Optical phased array main lobe centroid identification method and system based on CBAM-CNN

Abstract

The invention discloses an optical phased array main lobe centroid identification method and system based on CBAM-CNN, and the method comprises the following steps: S1, adjusting an optical antenna unit, S2, adjusting the hardware exposure parameters of an image collection unit, S3, converting a gray value image into a binary image, S4, inputting the binary image into a CBAM module-based convolutional neural network, and S5, carrying out the recognition of the main lobe centroid of an optical phased array. S5, judging the position of the mass center of the light spot in the main lobe area through a method, and outputting the coordinate of the mass center of the main lobe area; and S6, judging whether the far-field light spot of the optical phased array exceeds the hardware acquisition view field of the image acquisition unit. According to the invention, the convolutional neural network combines a channel attention mechanism and a space attention mechanism, carries out the accurate image region recognition of an optical phased array main lobe region, an optical phased array side lobe region and a stray light region, converts a multi-light-spot recognition problem into a single-light-spot recognition problem in the optical phased array main lobe region, and achieves the recognition of a single light spot in the optical phased array main lobe region. And the optical phased array main lobe centroid can be accurately identified under the condition of multi-side-lobe interference.

Description

Technical Field

[0001] The present invention belongs to the technical field of space laser communication, and in particular relates to a method and system for identifying the centroid of a main lobe of an optical phased array based on CBAM-CNN. Background Art

[0002] Optical phased array technology has the advantages of fast response speed, high pointing accuracy, light weight and miniaturization. It has great development potential in the fields of free-space laser communication, solid-state lidar and long-distance three-dimensional imaging.

[0003] The existing methods for identifying the centroid of a light spot mainly rely on an image gray value threshold correlation algorithm to locate the centroid position of the light spot. For example, the invention patent with publication number CN111462225 A discloses a method for identifying and locating the centroid of an infrared light spot image, including: using the image of the infrared light spot centroid area to be identified and located as the image of the region of interest; performing deep conversion and grayscale linear stretching on the image of the region of interest to generate a preprocessed image; performing adaptive threshold calculation on the preprocessed image to obtain a binarization threshold of the preprocessed image; binarizing the preprocessed image, and finding the contour of the object in the area according to the contour tracking algorithm, and selecting the maximum contour point set as the light spot contour; and locating the infrared light spot centroid according to the light spot contour. For example, the invention patent with the publication number CN100371676C discloses a fast and high-precision spot image centroid positioning method and device, including: performing Gaussian convolution operation on the pixel gray value, and judging whether the pixel gray value after the Gaussian convolution operation is greater than a preset threshold value, if so, marking the current pixel, identifying the corresponding spot, and calculating the product of the current pixel gray value and coordinate value and the cumulative value of the product of all the processed pixel gray values ​​and coordinate values ​​of the same spot, the cumulative value of the current pixel gray value and the cumulative value of all the processed pixel gray values ​​of the same spot, saving the obtained cumulative value, otherwise marking the current pixel as a background pixel and processing; after processing the entire output image, calculating the quotient of the gray value and coordinate value product cumulative value and the gray value cumulative value, and outputting the calculation result as the coordinate value of the centroid of the spot image. The present invention also discloses a centroid positioning device, which can improve the data processing speed and noise resistance in the centroid positioning of the spot image, and can process multiple spot images.

[0004] The light spot formed by the optical phased array outgoing light beam in the far field includes the main lobe and side lobes. The presence of the optical phased array side lobes and stray light in the optical system will interfere with the identification of the centroid of the far-field light spot, resulting in the inability of the existing positioning algorithm that relies on image gray value threshold correlation to accurately identify the optical phased array far-field light spot. A method that can accurately identify the centroid of the optical phased array main lobe in the presence of multiple side lobes and stray light interference has become a problem that needs to be urgently solved in the engineering application of optical phased array technology. Summary of the invention

[0005] The purpose of the present invention is to provide a method and system for identifying the main lobe centroid of an optical phased array based on CBAM-CNN. A convolutional neural network (CNN) model based on a channel and spatial attention mechanism (CBAM) is used to perform regional recognition of the main lobe and side lobe distribution of the far-field light spot of the optical phased array, and the centroid of the light spot is located in the identified main lobe area of ​​the far-field light spot of the optical phased array, thereby solving the problem of accurately identifying the main lobe centroid of the optical phased array in the presence of multiple sidelobe interference.

[0006] To achieve the above object, the technical solution adopted by the present invention is:

[0007] A method for identifying the centroid of the main lobe of an optical phased array based on CBAM-CNN comprises the following steps:

[0008] S1. Adjust the optical antenna unit so that the far-field light spot of the optical phased array enters the field of view of the image acquisition unit, and adjust the optical antenna unit so that the main lobe of the far-field light spot of the optical phased array appears in the center of the field of view of the camera;

[0009] S2, adjusting the hardware exposure parameters of the image acquisition unit to display as many light spots as possible in its field of view without overexposure;

[0010] S3, the image processing unit reads the current gray value image of the image acquisition unit, and converts the acquired gray value image into a binary image by setting a gray value threshold;

[0011] S4, the image processing unit inputs the obtained binary image into the convolutional neural network based on the CBAM module, and outputs the main lobe area, the side lobe area and the stray light area of ​​the optical phased array based on the convolutional neural network of the CBAM module;

[0012] S5, the centroid identification unit uses the centroid method to determine the centroid position of the main lobe area spot, and outputs the centroid coordinates of the main lobe area of ​​the optical phased array far-field spot;

[0013] S6. The negative feedback adjustment unit determines whether the far-field spot of the optical phased array exceeds the hardware acquisition field of view of the image acquisition unit. If so, jump to step S1; if not, jump to step S3.

[0014] As a further optimization of the above method,

[0015] Preferably, in step S4, the recognition of the main lobe area and the side lobe area based on the CBAM module convolutional neural network can be achieved by the following steps:

[0016] S41, obtaining an optical phased array far-field spot binary image dataset for training a convolutional neural network;

[0017] S42, establishing a convolutional neural network based on the CBAM module, and training the neural network using an optical phased array far-field spot binary image data set to obtain a trained convolutional neural network based on the CBAM module;

[0018] S43, inputting the binary image of the far-field spot of the optical phased array to be identified into the trained convolutional neural network based on the CBAM module.

[0019] Furthermore, the CBAM-based convolutional neural network in step S42 is a multi-layer convolutional neural network, the first, third, and fifth layers in the multi-layer convolutional neural network are convolutional layers, the second, fourth, and sixth layers in the multi-layer convolutional neural network are CBAM layers, the seventh layer in the multi-layer convolutional neural network is a fully connected layer, and the eighth layer in the multi-layer convolutional neural network is a Softmax layer.

[0020] The CBAM layer consists of a channel attention module and a spatial attention module.

[0021] The channel attention module has the following process description formula:

[0022]

[0023] ①In the formula, M C (F) is the weight coefficient of the channel attention module of the intermediate feature F, σ is the Sigmoid function, MLP is a multi-layer neural network, AvgPool is average pooling, MaxPool is maximum pooling, W0 and W1 are the weights of the multi-layer neural network MLP, is the average pooling value of the intermediate feature F in the channel attention module, is the maximum pooling value of the intermediate feature F in the channel attention module.

[0024] The spatial attention module has the following process description formula:

[0025]

[0026] ②In the formula, M S (F) is the weight coefficient of the spatial attention module of the intermediate feature F, f 5×5 The convolution kernel size is 5×5, is the average pooling value of the intermediate feature F in the spatial attention module, is the maximum pooling value of the intermediate feature F in the spatial attention module.

[0027] The CBAM layer has the following process description formula:

[0028]

[0029] In formula ③, F′ is the intermediate quantity of the intermediate feature F after the channel attention module, and F″ is the output quantity of the intermediate feature F after the entire channel attention module and the spatial attention module.

[0030] An optical phased array main lobe centroid recognition system based on a CBAM module convolutional neural network includes an optical antenna unit, an image acquisition unit, an image processing unit, a centroid recognition unit, and a negative feedback regulation unit;

[0031] The optical antenna unit is used to receive the target light beam and transmit the light beam to the image acquisition unit. The relative position of the target light beam irradiated on the image acquisition unit can be changed by adjusting the optical antenna unit;

[0032] The image acquisition unit is used to acquire the spot image generated by the target light beam irradiating thereon, and the information such as the spot brightness of the acquired image can be changed by changing the parameters such as the hardware exposure time of the image acquisition unit;

[0033] The image processing unit is used to convert the collected light spot gray value image into a binary image, and realize image area recognition of the main lobe area of ​​the optical phased array, the side lobe area of ​​the optical phased array, and the stray light area through a convolutional neural network based on the CBAM module;

[0034] The centroid identification unit uses the centroid method to determine the centroid position of the light spot in the main lobe area of ​​the optical phased array, and outputs the centroid coordinates of the main lobe area of ​​the far-field light spot of the optical phased array;

[0035] The negative feedback adjustment unit is used to compare the deviation between the centroid coordinates of the main lobe area of ​​the far-field light spot of the optical phased array and the center coordinates of the field of view of the image acquisition unit, calculate the relevant control parameters of the optical antenna unit, and control the optical antenna unit to move the centroid of the light spot to the center of the field of view of the image acquisition unit.

[0036] Preferably, when the image processing unit performs region recognition, the convolutional neural network target detection method based on the CBAM module is a supervised machine learning method, and its training method is as follows:

[0037] (1) Establishing a Dataset

[0038] Collect binary images including the main lobe, side lobe and stray light spot of the optical phased array and establish a data set;

[0039] (2) Image Annotation

[0040] The data set is manually labeled to mark the main lobe area, side lobe area, and stray light area of ​​the optical phased array. The data set with the spot type and spot position marked is used for subsequent model training and evaluation;

[0041] (3) Model training

[0042] The convolutional neural network based on the CBAM module is a multi-layer convolutional neural network, the first, third and fifth layers in the multi-layer convolutional neural network are convolutional layers, the second, fourth and sixth layers in the multi-layer convolutional neural network are CBAM layers, the seventh layer in the multi-layer convolutional neural network is a fully connected layer, and the eighth layer in the multi-layer convolutional neural network is a Softmax layer.

[0043] The CBAM module enhances the ability of convolutional neural networks to extract features from different channels and positions by introducing channel attention modules and spatial attention modules.

[0044] (4) Result evaluation

[0045] The mean average precision (mAP) is used to evaluate the accuracy of the model's prediction results. The calculation formula is as follows:

[0046]

[0047] Where C represents the total number of categories, AP i Represents the AP value of the i-th category target.

[0048] Preferably, steps S1 to S6 are executed cyclically, and the centroid coordinates of the main lobe area of ​​the optical phased array far-field spot are output once each round, and the cycle ends when the number of cycles reaches a predetermined value.

[0049] Compared with the prior art, the present invention has the following advantages:

[0050] (1) The convolutional neural network of the present invention combines the channel attention mechanism and the spatial attention mechanism, and accurately recognizes the image area for the main lobe area, side lobe area and stray light area of ​​the optical phased array, converts the multi-spot recognition problem into a single-spot recognition problem in the main lobe area of ​​the optical phased array, and can accurately identify the centroid of the main lobe of the optical phased array in the presence of multiple sidelobe interferences, eliminates multiple sidelobe interferences, and only calculates the centroid coordinates of the main lobe spot, thus solving the problem of accurately identifying the centroid of the main lobe of the optical phased array in the presence of multiple sidelobe interferences.

[0051] (2) The present invention can further improve the accuracy of spot centroid identification by eliminating the interference of stray light on spot centroid identification through regional identification of stray light.

[0052] (3) The present invention introduces a negative feedback adjustment unit, which controls the optical antenna unit to move the centroid of the light spot to the center of the field of view of the image acquisition unit by comparing the deviation between the centroid coordinates of the main lobe area of ​​the far-field light spot of the optical phased array and the center coordinates of the field of view of the image acquisition unit. This ensures that the main lobe of the light spot is always near the center of the field of view, thereby ensuring the accuracy of the identification of the centroid of the light spot of the optical phased array.

[0053] (4) The present invention aims at the problem of identifying the centroid of multiple light spots formed by the linearly symmetrically arranged main lobe and multiple side lobes formed in the far field by the optical phased array output light beam. This solution can be used for any problem of identifying the centroid of multiple light spots that satisfies the requirement that the side lobes are symmetrically distributed relative to the center of the main lobe. It is independent of the specific technical system adopted by the optical phased array and can be applied to optical phased array devices under various technical systems such as optical waveguide optical phased array and liquid crystal optical phased array, and has a wide range of applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 This is a module block diagram of an optical phased array main lobe centroid identification system based on a CBAM module convolutional neural network of the present invention;

[0055] Figure 2 It is a schematic diagram of image partitioning in the field of view of the image acquisition unit of the present invention;

[0056] Figure 3 Schematic diagram of the channel attention module structure of the present invention;

[0057] Figure 4 Schematic diagram of the spatial attention module structure of the present invention;

[0058] Figure 5 It is a schematic diagram of the CBAM module structure of the present invention;

[0059] Figure 6 The figure is a flow chart of an embodiment of the present invention. DETAILED DESCRIPTION

[0060] In order to make the purpose and advantages of the present invention more clear, the present invention is described in detail below with reference to the accompanying drawings and embodiments.

[0061] like Figure 1 As shown, the present invention discloses an optical phased array main lobe centroid recognition system based on CBAM-CNN, including an optical antenna unit, an image acquisition unit, an image processing unit, a centroid recognition unit and a negative feedback regulation unit.

[0062] The optical antenna unit is responsible for receiving the target light beam and transmitting the light beam to the image acquisition unit. The relative position of the target light beam irradiated on the image acquisition unit can be changed by adjusting the optical antenna unit.

[0063] The image acquisition unit is responsible for acquiring the spot image generated by the target light beam irradiating it. The spot brightness and other information of the acquired image can be changed by changing parameters such as the exposure time of the image acquisition unit hardware. The image acquisition unit usually uses CMOS / CCD as its underlying hardware. Figure 2 Schematic diagram of image partitioning in the field of view of the image acquisition unit.

[0064] The image processing unit reads the spot grayscale value image currently acquired by the image acquisition unit, and converts the acquired grayscale value image into a binary image by setting a grayscale value threshold; the obtained binary image is input into a trained convolutional neural network based on the CBAM module, and the neural network outputs image area recognition results for the main lobe area of ​​the optical phased array, the side lobe area of ​​the optical phased array, and the stray light area.

[0065] The centroid recognition unit uses the centroid method to determine the centroid position of the light spot in the binary image, and outputs the centroid coordinates of the main lobe area of ​​the optical phased array far-field light spot. After converting the multi-spot centroid recognition problem into a single-spot centroid recognition problem, the solution can use the centroid method to determine the spot centroid, or it can use the geometric method, centroid method, etc. to determine the spot centroid parameters according to actual application requirements.

[0066] The negative feedback adjustment unit compares the deviation between the coordinates of the centroid of the main lobe area of ​​the far-field spot of the optical phased array and the coordinates of the center of the field of view of the image acquisition unit, calculates the relevant control parameters of the optical antenna unit, and controls the optical antenna unit to move the centroid of the spot to the center of the field of view of the image acquisition unit.

[0067] Furthermore, in the image processing unit region recognition strategy, the convolutional neural network target detection method based on the CBAM module is a supervised machine learning method, and its training method is as follows:

[0068] (1) Establishing a Dataset

[0069] Collect binary images containing the main lobe, side lobe and stray light spot of the optical phased array and establish a data set. Since the pixel value in the binary image is 0 or 1, there is no need to perform image normalization.

[0070] (2) Image Annotation

[0071] The data set is manually labeled to mark the optical phased array main lobe area, optical phased array side lobe area, and stray light area. The data set with the spot type and spot position labeled is used for subsequent model training and evaluation.

[0072] (3) Model training

[0073] The convolutional neural network based on the CBAM module is a multi-layer convolutional neural network, the first, third and fifth layers in the multi-layer convolutional neural network are convolutional layers, the second, fourth and sixth layers in the multi-layer convolutional neural network are CBAM layers, the seventh layer in the multi-layer convolutional neural network is a fully connected layer, and the eighth layer in the multi-layer convolutional neural network is a Softmax layer.

[0074] The CBAM (Convolutional Block Attention Module) module enhances the ability of convolutional neural networks to extract features from different channels and locations by introducing channel attention modules and spatial attention modules. Figure 3 As shown in , the channel attention module can enhance the feature expression of each channel; Figure 4 As shown in , the Spatial Attention Module can enhance the feature expression of different positions in the image; Figure 5 As shown in Figure 2, the CBAM module multiplies the output features of the channel attention module and the spatial attention module element by element to obtain the final attention-enhanced features.

[0075] This embodiment adopts a random partitioning method to divide 60% of the labeled binary image data set into a training set for training the convolutional neural network based on the CBAM module to learn data features; 20% of the data set is divided into a validation set for adjusting the hyperparameters of the model and preventing the model from overfitting; and 20% of the data set is divided into a test set for evaluating the generalization ability of the trained model.

[0076] (4) Result evaluation

[0077] This example uses mean average precision (mAP) to evaluate the accuracy of the model prediction results. The calculation formula is as follows:

[0078]

[0079] Where C represents the total number of categories, AP i Represents the AP value of the i-th category target.

[0080] The present invention also discloses a method for identifying the main lobe centroid of an optical phased array based on CBAM-CNN. The core of the method is to realize image area recognition of the main lobe area, side lobe area and stray light area of ​​the optical phased array through a convolutional neural network based on the CBAM module, and transform the multi-spot recognition problem into a single-spot recognition problem in the main lobe area of ​​the optical phased array.

[0081] Here, a specific embodiment is cited for detailed description.

[0082] The algorithm flow of the embodiment provided by the present invention is as follows Figure 6 As shown:

[0083] Step 1: Set the termination condition and end the algorithm when the number of loop cycles reaches the predetermined value.

[0084] Step 2: Adjust the optical antenna system so that the far-field light spot of the optical phased array enters the field of view of the image acquisition unit, and further adjust the optical antenna system so that the main lobe of the far-field light spot of the optical phased array appears in the center of the camera's field of view.

[0085] Step 3: The default exposure time parameter of the image acquisition unit CCD / CMOS is E0. If the image processing unit determines that the image is overexposed, the exposure time parameter E is reduced and the acquired image is transmitted to the image processing unit.

[0086] Step 4: The image processing unit determines whether the captured image spot is overexposed. If the image is overexposed, the process jumps to step 3.

[0087] Step 5: The image processing unit sets the image gray value threshold and converts the collected gray value image into a binary image (pixels exceeding the current gray value threshold G are white, and pixels below the current gray value threshold are black).

[0088] Step 6: The image processing unit inputs the binarized image into a convolutional neural network based on the CBAM module, and the neural network outputs the image area coordinates of the main lobe area of ​​the optical phased array, the side lobe area of ​​the optical phased array, and the stray light area.

[0089] Step 7: The centroid recognition unit uses the centroid method to determine the centroid position of the light spot in the main lobe area of ​​the optical phased array, and outputs the centroid coordinates (x, y) of the main lobe area of ​​the far-field light spot of the optical phased array.

[0090] Step 8: Determine whether the termination condition of the algorithm is met. If the termination condition is met, the algorithm ends.

[0091] Step 9: The negative feedback adjustment unit determines whether the far-field spot of the optical phased array exceeds the hardware acquisition field of view of the image acquisition unit according to the centroid coordinates of the main lobe area of ​​the far-field spot of the optical phased array. If it exceeds its field of view, it jumps to the second step; if it does not exceed its field of view, it jumps to the fourth step.

[0092] Preferably, the recognition of the main lobe area and the side lobe area based on the CBAM module convolutional neural network can be achieved by the following steps:

[0093] First, obtain a data set of optical phased array far-field spot binarization images for training convolutional neural networks. Secondly, establish a convolutional neural network based on the CBAM module, and use the optical phased array far-field spot binarization image data set to train the neural network to obtain a trained convolutional neural network based on the CBAM module; the first, third, and fifth layers in the multi-layer convolutional neural network are convolutional layers, the second, fourth, and sixth layers in the multi-layer convolutional neural network are CBAM layers, the seventh layer in the multi-layer convolutional neural network is a fully connected layer, and the eighth layer in the multi-layer convolutional neural network is a Softmax layer. The CBAM layer consists of a channel attention module and a spatial attention module. The channel attention module, the corresponding process description formula is:

[0094]

[0095] ①In the formula, M C (F) is the weight coefficient of the channel attention module of the intermediate feature F, σ is the Sigmoid function, MLP is a multi-layer neural network, AvgPool is average pooling, MaxPool is maximum pooling, W0 and W1 are the weights of the multi-layer neural network MLP, is the average pooling value of the intermediate feature F in the channel attention module, is the maximum pooling value of the intermediate feature F in the channel attention module. The corresponding process description formula of the spatial attention module is:

[0096]

[0097] ②In the formula, M S (F) is the weight coefficient of the spatial attention module of the intermediate feature F, f 5×5 The convolution kernel size is 5×5, is the average pooling value of the intermediate feature F in the spatial attention module, is the maximum pooling value of the intermediate feature F in the spatial attention module. The corresponding process description formula of the CBAM layer is:

[0098]

[0099] In formula ③, F′ is the intermediate quantity of the intermediate feature F after the channel attention module, and F″ is the output quantity of the intermediate feature F after the entire channel attention module and the spatial attention module.

[0100] Finally, the binary image of the far-field spot of the optical phased array to be identified is input into the trained convolutional neural network based on the CBAM module.

[0101] The above embodiments are only specific examples for further describing the purpose, technical solutions and beneficial effects of the present invention, and the present invention is not limited thereto. Any modification, equivalent replacement, improvement, etc. made within the scope of the disclosure of the present invention are included in the protection scope of the present invention.

Claims

1. A method for identifying the main lobe centroid of an optical phased array based on CBAM-CNN, comprising the following steps: S1. Adjust the optical antenna unit so that the far-field light spot of the optical phased array enters the field of view of the image acquisition unit, and adjust the optical antenna unit so that the main lobe of the far-field light spot of the optical phased array appears in the center of the field of view of the camera; S2, adjusting the hardware exposure parameters of the image acquisition unit to display as many light spots as possible in its field of view without overexposure; S3, the image processing unit reads the current gray value image of the image acquisition unit, and converts the acquired gray value image into a binary image by setting a gray value threshold; S4, the image processing unit inputs the obtained binary image into the convolutional neural network based on the CBAM module, and outputs the main lobe area, the side lobe area and the stray light area of ​​the optical phased array based on the convolutional neural network of the CBAM module; S5, the centroid identification unit uses the centroid method to determine the centroid position of the main lobe area spot, and outputs the centroid coordinates of the main lobe area of ​​the optical phased array far-field spot; S6, the negative feedback adjustment unit determines whether the far-field spot of the optical phased array exceeds the hardware acquisition field of view of the image acquisition unit. If it exceeds the field of view, jump to step S1; If it does not exceed its field of view, jump to step S3.

2. The optical phased array main lobe centroid identification method based on CBAM-CNN according to claim 1, characterized in that: In step S4, the recognition of the main lobe area and the side lobe area based on the CBAM module convolutional neural network can be achieved by the following steps: S41, obtaining an optical phased array far-field spot binary image dataset for training a convolutional neural network; S42, establishing a convolutional neural network based on the CBAM module, and training the neural network using an optical phased array far-field spot binary image data set to obtain a trained convolutional neural network based on the CBAM module; S43, inputting the binary image of the far-field spot of the optical phased array to be identified into the trained convolutional neural network based on the CBAM module.

3. The optical phased array main lobe centroid identification method based on CBAM-CNN according to claim 2, characterized in that: The CBAM-based convolutional neural network in step S42 is a multi-layer convolutional neural network, the first, third and fifth layers in the multi-layer convolutional neural network are convolutional layers, the second, fourth and sixth layers in the multi-layer convolutional neural network are CBAM layers, the seventh layer in the multi-layer convolutional neural network is a fully connected layer, and the eighth layer in the multi-layer convolutional neural network is a Softmax layer.

4. The optical phased array main lobe centroid identification method based on CBAM-CNN according to claim 3 is characterized in that: The CBAM layer consists of a channel attention module and a spatial attention module.

5. The optical phased array main lobe centroid identification method based on CBAM-CNN according to claim 4 is characterized in that: The channel attention module has the following process description formula: ①In the formula, M C (F) is the weight coefficient of the channel attention module of the intermediate feature F, σ is the Sigmoid function, MLP is a multi-layer neural network, AvgPool is average pooling, MaxPool is maximum pooling, W0 and W1 are the weights of the multi-layer neural network MLP, is the average pooling value of the intermediate feature F in the channel attention module, is the maximum pooling value of the intermediate feature F in the channel attention module.

6. The optical phased array main lobe centroid identification method based on CBAM-CNN according to claim 4, characterized in that: The spatial attention module has the following process description formula: ②In the formula, M S (F) is the weight coefficient of the spatial attention module of the intermediate feature F, f 5×5 The convolution kernel size is 5×5, is the average pooling value of the intermediate feature F in the spatial attention module, is the maximum pooling value of the intermediate feature F in the spatial attention module.

7. The optical phased array main lobe centroid identification method based on CBAM-CNN according to claim 4, characterized in that: The CBAM layer has the following process description formula: In formula ③, F′ is the intermediate quantity of the intermediate feature F after the channel attention module, and F″ is the output quantity of the intermediate feature F after the entire channel attention module and the spatial attention module.

8. The optical phased array main lobe centroid identification method based on CBAM-CNN according to claim 1, characterized in that: Steps S1 to S6 are executed cyclically, and the centroid coordinates of the main lobe area of ​​the optical phased array far-field spot are output once each round. The cycle ends when the number of cycles reaches a predetermined value.

9. The optical phased array main lobe centroid identification system based on CBAM-CNN is characterized by: It includes an optical antenna unit, an image acquisition unit, an image processing unit, a centroid recognition unit, and a negative feedback regulation unit; The optical antenna unit is used to receive the target light beam and transmit the light beam to the image acquisition unit. The relative position of the target light beam irradiated on the image acquisition unit can be changed by adjusting the optical antenna unit; The image acquisition unit is used to acquire the spot image generated by the target light beam irradiating thereon, and the information such as the spot brightness of the acquired image can be changed by changing the parameters such as the hardware exposure time of the image acquisition unit; The image processing unit is used to convert the collected light spot gray value image into a binary image, and realize image area recognition of the main lobe area of ​​the optical phased array, the side lobe area of ​​the optical phased array, and the stray light area through a convolutional neural network based on the CBAM module; The centroid identification unit uses the centroid method to determine the centroid position of the light spot in the main lobe area of ​​the optical phased array, and outputs the centroid coordinates of the main lobe area of ​​the far-field light spot of the optical phased array; The negative feedback adjustment unit is used to compare the deviation between the centroid coordinates of the main lobe area of ​​the far-field light spot of the optical phased array and the center coordinates of the field of view of the image acquisition unit, calculate the relevant control parameters of the optical antenna unit, and control the optical antenna unit to move the centroid of the light spot to the center of the field of view of the image acquisition unit.

10. The optical phased array main lobe centroid identification system based on CBAM-CNN according to claim 9, characterized in that: When the image processing unit performs region recognition, the convolutional neural network target detection method based on the CBAM module is a supervised machine learning method, and its training method is as follows: (1) Establishing a Dataset Collect binary images including the main lobe, side lobe and stray light spot of the optical phased array and establish a data set; (2) Image Annotation The data set is manually labeled to mark the main lobe area, side lobe area, and stray light area of ​​the optical phased array. The data set with the spot type and spot position marked is used for subsequent model training and evaluation; (3) Model training The convolutional neural network based on the CBAM module is a multi-layer convolutional neural network, the first, third and fifth layers in the multi-layer convolutional neural network are convolutional layers, the second, fourth and sixth layers in the multi-layer convolutional neural network are CBAM layers, the seventh layer in the multi-layer convolutional neural network is a fully connected layer, and the eighth layer in the multi-layer convolutional neural network is a Softmax layer. The CBAM module enhances the ability of convolutional neural networks to extract features from different channels and positions by introducing channel attention modules and spatial attention modules. (4) Result evaluation The mean average precision (mAP) is used to evaluate the accuracy of the model's prediction results. The calculation formula is as follows: Where C represents the total number of categories, AP i Represents the AP value of the i-th category target.

Images