CN118691947A - A control method for laser denial system - Google Patents

Abstract

The invention discloses a control method of a laser rejection system, and belongs to the technical field of laser rejection. A control method of a laser rejection system, comprising: collecting an image; identifying a target in the image by utilizing a pre-constructed YOLOv-tiny model; calculating the off-target amount of the target; and controlling a target refusing module in the laser refusing system to refuse the target according to the target-off amount of the target. According to the invention, the target is identified and tracked through the image identification technology, the target miss distance is calculated, and the target rejection module is controlled to reject according to the target miss distance, so that the automatic control of target discovery and rejection is realized, and the target rejection accuracy and efficiency are improved.

Description

A control method for laser denial system

Technical Field

The present invention belongs to the technical field of laser denial, and in particular relates to a control method of a laser denial system.

Background Art

The laser denial system is a system that uses laser technology to achieve specific denial functions, such as using laser technology to drive away or flash targets. Traditional laser denial systems rely on manual detection of targets and then operate corresponding equipment for denial, which has low accuracy and efficiency.

Summary of the invention

The purpose of the present invention is to overcome the deficiencies of the prior art and provide a control method for a laser denial system.

The objective of the present invention is achieved through the following technical solution: A control method for a laser denial system, comprising:

Collect images;

Identify the object in the image using a pre-built YOLOv3-tiny model;

Calculating the miss amount of the target;

The target denial module in the laser denial system is controlled according to the miss distance of the target to deny the target.

Furthermore, the YOLOv3-tiny model abandons the stage of generating candidate boxes and directly adopts a regression analysis method to obtain the position and confidence of the prediction box directly through the convolutional neural network operation in the input image.

Furthermore, the loss function of the YOLOv3-tiny model is:

in, represents the loss function of the YOLOv3-tiny model, represents the loss function of target confidence, represents the loss function of the target category, represents the position loss function, is the balance coefficient;

The loss function of target confidence is:

in, , represents the value of the intersection-over-union (IOU) between the predicted target bounding box and the true bounding box; is the predicted value, is the prediction confidence, , is the number of positive and negative samples;

The loss function for the target category is:

in, , represents the predicted target bounding box Does it exist in Class target; is the predicted value, is the target probability, , is the number of positive samples;

The position loss function is:

in, , , , , , , , , is the center offset of the predicted bounding box, is the aspect ratio of the predicted bounding box, is the number of positive samples.

Further, calculating the off-target amount of the target includes:

According to the coordinates of the center point of the target, the X-axis pixel offset and the Y-axis pixel offset of the target's coordinate position relative to the image center are calculated;

Calculate the current field of view angle based on the X-axis pixel offset and the Y-axis pixel offset, as well as the camera zoom factor;

Calculates the target's X and Y angular offsets based on the current field of view.

Furthermore, the laser denial system comprises:

The target recognition module is used to collect images, track and identify targets in the images, and calculate the target's miss distance;

Target Denial Module;

A control module, used for controlling the target rejection module to reject the target according to the miss distance of the target;

A power module, used to supply power to the target identification module, the target denial module and the control module;

Wherein, the target denial module includes a white light module, a green light module and an infrared module;

The white light module comprises a white laser light source and a first lens, the first lens comprises a first fixed lens group and a first movable lens group, the first fixed lens group and the first movable lens group are both arranged on one side of the light emitting surface of the white laser light source, the first movable lens group is arranged between the white laser light source and the first fixed lens group, the transmittance of the first lens to light within the white light spectrum range is greater than 80%, and the maximum stroke of the first movable lens group is 90 mm;

The green light module includes a green laser light source and a second lens, the second lens includes a second fixed lens group, a third fixed lens group, a second movable lens group and a third movable lens group, the second fixed lens group, the third fixed lens group, the second movable lens group and the third movable lens group are all arranged on one side of the light exit surface of the green laser light source, the second fixed lens group is arranged between the green laser light source and the third fixed lens group, the second movable lens group is arranged between the second fixed lens group and the third fixed lens group, the third movable lens group is arranged between the second movable lens group and the third fixed lens group, the transmittance of the second lens to light within the green light spectrum range is greater than 95%, the maximum aperture of the second lens is 45 mm, and the optical path focal length range of the second lens is 1 mm to 78 mm;

The infrared module includes an infrared laser light source and a third lens. The third lens includes a fourth fixed lens group, a fifth fixed lens group, a fourth movable lens group and a fifth movable lens group. The fourth fixed lens group, the fifth fixed lens group, the fourth movable lens group and the fifth movable lens group are all arranged on one side of the light exit surface of the infrared laser light source. The fourth fixed lens group is arranged between the infrared laser light source and the fifth fixed lens group. The fourth movable lens group is arranged between the fourth fixed lens group and the fifth fixed lens group. The fifth movable lens group is arranged between the fourth movable lens group and the fifth fixed lens group. The transmittance of the third lens to light within the infrared spectrum range is greater than 95%. The maximum light aperture of the third lens is 25 mm. The optical path focal length range of the third lens is 0.4 mm to 30 mm.

Furthermore, the target recognition module includes:

A camera, used to capture images;

An image tracking board is used to track and identify the target in the image and calculate the target's miss distance;

A floodlight is used to provide fill light for the camera.

Furthermore, the white light module also includes a deceleration linear motor with a built-in PI type displacement sensor, and the deceleration linear motor is used to drive the first movable lens group to move.

Furthermore, the green laser light source is manufactured by multi-core coupled optical fiber output, and the output characteristics are optical fiber core diameter 400um, NA0.22, and central wavelength 525±10nm.

Furthermore, the infrared laser light source is manufactured by adopting a multi-core coupled optical fiber output method, and the light output from multiple cores is coupled into a 400um optical fiber NA0.22 output through an optical path.

The beneficial effects of the present invention are:

(1) The present invention uses image recognition technology to identify and track the target, then calculates the target's miss distance, and controls the target denial module to perform denial according to the miss distance, thereby realizing the automatic control of target discovery and denial, and improving the accuracy and efficiency of target denial;

(2) The YOLOv3-tiny model in the present invention abandons the stage of generating candidate frames and directly adopts the regression analysis method, that is, the position and confidence of the prediction frame can be obtained directly through the convolutional neural network operation in the input image, thereby improving the detection speed of the target;

(3) The present invention improves the loss function of the YOLOv3-tiny model, achieves high-precision recognition of the target, and effectively balances the recognition speed and accuracy of the target.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a flow chart of a control method in the present invention;

FIG2 is a schematic diagram of the network structure of the YOLOv3-tiny model in the present invention;

FIG3 is a schematic diagram of a feature extraction network architecture in the present invention;

FIG4 is a schematic diagram of the prediction bounding box calculation in the present invention;

FIG5 is a schematic diagram of the parameter composition in the YOLOv3-tiny model;

FIG6 is a schematic diagram of calculating the coordinate offset of the identified target in the present invention;

FIG7 is a schematic diagram of a laser denial system;

FIG8 is a schematic diagram of the optical path of the white light module;

FIG9 is a schematic diagram of the optical path of the green light module;

FIG10 is a schematic diagram of the optical path of the infrared module;

In the figure, 1-first fixed lens group, 2-first movable lens group, 3-white laser light source, 4-second fixed lens group, 5-second movable lens group, 6-third movable lens group, 7-third fixed lens group, 8-fourth fixed lens group, 9-fourth movable lens group, 10-fifth movable lens group, 11-fifth fixed lens group.

DETAILED DESCRIPTION

The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

Referring to FIGS. 1 to 10 , the present invention provides a control method for a laser denial system:

As shown in FIG. 1 , a control method of a laser rejection system includes steps S100 to S400 .

Step S100: Capture images.

For example, an image is collected using an object recognition module.

Step S200: Using a pre-built YOLOv3-tiny model to identify the target in the image.

In some embodiments, the YOLOv3-tiny model abandons the stage of generating candidate boxes and directly adopts a regression analysis method to obtain the position and confidence of the predicted box directly through the convolutional neural network operation in the input image. This end-to-end detection method improves the detection speed of the target, so that the detection speed of the target reaches 30FPS.

The network structure of the YOLOv3-tiny model is shown in Figure 2. The backbone network used by YOLOv3-tiny in the present invention for feature extraction is a combination of the Darknet-19 network and the maximum pooling layer based on YOLOv2. This feature extraction network has a total of 7 convolutional layers, the size of the convolution kernel is 3×3, and the structure is shown in the figure. It can be seen from Figure 3 that feature extraction only requires 7 convolution operations, which reduces the amount of calculation and improves the detection speed. The backbone network also improves the utilization rate of the NPU through floating-point operations, thereby improving the training speed.

YOLOv3-tiny uses dimension clusters as anchor boxes to directly predict bounding boxes. YOLOv3-tiny only assigns a bounding box to each real object. If the bounding box does not correspond to the real object, only the target loss is calculated, and the coordinate and category losses are not calculated. The network predicts the four coordinate parameters of each bounding box as follows: If the grid is offset from the top left corner of the image by , and the width and height of the bounding box are , the prediction of the target bounding box is shown in Figure 4. The dotted rectangular box is the preset bounding box, and the solid rectangular box is the predicted bounding box calculated by the offset predicted by the network; the center offset of the bounding box predicted by the network is and aspect ratio , is the final predicted target bounding box. The conversion process from the preset bounding box to the final predicted bounding box is shown in the formula on the right, where Represents the sigmoid function, which maps the predicted offset between 0 and 1.

The YOLOv3-tiny network uses multi-scale prediction, predicting 3 bounding boxes at each scale. For the self-built vehicle head and tail dataset, the parameter size of each scale is ,in It represents the size of the feature layer for prediction. For example, the input image size of the YOLOv3-tiny model is 416×416. The sizes of the two scale feature maps after the convolution operation of the feature extraction network are 13×13 and 26×26 respectively. Each scale has 4 position offset parameters, 1 confidence score, and 2 target categories. The specific structure is shown in Figure 5.

Therefore, by predicting feature maps of three sizes on two feature layers, a total of six sizes will be obtained. The k-means clustering algorithm is used for the vehicle head orientation dataset to obtain these six anchors, which are (52× 45), (107×93), (169×148), (314×210), (250×272), and (343×322). They are then divided into the corresponding two feature layers. The specific classification is shown in Table 1.

Table 1 Bounding box classification

Feature Map size Bounding box size Number of bounding boxes Feature map 1 13×13 (314×210), (250×272), (343×322) 13×13×3 Feature map 3 26×26 (52×45), (107×93), (169×148) 26×26×3

In some embodiments, the loss function of the YOLOv3-tiny model is:

in, represents the loss function of the YOLOv3-tiny model, represents the loss function of target confidence, represents the loss function of the target category, represents the position loss function, is the balance coefficient;

The loss function of target confidence is:

in, , represents the value of the intersection-over-union (IOU) between the predicted target bounding box and the true bounding box; is the predicted value, is the prediction confidence, , is the number of positive and negative samples;

The loss function for the target category is:

in, , represents the predicted target bounding box Does it exist in Class target; is the predicted value, is the target probability, , is the number of positive samples;

The position loss function is:

in, , , , , , , , , is the center offset of the predicted bounding box, is the aspect ratio of the predicted bounding box, is the number of positive samples.

Step S300: Calculate the miss distance of the target.

In some embodiments, calculating the miss distance of the target includes: calculating the X-axis pixel offset and the Y-axis pixel offset of the coordinate position of the target relative to the image center according to the center point coordinates of the target; calculating the current field of view angle based on the X-axis pixel offset and the Y-axis pixel offset, and the zoom ratio of the camera; and calculating the angular offset of the target X-axis and Y-axis based on the current field of view angle.

Assuming that the current camera magnification is 30 times, the table shows that the field of view angle at this time is: horizontal 18.22° (≈0.32 radians = 19.2 arc minutes), vertical 10.3° (≈0.18 radians = 10.8 arc minutes), the camera target surface is 1920x1080 pixels, and the relationship between the pixel and the field of view angle at this time is calculated to be: horizontal: 19.2/1920 = 0.01 arc minutes/Pixel; vertical: 10.8/1080 = 0.01 arc minutes/Pixel. At this time, according to the vertical and horizontal cheap pixel values, the horizontal X-axis and vertical Y-axis offset angles of the real-time target offset can be calculated in real time. The schematic diagram of the pixel calculation angle deviation miss amount is shown in Figure 6.

Step S400: Controlling the target denial module in the laser denial system to deny the target according to the miss distance of the target.

For example, according to the miss distance of the target, the target rejection module is controlled to achieve strong light drive away, flashing, etc.

In some embodiments, as shown in FIG7 , the laser dazzle denial system includes a target recognition module, a target denial module, a control module and a power module. The target recognition module is used to collect images, track and identify targets in the images, and calculate the miss distance of the targets; the control module is used to control the target denial module to deny the targets according to the miss distance of the targets; the power module is used to supply power to the target recognition module, the target denial module and the control module. The laser dazzle denial system is mainly used for searchlighting, flashing, communication, etc. of sea surface targets.

The target rejection module includes a white light module, a green light module and an infrared module.

The white light module includes a white laser light source 3 and a first lens. As shown in FIG8 , the first lens includes a first fixed lens group 1 and a first movable lens group 2, wherein the first fixed lens group 1 and the first movable lens group 2 are both arranged on one side of the light-emitting surface of the white laser light source 3, and the first movable lens group 2 is arranged between the white laser light source 3 and the first fixed lens group 1, wherein the transmittance of the first lens to light within the white light spectrum range is greater than 80%, and the maximum stroke of the first movable lens group 2 is 90 mm. In this embodiment, the angle change is achieved by moving the first movable lens group 2 forward and backward, thereby achieving continuous zooming of the white light module.

The white laser light source 3 adopts a high-energy blue laser to excite high-temperature resistant phosphor solution, which has the advantages of high reliability, small light output angle, high cost performance, good environmental practicality, etc. The white laser light source 3 in this embodiment covers the entire white light spectrum band, and the central brightness reaches 1260cd/ ^mm2 . Through the cooperation of the white laser light source 3 and the first lens in this embodiment, the loss in the light propagation process is reduced, and the concentration of the light emitted by the white light module is improved, so that the central illumination of the white light module at 25m is greater than or equal to 28000 lux, and the variable beam divergence index of the lens is 1.2°~8°.

In some embodiments, a deceleration linear motor is used to realize the forward and backward movement of the first movable lens group 2, and a built-in position sensor PI is provided. When the position sensor PI signal is triggered, the address code of the deceleration linear motor is set to zero to define the starting position of the optical path zoom.

The green light module includes a green laser light source and a second lens. The second lens includes a second fixed lens group 4, a third fixed lens group 7, a second movable lens group 5 and a third movable lens group 6. The second fixed lens group 4, the third fixed lens group 7, the second movable lens group 5 and the third movable lens group 6 are all arranged on the light-emitting surface side of the green laser light source. The second fixed lens group 4 is arranged between the green laser light source and the third fixed lens group 7. The second movable lens group 5 is arranged between the second fixed lens group 4 and the third fixed lens group 7. The third movable lens group 6 is arranged between the second movable lens group 5 and the third fixed lens group 7. The transmittance of the second lens to the light within the green light spectrum range is greater than 95%. The maximum aperture of the second lens is 45mm, which reduces the scattering and absorption of light inside the lens and improves. In this embodiment, the continuous zoom of the green light module can be achieved by moving the second movable lens group 5 and the third movable lens group 6 forward and backward, and the focal length range of the optical path is 1mm to 78mm. From top to bottom in FIG. 9, the green light paths with focal lengths of 1mm, 45mm and 78mm correspond respectively.

The green laser light source is manufactured by multi-core coupled optical fiber output, and the output characteristics are optical fiber core diameter 400um, NA0.22, and central wavelength 525±10nm. On the one hand, a low-power green laser light source chip can be used to achieve high optical power output. On the other hand, after multiple cores are coupled, the light outputs of multiple laser cores interfere with each other, which is conducive to eliminating laser speckle and reducing the influence of laser speckle on lighting imaging.

The light output power of the green laser light source is 25W. Through the cooperation of the green laser light source and the second lens, the loss in the process of light propagation is reduced, and the concentration of the light emitted by the green light module is improved, so that the light output power of the green light module is greater than 18W, the central illumination at 25m is greater than or equal to 500,000 lux, and the peak illumination at 10 kilometers is greater than or equal to 1lux, so that the variable beam divergence index of the lens is 0.3~20°.

The infrared module includes an infrared laser light source and a third lens. The third lens includes a fourth fixed lens group 8, a fifth fixed lens group 11, a fourth movable lens group 9 and a fifth movable lens group 10. The fourth fixed lens group 8, the fifth fixed lens group 11, the fourth movable lens group 9 and the fifth movable lens group 10 are all arranged on the light-emitting surface side of the infrared laser light source. The fourth fixed lens group 8 is arranged between the infrared laser light source and the fifth fixed lens group 11. The fourth movable lens group 9 is arranged between the fourth fixed lens group 8 and the fifth fixed lens group 11. The fifth movable lens group 10 is arranged between the fourth movable lens group 9 and the fifth fixed lens group 11. The transmittance of the third lens to light within the infrared spectrum range is greater than 95%, and the maximum aperture of the second lens is 25 mm. In this embodiment, the continuous zoom of the infrared module can be achieved by moving the fourth movable lens group 9 and the fifth movable lens group 10 forward and backward. The focal length range of the optical path is 0.4 mm to 30 mm, so that the beam divergence index of the lens is 0.8 to 45°. FIG10 shows infrared light paths with focal lengths of 0.4 mm, 5 mm, and 30 mm from top to bottom.

The infrared laser light source is manufactured by adopting a multi-core coupled optical fiber output method, and the light output of multiple cores is coupled into a 400um optical fiber NA0.22 output through an optical path. On the one hand, a low-power infrared laser light source chip can be used to achieve a large optical power output. On the other hand, after the multiple cores are coupled, the light outputs of the multiple laser cores interfere with each other, which is beneficial to eliminate laser speckle and reduce the influence of laser speckle on lighting imaging.

The output power of the infrared laser light source is 20W. Through the cooperation of the infrared laser light source and the third lens, the loss in the process of light propagation is reduced, and the concentration of the light emitted by the infrared module is improved, so that the output power of the infrared module is greater than or equal to 15W. The angle of the infrared light night vision imaging beam is: 0.8°~45° (adjustable between the minimum angle and the maximum angle).

In some embodiments, the target recognition module includes a camera, an image tracking board and a floodlight. The camera includes a visible light camera and an infrared thermal imaging camera for collecting images; the image tracking board is used to track and identify the target in the image and calculate the target's miss distance; the floodlight is used to provide supplementary light for the camera.

The parameters of infrared thermal imaging cameras are based on the formula Determine, among which, represents the target size, Indicates the target recognition distance, represents the focal length of the optical system, Indicates the number of pixels occupied by the target. Indicates the pixel size (12um).

The number of pixels occupied by the target is based on the Johnson criterion. Specifically: Detection: A target is found in the field of view. At this time, the image formed by the target must occupy more than 1.5 pixels in the critical size direction; Recognition: The target can be classified, that is, the target can be identified as a tank, truck or person. At this time, the image formed by the target must occupy more than 6 pixels in the critical size direction; Identification: The model and other characteristics of the target can be distinguished, such as distinguishing between the enemy and the friend based on the appearance characteristics. This is when the image formed by the target must occupy more than 12 pixels in the critical size direction.

Due to the mirror reflection and refraction effects of the water surface, using white light for fill-in illumination in the near distance will make it difficult to form an image and observe with the naked eye. In this embodiment, the solution of using a floodlight for fill-in illumination in the near distance overcomes this problem.

In order to improve the compactness of the system, the floodlight shares the structural space with the infrared thermal imaging camera.

The above is only a preferred embodiment of the present invention. It should be understood that the present invention is not limited to the form disclosed herein, and should not be regarded as excluding other embodiments, but can be used in various other combinations, modifications and environments, and can be modified within the scope of the concept described herein through the above teachings or the technology or knowledge of the relevant field. The changes and modifications made by those skilled in the art shall not deviate from the spirit and scope of the present invention, and shall be within the scope of protection of the claims attached to the present invention.

Claims (9)

1. A control method of a laser rejection system, characterized by comprising:

Collecting an image;

Identifying a target in the image by utilizing a pre-constructed YOLOv-tiny model;

calculating the off-target amount of the target;

And controlling a target refusing module in the laser refusing system to refuse the target according to the target-off amount of the target.

2. The method according to claim 1, wherein the YOLOv-tini model omits a stage of generating a candidate frame, and a regression analysis method is directly adopted, so that the position and the confidence of the predicted frame are obtained in the input image directly through convolutional neural network operation.

3. The method for controlling a laser rejection system according to claim 1, wherein the YOLOv-tini model has a loss function of:

；

wherein, Representing the penalty function of the YOLOv-tiniy model,A loss function representing the confidence of the target,A loss function representing a class of objects is presented,The position loss function is represented as a function of position loss,Is a balance coefficient;

the loss function for target confidence is:

; wherein, A value representing the intersection ratio IOU of the prediction target bounding box and the real bounding box; in order to be able to predict the value, In order to predict the degree of confidence,，The number of positive and negative samples;

The loss function for the target class is:

; wherein, Representing a prediction target bounding boxWhether or not there isA class target; in order to be able to predict the value, For the target probability of being a target,，The number of positive samples;

The position loss function is:

；

wherein, ，，，，，，，，To predict the center offset of the bounding box,To predict the wide-to-high scaling of the bounding box,Is the number of positive samples.

4. The method of controlling a laser rejection system according to claim 1, wherein calculating the off-target amount of the target includes:

Calculating the X-axis pixel offset and the Y-axis pixel offset of the coordinate position of the target relative to the center of the image according to the center point coordinates of the target;

calculating a current field angle based on the X-axis pixel offset and the Y-axis pixel offset, and a zoom magnification of the camera;

The angular offsets of the target X-axis and Y-axis are calculated based on the current field angle.

5. The method of controlling a laser rejection system according to claim 1, wherein the laser rejection system comprises:

The target recognition module is used for collecting images, tracking and recognizing targets in the images and calculating the off-target quantity of the targets;

A target rejection module;

the control module is used for controlling the target refusing module to refuse the target according to the target-off amount of the target;

the power supply module is used for supplying power to the target identification module, the target refusing module and the control module;

The target rejection module comprises a white light module, a green light module and an infrared module;

The white light module comprises a white laser light source and a first lens, the first lens comprises a first fixed lens group and a first movable lens group, the first fixed lens group and the first movable lens group are arranged on one side of a light emitting surface of the white laser light source, the first movable lens group is arranged between the white laser light source and the first fixed lens group, the transmittance of the first lens to light rays in a white light spectrum range is greater than 80%, and the maximum travel of the first movable lens group is 90mm;

The green light module comprises a green laser light source and a second lens, the second lens comprises a second fixed lens group, a third fixed lens group, a second movable lens group and a third movable lens group, the second fixed lens group, the third fixed lens group, the second movable lens group and the third movable lens group are all arranged on one side of a light emitting surface of the green laser light source, the second fixed lens group is arranged between the green laser light source and the third fixed lens group, the second movable lens group is arranged between the second fixed lens group and the third fixed lens group, the third movable lens group is arranged between the second movable lens group and the third fixed lens group, the transmittance of the second lens to light rays in a green light spectrum range is greater than 95%, the maximum light transmitting caliber of the second lens is 45mm, and the optical path focal length range of the second lens is 1 mm-78 mm;

the infrared module comprises an infrared laser light source and a third lens, the third lens comprises a fourth fixed lens group, a fifth fixed lens group, a fourth movable lens group and a fifth movable lens group, the fourth fixed lens group, the fifth fixed lens group, the fourth movable lens group and the fifth movable lens group are all arranged on one side of a light emitting surface of the infrared laser light source, the fourth fixed lens group is arranged between the infrared laser light source and the fifth fixed lens group, the fourth movable lens group is arranged between the fourth fixed lens group and the fifth fixed lens group, the fifth movable lens group is arranged between the fourth movable lens group and the fifth fixed lens group, the transmittance of the third lens on light rays in an infrared spectrum range is greater than 95%, the maximum light transmitting caliber of the third lens is 25mm, and the optical path focal length range of the third lens is 0.4-30 mm.

6. The method for controlling a laser beam reject system according to claim 5, the target recognition module is characterized by comprising:

A camera for capturing images;

the image tracking plate is used for tracking and identifying the target in the image and calculating the off-target quantity of the target;

and the floodlight is used for supplementing light for the camera.

7. The method according to claim 5, wherein the white light module further comprises a deceleration linear motor with a PI-type displacement sensor, and the deceleration linear motor is used for driving the first moving lens group to move.

8. The method for controlling a laser rejection system according to claim 5, wherein the green laser source is manufactured by adopting a multi-die coupling optical fiber output mode, and the output characteristic is 400um, NA0.22 and 525+ -10 nm of central wavelength.

9. The method of claim 5, wherein the infrared laser source is manufactured by multi-die coupling optical fiber output, and the light emitted from the multiple dies is coupled into the 400um optical fiber NA0.22 via the optical path.