CN117853932A - Sea surface target detection method, detection platform and system based on photoelectric pod - Google Patents

Abstract

The invention discloses a sea surface target detection method, a detection platform and a system based on a photoelectric pod, which belong to the field of sea surface target detection and comprise the following steps: for the aggregated multiband sea surface image data transmitted by the optoelectronic pod, performing: depolymerizing it into a plurality of single-pass image data; respectively executing sea clutter self-adaptive suppression steps on each single-path image data to obtain a denoised image; respectively carrying out target detection on the denoised images under the same focal length, and fusing detection results; the sea clutter self-adaptive suppression step comprises the following steps: performing top hat transformation on the image data to be processed, calculating gradients of the image data in different directions, and taking the minimum pixel value of a plurality of gradient images at the same position as the pixel value of the denoised image at the position; and adjusting the pose of the tele lens by using the detection result under the wide-angle lens. The invention can fully utilize the multichannel sea surface image data acquired by the photoelectric pod, inhibit sea surface clutter therein and improve the detection precision of sea surface targets.

Description

A sea surface target detection method, detection platform and system based on optoelectronic pod

Technical Field

The present invention belongs to the field of sea surface target detection, and more specifically, relates to a sea surface target detection method, a detection platform and a system based on an optoelectronic pod.

Background technique

Maritime target detection technology has great research significance in both civil and military fields. On the one hand, with the continuous expansion and deepening of climate factors such as typhoons and marine oil and gas development activities, the number of water traffic accidents is on the rise. The harsh marine environment is extremely likely to cause major economic losses and casualties. Therefore, it is very important to quickly identify and track targets in distress at sea and ensure a safe navigation environment. On the other hand, maritime protection is of great significance. Rapid positioning, monitoring and combating of invading ships can better control the initiative at sea. Therefore, high-performance and all-round monitoring of the complex sea environment is required. Intelligent detection of maritime targets is the key to protecting coastal defense security.

Marine target detection usually has the problem of complex background. The complex clutter of the sea surface background includes not only high-brightness sea surface bright strips and dense sea surface fish-scale bright spots, but also background clutter noise such as the heterogeneous and complex background of the sea and sky and island clutter. The target is confused by the complex sea surface clutter background noise, thereby increasing the difficulty of sea surface target detection.

With the miniaturization and diversification of photoelectric detectors such as short-wave infrared, visible light, medium-wave infrared, and long-wave infrared, it is easier to obtain image data in different bands. There are also some differences in the information in images of different bands. For example, the penetration of infrared light allows infrared images to distinguish targets from backgrounds based on radiation differences. Therefore, infrared images can obtain more effective information at night, while visible light can obtain texture and color information that is more in line with the human visual system. Similarly, the choice of lenses is more diverse: wide-angle lenses have the characteristics of short focal length and large viewing angle. It can capture more scenes, while telephoto lenses have long focal lengths and small viewing angles, and can clearly image distant targets. For detectors of different bands and different lens combinations, it is particularly important to process their image data so that the characteristics of multi-channel images can be more efficiently utilized to obtain more accurate detection and recognition results.

The optoelectronic pod can be flexibly configured according to the mission requirements, and can respectively adopt different combinations of short-wave infrared, visible light, and long-wave infrared detectors and telephoto and wide-angle lenses to collect sea surface image data of different bands and different focal lengths. In the patent document with application publication number CN116248705A, a multi-channel image transmission and processing system for a micro optoelectronic pod is disclosed, as shown in FIG1, which includes a multi-channel photoelectric detector, an FPGA unit A, and an image aggregation unit. The multi-channel photoelectric detector is used to collect photon signals of different bands, generate image electrical signals through a photoelectric conversion circuit, and output the multi-channel image signals to the FPGA unit A; the FPGA unit A is used to realize the collection of multi-channel image signals, generate multi-channel image data and output it to the image aggregation unit; the image aggregation unit aggregates the multi-channel image data into aggregated image data and then outputs it to a single-channel high-speed link connecting the micro optoelectronic pod and a remote data processing platform (such as an aircraft platform), thereby realizing the synchronous real-time transmission, processing, forwarding or storage of multi-channel detector image data, which has the characteristics of good real-time performance, high transmission rate, low link cost, and good flexibility. Using the micro optoelectronic pod to collect sea surface image data is beneficial to improving the sea surface target detection effect, but there is still a lack of effective methods for effectively processing multi-channel image data to accurately complete sea surface target detection.

Summary of the invention

In view of the defects of the prior art and the need for improvement, the present invention provides a sea surface target detection method, detection platform and system based on an optoelectronic pod, the purpose of which is to make full use of the multi-channel sea surface image data collected by the optoelectronic pod and suppress the sea surface clutter therein to improve the detection accuracy of sea surface targets.

To achieve the above object, according to one aspect of the present invention, a method for detecting sea surface targets based on an optoelectronic pod is provided, comprising:

Continuously receive the sea surface image data sent by the optoelectronic pod; for each frame of aggregated multi-band sea surface image data received, perform the following steps:

(S1) deaggregating it into a plurality of single-channel image data; each single-channel image data corresponds to image data of a band at a focal length collected by the optoelectronic pod;

(S2) performing a sea surface clutter adaptive suppression step on each single channel image data to obtain a corresponding denoised image; the sea surface clutter adaptive suppression step includes:

(S21) performing a top-hat transformation on the image data to be processed to obtain a preliminary denoised image;

(S22) calculating the gradient of the preliminary denoised image in different directions to obtain a plurality of gradient images;

(S23) taking the minimum pixel value of the multiple gradient images at the same position as the pixel value of the denoised image at the position, and obtaining the denoised image corresponding to the image to be processed;

(S3) After performing target detection on the denoised images at the same focal length, the detection results are fused and the fused results are used as the sea surface target detection results at the corresponding focal length.

Further, step (S22) includes:

Mark the connected domains of the preliminary denoised image, remove the connected domains smaller than the preset threshold, and calculate the average size C of the remaining connected domains;

Calculate the gradient convolution kernel size ks according to ks = C // M , and determine the downward gradient convolution kernels in all directions according to the calculated gradient convolution kernel size;

The preliminary denoised image is convolved using the downward gradient convolution kernel in all directions to calculate the gradient of the preliminary denoised image in different directions, thereby obtaining multiple gradient images;

Wherein, “//” represents integer division operation; M is a preset positive integer.

Furthermore, M =5.

Furthermore, the method for detecting sea surface targets based on an optoelectronic pod provided by the present invention further comprises, after step (S3):

(S4) calculating the center position of the sea surface target according to the sea surface target detection result at the shorter focal length, and calculating the offset of the center position of the sea surface target relative to the center of the image;

(S5) Sending a command to the optoelectronic pod so that the optoelectronic pod adjusts the posture of the lens with the longer focal length according to the offset, so that the center position of the sea surface target at the focal length coincides with the center position of the image.

Furthermore, in step (S3), the target detection results at the same focal length are fused, including:

Calculate the overlap between the detection frames obtained by target detection in each denoised image at the focal length, output the detection frames with an overlap greater than a preset high threshold as credible targets, and remove the detection frames with an overlap less than a preset low threshold as false alarms;

The trajectory of the sea surface target is determined according to the sea surface target detection result sequence of the sea surface image data before the current frame. For the detection frame with an overlap between the low threshold and the high threshold, if it is located on the trajectory of the sea surface target, it is output as a credible target, otherwise it is rejected as a false alarm.

Furthermore, in step (S3), if the denoised image is a visible light band image, YOLOv5 is used for target detection; if the denoised image is an infrared band image, a multi-scale block contrast measurement algorithm is used for detection.

According to another aspect of the present invention, a sea surface target detection platform based on an optoelectronic pod is provided, comprising:

A computer-readable storage medium for storing a computer program;

and a processor, which is used to read a computer program stored in a computer-readable storage medium and execute the sea surface target detection method based on an optoelectronic pod provided by the present invention.

According to another aspect of the present invention, there is provided a sea surface target detection system, comprising:

The above-mentioned sea surface target detection platform based on optoelectronic pod provided by the present invention;

and optoelectronic pods;

The image data stream is transmitted between the optoelectronic pod and the sea surface target detection platform via a single-channel link.

In general, the above technical solutions conceived by the present invention can achieve the following beneficial effects.

(1) The present invention first performs a top-hat transformation on the sea surface image data collected by the optoelectronic pod, which can effectively filter out sea clutter noise with uniform brightness and simple texture, such as fish scale waves and bright stripes on the sea surface. Then, the specific value of each pixel in the image is determined based on the directional gradient calculation, thereby effectively suppressing the noise in the image that is close to the target size but inconsistent with the shape. The combination of the two can effectively suppress the sea clutter noise in the sea surface image data. On this basis, target detection is performed on the denoised multi-channel image data respectively, and the target detection results of images of different bands at the same focal length are used as the sea surface target detection results at the corresponding focal length. In this way, the information of images at different bands and focal lengths can be fully utilized, and the accuracy of sea surface target detection can be effectively improved.

(2) In the preferred embodiment of the present invention, when suppressing noise in an image that is close in size to the target but inconsistent in shape based on directional gradient calculation, the image is first marked for connected domains, and connected domains that are too small are filtered out. Then, the size of the gradient convolution kernel used to calculate the gradient is determined based on the average size of the remaining connected domains. This can adapt to the size of the sea surface target to be detected in different scenarios, effectively improving the accuracy and robustness of sea surface target detection. In a further preferred embodiment of the present invention, the gradient convolution kernel is further set to the result of the average size of the connected domain divided by 5. Experimental data show that the gradient calculated based on the gradient convolution kernel size can suppress the sea surface background clutter to the greatest extent.

(3) At a shorter focal length, the field of view is larger and there are more noise and interference, the target size is small, and the detection difficulty is large. At a longer focal length, the field of view is smaller, the noise and interference are smaller, the target is clear, and the detection difficulty is small. However, the sea surface target may be located outside the field of view. In the preferred embodiment of the present invention, the offset of the sea surface target center relative to the image center calculated from the sea surface target detection result at a shorter focal length is used to adjust the lens posture of the longer focal length, thereby obtaining higher quality target image data using a telephoto lens, thereby obtaining more accurate sea surface target detection results at a longer focal length.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a schematic diagram of the data connection relationship between the existing optoelectronic pod and the aircraft platform.

FIG2 is a schematic diagram of a method for detecting sea surface targets based on an optoelectronic pod according to an embodiment of the present invention.

FIG. 3 is an original image before top-hat transformation provided by an embodiment of the present invention.

FIG. 4 is an image of the original image shown in FIG. 3 after being top-hat transformed.

FIG5 is a schematic diagram of gradient convolution kernels in eight directions provided by an embodiment of the present invention.

FIG. 6 is a gradient diagram of the image shown in FIG. 4 in the 0° direction.

FIG. 7 is a gradient diagram of the image shown in FIG. 4 in the 90° direction.

Detailed ways

In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

In the present invention, the terms "first", "second", etc. (if any) in the present invention and the drawings are used to distinguish similar objects but not necessarily to describe a specific order or sequence.

In order to effectively improve the detection accuracy of sea surface targets, the present invention provides a sea surface target detection method, detection platform and system based on an optoelectronic pod. The overall idea is to use the optoelectronic pod to collect sea surface images in different bands and focal lengths, and based on the characteristics of sea surface clutter in the sea surface images, an effective sea surface clutter suppression method is proposed, thereby making full use of the information of multi-channel image data while effectively suppressing the sea surface clutter, thereby improving the accuracy of sea surface target detection.

In practical applications, optoelectronic pods are mostly micro or small structures, in which the multi-channel photoelectric detectors are used to collect photon signals of different bands. For input photon signals of different bands, image electrical signals are generated through photoelectric conversion circuits, and then multiple single-channel image data are aggregated into one high-speed image data through an image aggregation unit, and high-speed transmission of image data streams is achieved through a single-channel high-speed link. The photoelectric detectors in the optoelectronic pod can be flexibly configured according to task requirements, and can respectively use combinations of detectors of different bands such as short-wave infrared, visible light, and long-wave infrared and lenses of different focal lengths such as wide-angle lenses and telephoto lenses. Without loss of generality, in the following embodiments, the optoelectronic pod used includes four photoelectric detectors, namely, a combination of a visible light detector and a wide-angle lens, a combination of an infrared detector and a wide-angle lens, a combination of a visible light detector and a telephoto lens, and a combination of an infrared detector and a telephoto lens. Therefore, the optoelectronic pod can collect 4 different sea surface image data at a time. For ease of description, in the following embodiments, these four detectors are respectively recorded as detector 1, detector 2, detector 3, and detector 4. It should be noted that the multi-channel detectors herein are merely exemplary. In other embodiments of the present invention, an optoelectronic pod with different multi-channel detectors may also be used to collect sea surface image data.

The following are examples.

Embodiment 1:

A method for detecting sea surface targets based on an optoelectronic pod, as shown in FIG2 , includes:

Continuously receive the sea surface image data sent by the optoelectronic pod; for each frame of aggregated multi-band sea surface image data received, perform the following steps (S1) to (S3).

The step (S1) of this embodiment includes: deaggregating it into multiple single-channel image data; each single-channel image data corresponds to image data of a band at a focal length collected by the optoelectronic pod; in this embodiment, after deaggregating a frame of sea surface image data received, four channels of image data can be obtained, namely, a wide-angle image of a visible light band, a wide-angle image of an infrared band, a telephoto image of a visible light band, and a telephoto image of an infrared band. Optionally, in this embodiment, the size of the visible light image is 1920*1080, and the size of the infrared image is 640*512.

The step (S2) of this embodiment includes: performing a sea surface clutter adaptive suppression step on each single channel of image data to obtain a corresponding denoised image.

Sea clutter noise with uniform brightness and simple texture on the sea surface, such as bright stripes and sea fish scale waves, is the main component of sea clutter and will seriously interfere with the detection of sea surface targets. Therefore, this embodiment will first suppress these noises. Analysis and related experimental data show that top hat transformation can effectively suppress such sea clutter noise. After the sea clutter noise with uniform brightness and simple texture in the sea surface image is suppressed, there will still be some noise in the image that is close to the target size but inconsistent in shape. These noises will still interfere with the sea surface target detection result. The present invention further analyzes and finds that since the background clutter has a large gradient only in a certain direction, and the target is isotropic, the gradient changes in different directions around it are small. Based on this finding, the present invention proposes to calculate the gradients in each direction on the basis of suppressing the bright stripes and sea fish scale wave noises, and take the minimum value of the same pixel position in each gradient map as the pixel value of the position on the image, so as to achieve the effect of effectively suppressing the background clutter.

Based on the above analysis, as shown in FIG2 , in order to effectively suppress sea surface clutter, in this embodiment, the sea surface clutter adaptive suppression step includes:

(S21) performing a top-hat transformation on the image data to be processed to obtain a preliminary denoised image;

(S22) calculating the gradient of the preliminary denoised image in different directions to obtain a plurality of gradient images;

(S23) The minimum pixel value of the multiple gradient images at the same position is used as the pixel value of the denoised image at the position, so as to obtain the denoised image corresponding to the image to be processed.

This embodiment further takes into account that the target sizes vary in different scenarios, and therefore designs an adaptive gradient convolution kernel size to cope with different scenarios, further improving the sea clutter suppression effect. Specifically, step (S22) includes:

Mark the connected domains of the preliminary denoised image, remove the connected domains smaller than the preset threshold, and calculate the average size C of the remaining connected domains;

Calculate the gradient convolution kernel size ks according to ks = C // M , and determine the downward gradient convolution kernels in all directions according to the calculated gradient convolution kernel size;

The preliminary denoised image is convolved using the downward gradient convolution kernel in all directions to calculate the gradient of the preliminary denoised image in different directions, thereby obtaining multiple gradient images;

Wherein, “//” represents integer division operation; M is a preset positive integer. As a preferred implementation, in this embodiment, M = 5. Experimental data show that by setting the relationship between the gradient convolution kernel size and the average size of the connected domain in this way, the effect of suppressing sea surface clutter can be maximized.

In this embodiment, the adaptive directional gradient calculation is aimed at suppressing the noise in the image that is close to the target size but different in shape. After the connected domain is marked, each connected domain corresponds to a noise or a sea surface target. Therefore, the noise that is too small is first removed according to the preset threshold, which can ensure that the average size of the connected domain is close to the size of the sea surface target; optionally, in this embodiment, considering that most of the actual sea surface targets are ships at sea, low-altitude drones, etc., based on the empirical size of these targets, this embodiment sets the preset threshold to 10. In some other embodiments, it can also be set to other values according to the size of the actual detected target, ensuring that the connected domain that can be removed is significantly smaller than the target size.

The following is a further explanation of the above-mentioned sea clutter adaptive suppression step by taking the processing process of an actual image as an example. As shown in FIG3, it is an original sea surface image data. In the figure, the object in the box is the sea surface target to be detected, and it is processed according to the sea surface clutter adaptive suppression step proposed in this embodiment. After the sea surface image shown in FIG3 is top-hat transformed, the resulting image is shown in FIG4. By comparing FIG3 and FIG4, it can be clearly seen that the bright bars and sea surface fish scale waves in the original image are effectively suppressed. After binarization of the image shown in FIG4, the connected domain is marked, and the connected domains less than 10 are eliminated. According to ks = C //5, the gradient convolution kernel size is calculated to be 7, and the gradient convolution kernels in the eight directions of 0°, 45°, 90°, 135°, 180°, 225°, 270° and 315° are calculated as shown in FIG5. The gradient map in the corresponding direction can be calculated using the gradient convolution kernels in different directions. Among them, after the gradient calculation is performed on the image shown in Figure 4, the gradient maps at 0° and 90° directions are shown in Figures 6 and 7 respectively. For the gradient maps at 8 different directions, the minimum pixel value of each gradient map at the same pixel position is used as the pixel value of the corresponding position in the image, and the denoised image with effectively suppressed sea clutter can be obtained.

In this embodiment, step (S3) includes: for the denoised images at the same focal length, after performing target detection respectively, fusing the detection results, and using the fusion results as the sea surface target detection results at the corresponding focal length.

When performing target detection on each image, a specific target detection algorithm can be selected according to the characteristics of the image in the corresponding band. In this embodiment, the size of the visible light image is 1920*1080, and the size of the infrared image is 640*512. Since there is more information in the visible light image, YOLOv5 is selected as the visible light target detection algorithm, and the traditional multiscale patch contrast measurement algorithm (Multiscale Patch Contrast Measurement, MPCM) is selected to realize the detection of infrared image targets. In some other embodiments of the present invention, the target detection algorithm for a specific image can also be flexibly selected as other algorithms.

By fusing the target detection results of images of different bands at the same focal length, the different information carried by images of different bands can be fully utilized to achieve multi-band image collaborative processing, and further improve the detection accuracy of sea surface targets. Since images of different bands at the same focal length have the same field of view, the detection frames of the detection results of the two images are matched. In order to effectively achieve the fusion of the detection results of images of different bands at the same focal length, this embodiment adopts a high and low dual threshold fusion method. Specifically, for an image at a certain focal length, the target detection results are fused, including:

Calculate the overlap (IOU) between the detection frames obtained by target detection in each denoised image at the focal length, output the detection frames with an overlap greater than a preset high threshold (e.g. 0.7) as credible targets, and remove the detection frames with an overlap less than a preset low threshold (e.g. 0.3) as false alarms;

The trajectory of the sea surface target is determined according to the sea surface target detection result sequence of the sea surface image data before the current frame. For the detection frame with an overlap between the low threshold and the high threshold, if it is located on the trajectory of the sea surface target, it is output as a credible target, otherwise it is rejected as a false alarm.

Among them, determining the trajectory of the sea surface target according to the sea surface target detection result sequence of the sea surface image data before the current frame can be completed based on the existing target tracking means.

Since the fields of view of lenses with different focal lengths are different, at a shorter focal length, the field of view is larger and there are more noise and interference, the target size is small, and the detection difficulty is large, while at a longer focal length, the field of view is smaller, the noise and interference are smaller, the target is clear, and the detection difficulty is small, but the sea surface target may be outside the field of view. Based on this, after simultaneously obtaining the sea surface target detection results at the shorter focal length and the sea surface target detection results at the longer focal length, this embodiment can directly take the sea surface target detection result at the longer focal length as the final sea surface target detection result; otherwise, if only the sea surface target detection result at one focal length is obtained, it is used as the final sea surface target detection result.

In order to further improve the detection accuracy of the sea surface target detection results, the present invention further proposes to regard the sea surface target detection results at a shorter focal length as low confidence results and use this to calculate the offset information of the target center relative to the image center, and use this information as guidance information for adjusting the longer lens posture at the optoelectronic pod end, so as to obtain higher quality target image data, and then obtain more accurate telephoto image target detection results. Accordingly, this embodiment further includes after step (S3):

(S4) calculating the center position of the sea surface target according to the sea surface target detection result at the shorter focal length, and calculating the offset of the center position of the sea surface target relative to the center of the image;

(S5) Sending a command to the optoelectronic pod so that the optoelectronic pod adjusts the posture of the lens with the longer focal length according to the offset, so that the center position of the sea surface target at the focal length coincides with the center position of the image.

In general, this embodiment uses the optoelectronic pod as the sea surface data collection end, which can collect multi-channel image data of different focal lengths and different bands, providing a more sufficient data basis for the accurate detection of sea surface targets; based on the characteristics of sea clutter, an adaptive suppression step for sea surface clutter is proposed, which can effectively suppress the sea clutter noise in the sea surface image data, thereby effectively improving the detection accuracy of sea surface targets. On this basis, the target detection results under the shorter lens are used to guide the posture adjustment of the lens under the longer focal length in the optoelectronic pod, so that higher quality target image data can be obtained, and then more accurate telephoto image target detection results can be obtained.

Embodiment 2:

A sea surface target detection platform based on an optoelectronic pod, comprising:

A computer-readable storage medium for storing a computer program;

and a processor, for reading a computer program stored in a computer-readable storage medium, and executing the sea surface target detection method based on an optoelectronic pod provided in the above-mentioned embodiment 1.

The sea surface target detection platform based on the optoelectronic pod provided in this embodiment can be any platform such as an aircraft platform, a remote control platform, etc. that can transmit data with the optoelectronic pod and has data processing capabilities.

Embodiment 3:

A sea surface target detection system, comprising:

The sea surface target detection platform based on the optoelectronic pod provided in the above-mentioned embodiment 2;

and optoelectronic pods;

The image data stream is transmitted between the optoelectronic pod and the sea surface target detection platform via a single-channel link.

It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims (8)

1. A method for detecting sea surface targets based on an optoelectronic pod, comprising: Continuously receiving the sea surface image data sent by the optoelectronic pod; for each frame of the received aggregated multi-band sea surface image data, performing the following steps: (S1) deaggregating it into a plurality of single-channel image data; each single-channel image data corresponds to image data of a band at a focal length collected by the optoelectronic pod; (S2) performing a sea surface clutter adaptive suppression step on each single channel of image data to obtain a corresponding denoised image; the sea surface clutter adaptive suppression step comprises: (S21) performing a top-hat transformation on the image data to be processed to obtain a preliminary denoised image; (S22) calculating the gradients of the preliminary denoised image in different directions to obtain a plurality of gradient images; (S23) taking the minimum pixel value of the plurality of gradient images at the same position as the pixel value of the denoised image at the position, and obtaining the denoised image corresponding to the image to be processed; (S3) After performing target detection on the denoised images at the same focal length, the detection results are fused and the fused results are used as the sea surface target detection results at the corresponding focal length. 2. The method for detecting sea surface targets based on an optoelectronic pod according to claim 1, wherein the step (S22) comprises: After binarizing the preliminary denoised image, connected domains are marked, connected domains smaller than a preset threshold are removed, and the average size C of the remaining connected domains is calculated; Calculate the gradient convolution kernel size ks according to ks = C // M , and determine the downward gradient convolution kernels in all directions according to the calculated gradient convolution kernel size; Using downward gradient convolution kernels in all directions to perform a convolution operation on the preliminary denoised image to calculate the gradients of the binary image in different directions to obtain a plurality of gradient images; Wherein, “//” represents integer division operation; M is a preset positive integer. 3. The sea surface target detection method based on the optoelectronic pod as described in claim 2 is characterized in that M = 5. 4. The method for detecting sea surface targets based on an optoelectronic pod according to any one of claims 1 to 3, characterized in that after step (S3), it also includes: (S4) calculating the center position of the sea surface target according to the sea surface target detection result at the shorter focal length, and calculating the offset of the center position of the sea surface target relative to the center of the image; (S5) Sending a command to the optoelectronic pod so that the optoelectronic pod adjusts the posture of the lens with the longer focal length according to the offset so that the center position of the sea surface target at the focal length coincides with the center position of the image. 5. The method for detecting sea surface targets based on an optoelectronic pod according to any one of claims 1 to 3, characterized in that in the step (S3), the target detection results at the same focal length are fused, comprising: Calculate the overlap between the detection frames obtained by target detection in each denoised image at the focal length, output the detection frames with an overlap greater than a preset high threshold as credible targets, and remove the detection frames with an overlap less than a preset low threshold as false alarms; The trajectory of the sea surface target is determined according to the sea surface target detection result sequence of the sea surface image data before the current frame. For the detection frame whose overlap degree is between the low threshold and the high threshold, if it is located on the trajectory of the sea surface target, it is output as a credible target, otherwise, it is eliminated as a false alarm. 6. The method for sea surface target detection based on an optoelectronic pod as described in claim 5 is characterized in that, in the step (S3), if the denoised image is a visible light band image, YOLOv5 is used for target detection; if the denoised image is an infrared band image, a multi-scale block contrast measurement algorithm is used for detection. 7. A sea surface target detection platform based on an optoelectronic pod, characterized by comprising: A computer-readable storage medium for storing a computer program; and a processor, configured to read the computer program stored in the computer-readable storage medium and execute the sea surface target detection method based on the optoelectronic pod according to any one of claims 1 to 6. 8. A sea surface target detection system, comprising: The sea surface target detection platform based on the optoelectronic pod as described in claim 7; and optoelectronic pods; The image data stream is transmitted between the optoelectronic pod and the sea surface target detection platform via a single-channel link.