invention, the image segmentation model trained by the above method may be used to identify the target surface image and the background image in the captured image. Then, in step S230 of the present invention, the background images identified in the first captured image and the second captured image are removed, and only the portion of the target surface image in the first captured image and the second captured image is retained for use in step S230 of the present invention. Bullet hole identification in the S240.

At this time, in the embodiment of the present invention, the automatic target reporting method also includes:

If it is determined according to the number of bullet holes in the first captured image and the second captured image that there are no newly added bullet holes, that is, a miss occurs, the shooting score is determined to be zero.

Among them, there are two types of misses in shooting: one is that the bullet deviates from the target, leaving no bullet holes on the target, and there is no bullet hole information in the collected image, and the score is zero; the other is that the bullet is on the target. Bullet holes are left on the surface, but the bullet holes are located outside the target surface area. Since in this embodiment, during the bullet hole identification, only the target surface image parts in the first and second acquisition images are retained, so when the bullet holes are identified Bullet holes located outside the target area will not be identified, resulting in a score of zero. In summary, the off-target situation can be well identified.

In the embodiment of the present invention, since it can be adapted to the extraction and segmentation of target surface images in a variety of different scenarios, the target surface images are extracted, and the background images in the collected images can be removed. Therefore, redundant background images other than the target surface images can be removed. Such as blue sky, white clouds, etc., in the subsequent bullet hole recognition process, the workload of image recognition processing can be reduced, and the interference of complex background images on bullet hole recognition can be avoided, and it has wide applicability. In addition, because in the embodiment of the present invention, the captured image for bullet hole identification is an image after the background image is removed, and the identification of bullet holes outside the target surface image is also avoided during bullet hole identification, which is a shooting process. The identification of the off-target situation in the judgment has laid an important foundation.

FIG. 4 shows a schematic flowchart of an automatic target reporting method provided by still another embodiment of the present invention. As shown in FIG. 4, different from the embodiment shown in FIG. 1, before step S340, it further includes:

S320, acquiring four corner points corresponding to the target area in the first captured image or the second captured image;

S330, according to the four corner points corresponding to the target area and the four corner points corresponding to the standard target surface image, calculate and obtain the image registration transformation matrix;

In step S350, after determining the coordinates of the newly added bullet holes, the method further includes:

According to the image registration transformation matrix, the coordinates of the newly added bullet holes are calibrated.

In this embodiment of the present invention, if the image capturing device or the video capturing device is not arranged in a position directly in front of the target surface, for example, the image capturing device or the video capturing device is arranged under the opposite side of the target, and is arranged from bottom to top to If the captured image is acquired by means of an elevation angle, the captured image will be compressed in the vertical direction, resulting in deformation of the captured captured image. Therefore, image registration is required.

In step S320 of the embodiment of the present invention, four corner points corresponding to the target area in the first captured image or the second captured image need to be acquired, where the target area may be the detection area of the target surface image. In addition, it is necessary to acquire the four corner points of the detection area of the standard target surface image. Specifically, a mask image corresponding to the target surface image in the acquired image can be obtained respectively, and four corner points of the detection area can be extracted, and a mask image of a standard target surface image can be obtained, and the four corner points of the detection area can be extracted. Thereby, four corner point pairs corresponding to one-to-one are obtained.

In step S330 of the embodiment of the present invention, using the four corner points corresponding to the target area in the acquired image and the four corner points corresponding to the standard target surface image, the spatial coordinate transformation parameters of the registration model of the two images can be obtained , these spatial coordinate transformation parameters can constitute the image registration transformation matrix. Among them, the adopted registration model is a projection transformation model.

In the actual shooting process, the main deformations of the target images collected by the automatic target reporting system are affine transformation and projection transformation. Among them, affine transformation is a special case of projective transformation.

In step S350 of the embodiment of the present invention, if it is determined that there are new bullet holes according to the number of bullet holes in the first collected image and the second collected image, then according to the number of bullet holes in the first collected image and the second collected image Coordinates, determine the coordinates of the newly added bullet holes, register the transformation matrix according to the image, calibrate the coordinates of the newly added bullet holes, and determine the shooting score according to the coordinates of the newly added bullet holes.

则有Therefore, when the coordinates of the newly added bullet holes are obtained, the image registration transformation matrix can be used to calculate the coordinates (Xnew, Ynew) of the newly added bullet holes to obtain the calibrated coordinates of the newly added bullet holes, denoted as (Xreg , Yreg), let the registration transformation matrix be

then there are

可以得到校准后的新增弹孔的坐标(Xreg,Yreg)，其计算公式为According to the calculated intermediate vector

The coordinates (Xreg, Yreg) of the newly calibrated bullet holes can be obtained, and the calculation formula is:

Using the above calculation formula, the coordinates (Xreg, Yreg) of the newly calibrated bullet holes can be calculated and used to determine the shooting score.

Therefore, the coordinates of the newly added bullet holes can be directly registered, thereby reducing the computational complexity of image processing.

It should be noted that, in this embodiment of the present invention, it is also possible to directly register the captured image, convert the captured image into an orthographic captured image, and then perform bullet hole identification after the registration of the captured image is achieved.

In the embodiment of the present invention, the coordinates of the bullseye and the distance between the target rings in the standard target surface image can be used to calculate the distance between the coordinates of the newly calibrated bullet holes and the coordinates of the bullseye, and then compare with the distance between the target rings to determine The shooting ring value of each bullet hole completes the judgment of the shooting score.

In the embodiment of the present invention, when the determination of the shooting score is completed, the coordinates of the bullet hole, the shooting ring value and the shooting score can be sent to each display terminal on the shooting range that is in wired or wireless communication with the shooting score computing device, and display The terminal can display the position of the bullet hole in the standard target surface image displayed by the display terminal in the form of a simulated image according to the coordinates of the bullet hole. At the same time, the display terminal can also display the shooting ring value and shooting performance, and can also broadcast the shooting ring value and shooting performance through voice.

To sum up, the embodiments of the present invention can solve the problems of large workload, low efficiency, poor safety, and inability to report shooting results in time in the existing shooting training, and avoid the influence of human factors, greatly improving the Shooting efficiency and fairness of shooting results. At the same time, in the embodiment of the present invention, only a common target can be used, which has great advantages in terms of low cost and environmental adaptability.

FIG. 5 shows a schematic structural diagram of an automatic target reporting device provided by an embodiment of the present invention. As shown in Figure 5, the automatic target reporting device includes:

The image acquisition unit 410 is configured to acquire the first captured image corresponding to the first moment and the second captured image corresponding to the second moment; wherein the first moment is the moment before the shooting behavior occurs, and the second moment is after the shooting behavior occurs. the moment;

The bullet hole identification unit 420 is configured to use a bullet hole identification model trained based on a convolutional neural network to identify the number of bullet holes and the coordinates of each bullet hole in the first captured image and the second captured image, respectively;

The achievement determination unit 430 is configured to, if it is determined according to the number of bullet holes in the first collected image and the second collected image that there are newly added bullet holes, then according to the coordinates of each bullet hole in the first collected image and the second collected image, Determine the coordinates of the newly added bullet holes, and determine the shooting results according to the coordinates of the newly added bullet holes.

In the embodiment of the present invention, the automatic target reporting device can obtain the first and second collected images of the target before and after the shooting behavior occurs, and then use the bullet hole recognition model trained based on the convolutional neural network to identify the first and second images respectively. The number of bullet holes and the coordinates of each bullet hole in the first captured image and the second captured image, and then based on the identified number of bullet holes and the coordinates of each bullet hole, determine the coordinates of the newly added bullet holes, and finally determine the shooting score, thus It can accurately and efficiently count the shooting performance of users. At the same time, since the embodiment of the present invention does not require manual target reporting, the safety of the shooting field can be improved. In addition, because the embodiments of the present invention can collect images of the target and identify the collected images to count shooting results, there is no need to use special targets, and can be applied to various training subjects with strong adaptability. Therefore, the embodiments of the present invention can better meet the needs of modern military science and technology, and have wider application prospects.

In the embodiment of the present invention, the automatic target reporting device may be applied to a shooting score calculation device, for example, a computer device and the like.

In this embodiment of the present invention, the first moment is the moment before the shooting behavior occurs, and the second moment is the moment after the shooting behavior occurs. It should be noted that the shooting behavior refers to the behavior of the user pulling the trigger once, that is, before and after each time the user pulls the trigger, a captured image must be collected to determine the shooting corresponding to each shooting behavior of the user. score.

In the embodiment of the present invention, an image acquisition device or a video acquisition device that communicates with the shooting score calculation device by wire or wireless may be set at a position opposite to the target to acquire the acquired image. Among them, the position opposite to the target refers to the position in front of the target face where the image of the target face can be collected. The image of the target surface can be collected.

In the embodiment of the present invention, the automatic target reporting device further includes an image processing unit, which is configured to: before recognizing the number of bullet holes and the coordinates of each bullet hole in the first captured image and the second captured image respectively, use the volume-based The image segmentation model obtained by integrating the segmentation network training, identify the target image and the background image in the first acquisition image and the second acquisition image respectively; remove the background image in the first acquisition image and the second acquisition image

In some embodiments of the present invention, the automatic target reporting device further includes a bullet hole analysis unit, which is configured to: if it is determined that the number of bullet holes corresponding to the second acquisition image is greater than the number of bullet holes corresponding to the first acquisition image, determine that the number of bullet holes corresponding to the second acquisition image is greater than the number of bullet holes corresponding to the first acquisition image. Added bullet holes.

In other embodiments of the present invention, the automatic target reporting device further includes a bullet hole analysis unit, which is configured to: if it is determined that the number of bullet holes corresponding to the second captured image is greater than the number of bullet holes corresponding to the first captured image, make the image The acquisition unit 410 acquires the third acquisition image corresponding to the third moment; wherein, the third moment corresponds to the second moment; then the bullet hole identification unit 420 uses the bullet hole identification model trained based on the convolutional neural network to identify the third time. The number of bullet holes in the captured image; finally, if it is determined that the number of bullet holes corresponding to the third captured image is greater than the number of bullet holes corresponding to the first captured image, it is determined that there are new bullet holes.

In this embodiment of the present invention, the achievement determination unit 430 is further configured to: compare the coordinates of each bullet hole corresponding to the second captured image with the coordinates of each bullet hole corresponding to the first captured image, and determine that only the second captured image The coordinates of the bullet holes appearing in are the coordinates of the newly added bullet holes.

In the embodiment of the present invention, the automatic target reporting device further includes an image registration unit and a data storage module, which are configured to: identify the number of bullet holes and the coordinates of each bullet hole in the first captured image and the second captured image, respectively Before, it also includes: acquiring four corner points corresponding to the target area in any captured image; calculating and obtaining an image registration transformation matrix according to the four corner points corresponding to the target area and the four corner points corresponding to the standard target surface image, The image registration transformation matrix is stored in the data storage module; after the coordinates of the newly added bullet holes are determined, the coordinates of the newly added bullet holes are calibrated according to the image registration transformation matrix.

In this embodiment of the present invention, the achievement determination unit 430 is further configured to: if it is determined that there are no newly added bullet holes according to the number of bullet holes in the first captured image and the second captured image, determine the shooting score as zero.

FIG. 6 shows a schematic diagram of a hardware structure of an automatic target reporting device provided by an embodiment of the present invention.

The automatic targeting device may include a processor 501 and a memory 502 storing computer program instructions.

Specifically, the above-mentioned processor 501 may include a central processing unit (CPU), or a specific integrated circuit (Application Specific Integrated Circuit, ASIC), or may be configured as one or more integrated circuits implementing the embodiments of the present invention.

Memory 502 may include mass storage for data or instructions. By way of example and not limitation, memory 502 may include a Hard Disk Drive (HDD), a floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, or Universal Serial Bus (USB) drive or two or more A combination of more than one of the above. Memory 502 may include removable or non-removable (or fixed) media, where appropriate. Storage 502 may be internal or external to the integrated gateway disaster recovery device, where appropriate. In certain embodiments, memory 502 is non-volatile solid state memory. In particular embodiments, memory 502 includes read only memory (ROM). Where appropriate, the ROM may be a mask programmed ROM, programmable ROM (PROM), erasable PROM (EPROM), electrically erasable PROM (EEPROM), electrically rewritable ROM (EAROM) or flash memory or A combination of two or more of the above.

The processor 501 reads and executes the computer program instructions stored in the memory 502 to implement any one of the automatic target reporting methods in the foregoing embodiments.

In one example, the automatic target reporting device may also include a communication interface 503 and a bus 510 . Among them, as shown in FIG. 6 , the processor 501 , the memory 502 , and the communication interface 503 are connected through the bus 510 and complete the mutual communication.

The communication interface 503 is mainly used to implement communication between modules, apparatuses, units and/or devices in the embodiments of the present invention.

The bus 510 includes hardware, software, or both, coupling the components of the automated targeting device to each other. By way of example and not limitation, the bus may include Accelerated Graphics Port (AGP) or other graphics bus, Enhanced Industry Standard Architecture (EISA) bus, Front Side Bus (FSB), HyperTransport (HT) Interconnect, Industry Standard Architecture (ISA) Bus, Infiniband Interconnect, Low Pin Count (LPC) Bus, Memory Bus, Micro Channel Architecture (MCA) Bus, Peripheral Component Interconnect (PCI) Bus, PCI-E5press (PCI-5) Bus, Serial Advanced Technology Attachment (SATA) bus, Video Electronics Standards Association Local (VLB) bus or other suitable bus or a combination of two or more of the above. Bus 510 may include one or more buses, where appropriate. Although embodiments of the present invention describe and illustrate a particular bus, the present invention contemplates any suitable bus or interconnect.

The automatic target reporting device can execute the automatic target reporting method in the embodiment of the present invention, thereby realizing the automatic target reporting method and device described in conjunction with FIG. 1 and FIG. 5 .

In addition, in combination with the automatic target reporting method in the above embodiment, the embodiment of the present invention may provide a computer storage medium for implementation. Computer program instructions are stored on the computer storage medium; when the computer program instructions are executed by the processor, any one of the automatic target reporting methods in the foregoing embodiments is implemented.

It is to be understood that the present invention is not limited to the specific arrangements and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above-described embodiments, several specific steps are described and shown as examples. However, the method process of the present invention is not limited to the specific steps described and shown, and those skilled in the art can make various changes, modifications and additions, or change the sequence of steps after comprehending the spirit of the present invention.

The functional blocks shown in the above-described structural block diagrams may be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, it may be, for example, an electronic circuit, an application specific integrated circuit (ASIC), suitable firmware, a plug-in, a function card, or the like. When implemented in software, elements of the invention are programs or code segments used to perform the required tasks. The program or code segments may be stored in a machine-readable medium or transmitted over a transmission medium or communication link by a data signal carried in a carrier wave. A "machine-readable medium" may include any medium that can store or transmit information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio frequency (RF) links, and the like. The code segments may be downloaded via a computer network such as the Internet, an intranet, or the like.

It should also be noted that the exemplary embodiments mentioned in the present invention describe some methods or systems based on a series of steps or devices. However, the present invention is not limited to the order of the above steps, that is, the steps may be performed in the order mentioned in the embodiments, or may be different from the order in the embodiments, or several steps may be performed simultaneously.

The above are only specific implementations of the present invention. Those skilled in the art can clearly understand that, for the convenience and brevity of the description, for the specific working process of the above-described systems, modules and units, reference may be made to the foregoing method embodiments. The corresponding process in , will not be repeated here. It should be understood that the protection scope of the present invention is not limited to this. Any person skilled in the art can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should all cover within the protection scope of the present invention.

Claims (8)

1. an automatic target reporting method, is characterized in that, comprises: Acquiring the first captured image corresponding to the first moment and the second captured image corresponding to the second moment; wherein, the first moment is the moment before the shooting behavior occurs, and the second moment is the moment after the shooting behavior occurs; Using a bullet hole identification model trained based on a convolutional neural network, identify the number of bullet holes and the coordinates of each bullet hole in the first captured image and the second captured image, respectively; If it is determined according to the number of bullet holes in the first collected image and the second collected image that there are newly added bullet holes, then according to the coordinates of each bullet hole in the first collected image and the second collected image, Determine the coordinates of the newly added bullet holes, and determine the shooting score according to the coordinates of the newly added bullet holes; If it is determined according to the number of bullet holes in the first captured image and the second captured image that there are no newly added bullet holes, the shooting score is determined to be zero; Before identifying the number of bullet holes and the coordinates of each bullet hole in the first captured image and the second captured image respectively, the method further includes: Using the image segmentation model trained based on the convolutional segmentation network, respectively identify the target surface image and the background image in the first acquisition image and the second acquisition image; The background images in the first captured image and the second captured image are removed. 2 . The automatic target reporting method according to claim 1 , wherein, determining that there are new bullet holes according to the number of bullet holes in the first captured image and the second captured image, comprising: 2 . If it is determined that the number of bullet holes corresponding to the second captured image is greater than the number of bullet holes corresponding to the first captured image, it is determined that the newly added bullet holes are present. 3. The automatic target reporting method according to claim 1, characterized in that, determining that there are newly added bullet holes according to the number of bullet holes in the first captured image and the second captured image, comprising: If it is determined that the number of bullet holes corresponding to the second captured image is greater than the number of bullet holes corresponding to the first captured image, a third captured image corresponding to the third moment is acquired; wherein the third moment and the second moment are Corresponding; Identify the number of bullet holes in the third captured image by using the bullet hole identification model trained based on the convolutional neural network; If it is determined that the number of bullet holes corresponding to the third captured image is greater than the number of bullet holes corresponding to the first captured image, it is determined that the newly added bullet holes are present. 4 . The automatic target reporting method according to claim 1 , wherein, according to the coordinates of each bullet hole in the first collected image and the second collected image, the size of the newly added bullet hole is determined. 5 . Coordinates, including: Comparing the coordinates of each bullet hole corresponding to the second captured image with the coordinates of each bullet hole corresponding to the first captured image, it is determined that the coordinates of the bullet hole only appearing in the second captured image are the Added the coordinates of bullet holes. 5. automatic target reporting method according to claim 1, is characterized in that: Before the respectively identifying the number of bullet holes and the coordinates of each bullet hole in the first captured image and the second captured image, the method further includes: acquiring four corner points corresponding to the target area in the first captured image or the second captured image; According to the four corner points corresponding to the target area and the four corner points corresponding to the standard target surface image, the image registration transformation matrix is obtained by calculating; After the determining the coordinates of the newly added bullet hole, the method further includes: The coordinates of the newly added bullet holes are calibrated according to the image registration transformation matrix. 6. an automatic target reporting device, is characterized in that, described device comprises: An image acquisition unit configured to acquire a first captured image corresponding to a first moment and a second captured image corresponding to a second moment; wherein the first moment is the moment before the shooting behavior occurs, and the second moment is the moment when the shooting behavior occurs the moment after the act of shooting; a bullet hole identification unit, configured to use a bullet hole identification model trained based on a convolutional neural network to identify the number of bullet holes and the coordinates of each bullet hole in the first captured image and the second captured image, respectively; The achievement determination unit is configured to, if it is determined according to the number of bullet holes in the first collection image and the second collection image that there are newly added bullet holes, then according to the first collection image and the second collection image The coordinates of each bullet hole in the newly added bullet hole are determined, and the shooting score is determined according to the coordinates of the newly added bullet hole; if the bullet holes in the first captured image and the second captured image are used If the number is determined to have no new bullet holes, the shooting score is determined to be zero; The apparatus further includes an image processing unit configured to perform training using a convolution-based segmentation network before identifying the number of bullet holes and the coordinates of each bullet hole in the first captured image and the second captured image, respectively. The obtained image segmentation model identifies the target surface image and the background image in the first captured image and the second captured image, respectively; The background images in the first captured image and the second captured image are removed. 7. An automatic target reporting device, characterized in that the device comprises: a processor and a memory storing computer program instructions; When the processor executes the computer program instructions, the automatic target reporting method according to any one of claims 1-5 is implemented. 8. A computer-readable storage medium, wherein computer program instructions are stored on the computer-readable storage medium, and when the computer program instructions are executed by a processor, any one of claims 1-5 is implemented. automatic target reporting method.

Images