A method and system for detecting a remote small target

Abstract

The application relates to the technical field of target detection, and provides a long-distance micro target detection method and system, which comprises the following steps: acquiring the diagonal line size, width-height ratio and resolution of a sensor; calculating the physical width of the sensor based on the diagonal line size and the width-height ratio of the sensor; calculating the physical size of each pixel based on the physical width and the resolution of the sensor; converting the pixel width occupied by a target in an image into the physical width of the target; acquiring the actual width of the target and the distance range between the target and the sensor; combining the physical width of the target to determine the focal length range; acquiring the image under the focal length range; obtaining multiple image blocks through an image segmentation method; after detecting the target in each image block by adopting an image detection model, integrating the targets in all the image blocks to obtain a detection result. The long-distance small target is efficiently detected and recognized.

Description

Technical Field

[0001] This invention belongs to the field of target detection technology, and in particular relates to a method and system for detecting small targets at long distances. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] With the rapid development of computer vision technology, object detection has been widely applied. However, existing technologies still face many challenges in detecting small objects at long distances. Traditional object detection methods have significant limitations in both accuracy and efficiency when dealing with both long-distance and small targets. YOLOv8, as a fast and efficient object detection framework, is widely used in real-time detection tasks.

[0004] However, YOLOv8 has shown shortcomings in long-range and small target detection. First, while the original backbone network Darknet53 performs well in many scenarios, its complexity and computational cost make it difficult to maintain high efficiency in real-time detection tasks. Second, the YOLOv8 algorithm has limited ability to detect small targets within a 200-meter range, failing to meet practical application requirements. Summary of the Invention

[0005] To address the technical problems mentioned above, this invention provides a method and system for detecting small targets at long distances. It proposes to determine the focal length range based on information such as the sensor's diagonal size, aspect ratio, actual width of the target, and the distance range between the target and the sensor. This enables the acquisition of high-resolution images of small targets at long distances, thereby achieving efficient detection and recognition of small targets at long distances.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] The first aspect of the present invention provides a method for detecting small targets at long distances, comprising:

[0008] Obtain the sensor's diagonal size, aspect ratio, and resolution. Based on the sensor's diagonal size and aspect ratio, calculate the sensor's physical width. Based on the sensor's physical width and resolution, calculate the physical size of each pixel. Convert the pixel width occupied by the target in the image into the target's physical width.

[0009] Obtain the actual width of the target and the distance range between the target and the sensor, and determine the focal length range by combining the physical width of the target;

[0010] The image within the specified focal length range is acquired, and multiple image blocks are obtained through image segmentation. After the target in each image block is detected using an image detection model, the targets in all image blocks are integrated to obtain the detection result.

[0011] Furthermore, the physical width of the sensor is:

[0012] W s =k×Q1

[0013] Wherein, scaling factor The resolution is P1×P2, the aspect ratio is Q1:Q2, and the diagonal size is H.

[0014] Furthermore, the physical width of the target:

[0015]

[0016] The physical width of the sensor is W. s The resolution is P1×P2, and W1 is the pixel width occupied by the target in the image.

[0017] Furthermore, the relationship between the focal length range and the distance range between the target and the sensor is as follows:

[0018]

[0019] In the formula, f is the focal length, D is the distance between the lens and the target, W3 is the actual width of the target, and W2 is the physical width of the target.

[0020] Furthermore, the image segmentation method is as follows: the image is divided into n1 parts along the long axis and into n2 parts along the short axis, and finally the image is divided into n1+n2 image blocks.

[0021] Furthermore, the image detection model adopts YOLOv8, and the backbone network of YOLOv8 adopts MobileNetV3.

[0022] Furthermore, the dataset used for training the image detection model is obtained by annotating images acquired at various angles and heights after brightness enhancement, rotation, and flipping.

[0023] A second aspect of the present invention provides a long-range small target detection system, comprising:

[0024] The target physical width calculation module is configured to: obtain the sensor diagonal size, aspect ratio and resolution; calculate the sensor physical width based on the sensor diagonal size and aspect ratio; calculate the physical size of each pixel based on the sensor physical width and resolution; and convert the pixel width occupied by the target in the image into the target physical width.

[0025] The focal length range determination module is configured to: acquire the actual width of the target and the distance range between the target and the sensor, and determine the focal length range in combination with the physical width of the target;

[0026] The target detection module is configured to: acquire images within the focal length range, obtain multiple image blocks through image segmentation methods, detect targets in each image block using an image detection model, and then integrate the targets in all image blocks to obtain the detection result.

[0027] A third aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method for detecting small targets at long distances as described above.

[0028] A fourth aspect of the present invention provides a computer device including a computer-readable storage medium, a processor, and a computer program stored on the computer-readable storage medium and executable on the processor, wherein the processor executes the program to implement the steps of a long-range small target detection method as described above.

[0029] Compared with the prior art, the beneficial effects of the present invention are:

[0030] This invention determines the focal length range based on information such as the sensor's diagonal size, aspect ratio, actual width of the target, and the distance range between the target and the sensor, in order to achieve continuous coverage from close range to a distance of 200 meters, acquire high-resolution images to be identified, and thus achieve efficient detection and identification of small targets at a distance.

[0031] This invention utilizes image segmentation methods to segment high-resolution images in order to accurately extract target regions.

[0032] To meet the needs of real-time detection, this invention improves the YOLOv8 backbone network by replacing its original Darknet53 backbone network with the lightweight MobileNetV3, thereby improving detection speed and efficiency.

[0033] This invention achieves efficient detection and recognition of small targets at a distance through the integrated application of network model optimization, multi-focal length shooting, and image segmentation techniques. Attached Figure Description

[0034] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0035] Figure 1 This is a structural diagram of the improved YOLOv8 network model according to Embodiment 1 of the present invention;

[0036] Figure 2 This is a schematic diagram of the structure of a computer device according to Embodiment 4 of the present invention. Detailed Implementation

[0037] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings.

[0038] It should be noted that the following detailed description is illustrative and intended to provide further explanation of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0039] Example 1

[0040] This embodiment provides a method for detecting small targets at long distances.

[0041] This embodiment provides a method for detecting small targets at long distances, in order to solve the problem of low accuracy in identifying small targets at long distances in the prior art.

[0042] This embodiment provides a method for detecting small targets at long distances, including the following steps:

[0043] S1. Acquire target images to form an initial dataset, preprocess the initial dataset to generate a processed dataset, and use the LabelImg tool to label the dataset.

[0044] Specifically, an initial dataset of small targets was generated by collecting images from various angles and heights. Images with unclear or overly complex target features were removed. The image resolution was 3840×2160. The initial dataset was preprocessed by brightness enhancement, rotation, and flipping to increase the diversity of the dataset and improve the generalization ability and robustness of the model. LabelImg was used to annotate the preprocessed dataset and saved as a txt file.

[0045] S2. Improve the YOLOv8 network model by replacing its original Darknet53 backbone network with the lightweight MobileNetV3 to improve detection speed and efficiency. Furthermore, train the improved YOLOv8 model to obtain the weight file.

[0046] Specifically, the original Darknet53 backbone network of YOLOv8 is replaced with the lightweight MobileNetV3, such as... Figure 1 As shown, the lightweight model employs depthwise separable convolutions to significantly reduce computation and the number of parameters. Depthwise convolutions are applied independently to each input channel, and then pointwise convolutions introduce mappings between different channels. It incorporates a lightweight attention mechanism (SE) module, which adaptively adjusts the weights of each channel to enhance important features and suppress irrelevant information. Furthermore, it integrates network architecture search (NAS). MobileNetV3 combines automated architecture search techniques to optimize the network structure, achieving a better balance between performance and efficiency, thus improving detection speed and efficiency overall. Finally, training is performed based on the improved lightweight network model to obtain the weight file best.pt.

[0047] S3. Set up 4 lenses with different focal lengths to achieve continuous coverage from close range to a distance of 200 meters and acquire high-resolution images to be identified.

[0048] Step S3 specifically includes the following:

[0049] Step 301: Obtain the sensor's diagonal dimensions, aspect ratio, and resolution.

[0050] For example, it is equipped with an IMX415 sensor for shooting, which has a diagonal size of 6.43mm, a resolution of 3840×2160, and an aspect ratio of 16:9.

[0051] Step 302: Calculate the physical width and height of the sensor based on its diagonal dimensions and aspect ratio.

[0052] First, calculate the scaling factor k:

[0053]

[0054] Therefore, the physical width Ws and height Hs of the sensor are respectively:

[0055] W s =16×k=5.6mm

[0056] H s =9×k=3.15mm

[0057] Step 303: Calculate the physical size of each pixel based on the sensor's physical width and resolution:

[0058]

[0059] Step 304: Convert the pixel width of the target in the image to its physical width:

[0060] W1×1.458×10 -3 mm / pixel = W2

[0061] In the formula, W1 is the pixel width occupied by the target in the image, in pixels; W2 is the physical width of the target, in mm.

[0062] In summary, the physical width of the target W s =k×Q1;

[0063] The resolution is P1×P2, where P1 is the horizontal pixel and P2 is the vertical pixel; the aspect ratio is Q1:Q2; and the diagonal size is H.

[0064] Step 305: Based on the physical width of the target, combined with the actual width of the target and the distance range between the target and the sensor, determine the focal length range.

[0065] The relationship between the focal length f and the range of distances between the target and the sensor is as follows:

[0066]

[0067] In the formula, D is the distance between the lens and the target, in meters; W3 is the actual width of the target, in meters.

[0068] Considering small targets such as birds and bees, the ratio of the target's pixel width to its temporal width in the image is approximately between 280 and 550 pixels / m. The following is the range for determining the lens focal length.

[0069] (1) The focal length range of the first lens is selected to be 21mm to 40mm. Lenses within this focal length range can provide a wide field of view, which is suitable for capturing a wide range of scenes. Therefore, this lens can shoot small targets within a close range (0-50 meters). Due to the wide field of view, the lens can ensure that the target occupies a large proportion of pixels in the image (image resolution is 3840×2160), which significantly improves the saliency and resolution of target features and helps the model to extract and identify small targets more accurately.

[0070] (2) The focal length range of the second lens is selected from 41mm to 80mm. Within this focal length range, the lens angle of view and focal length are relatively moderate, which can take into account both the shooting range and details, and is suitable for shooting small targets within a medium distance (50-100 meters). Therefore, this lens can provide relatively clear target details within this distance range, while still retaining enough environmental background information, which is convenient for target positioning and recognition. Its focal length selection ensures that the target has a moderate pixel density in the image (image resolution is 3840×2160), ensuring the optimization of feature extraction and recognition effects;

[0071] (3) The focal length range of the third lens is selected from 63mm to 120mm. Lenses within this focal length range provide a narrower field of view, which is suitable for long-distance shooting. Therefore, this lens can shoot small targets at a distance (101-150 meters). Using a lens with this focal length range can significantly increase the pixel ratio of the target in the image (image resolution is 3840×2160), improve the resolution of target features, and clearly capture target details even at a distance, thereby enhancing detection accuracy.

[0072] (4) The fourth lens has a focal length range of 84mm to 160mm, suitable for extremely long-distance shooting, and is specifically designed to capture small targets at a distance. Therefore, this lens can capture small targets at extremely long distances (151-200 meters). The high focal length lens can significantly magnify small targets at a distance, allowing them to occupy a sufficient pixel area in the image (image resolution of 3840×2160), thereby enhancing the salience and resolution of target features and ensuring efficient and accurate identification and detection of small targets at a distance.

[0073] S4. Based on the small target captured in step S3, the high-resolution image is divided into multiple image blocks using an image segmentation method, and then input into the improved YOLOv8 network model (640×640) for detection.

[0074] Specifically, the image captured in step S3 is segmented using an image segmentation method, dividing the high-resolution image into multiple image blocks. The image is divided into four parts along the long axis (3840 pixels), each 960 pixels, and into three parts along the short axis (2160 pixels), each 720 pixels, resulting in 12 image blocks. Each image block is 960×720 pixels in size, which, while not 640×640, is as close as possible to the desired input size. Each segmented image block is directly input into the improved YOLOv8 network model. The improved YOLOv8 model (image detection model) then processes each image block, detecting and identifying small targets. The detection results of each image block are integrated and mapped back to the coordinate system of the original high-resolution image, ensuring that each detected small target can be accurately located in the original full-frame image and maintaining the integrity of the detection results.

[0075] This embodiment provides a method for detecting small targets at long distances. First, target images are acquired to form an initial dataset, which is then preprocessed to generate a processed dataset. Subsequently, the LabelImg tool is used for annotation. To meet real-time detection requirements, the YOLOv8 backbone network is improved by replacing the original Darknet53 backbone network with the lightweight MobileNetV3, thereby improving detection speed and efficiency. Training is then performed on the improved YOLOv8 model. Given the significant limitations of the YOLOv8 algorithm in recognizing small targets at long distances, particularly those within a 200-meter range, lenses with different focal lengths are designed to achieve continuous coverage from near to far (up to 200 meters), acquiring high-resolution images to be identified. For situations where direct image scaling is insufficient for effective recognition of small targets, image segmentation methods are used to segment the high-resolution images to accurately extract the target region. Through the comprehensive application of network model optimization, multi-focal length shooting, and image segmentation techniques, efficient detection and recognition of small targets at long distances is achieved.

[0076] Example 2

[0077] This embodiment provides a long-range small target detection system, which specifically includes:

[0078] The target physical width calculation module is configured to: obtain the sensor diagonal size, aspect ratio and resolution; calculate the sensor physical width based on the sensor diagonal size and aspect ratio; calculate the physical size of each pixel based on the sensor physical width and resolution; and convert the pixel width occupied by the target in the image into the target physical width.

[0079] The focal length range determination module is configured to: acquire the actual width of the target and the distance range between the target and the sensor, and determine the focal length range in combination with the physical width of the target;

[0080] The target detection module is configured to: acquire images within the focal length range, obtain multiple image blocks through image segmentation methods, detect targets in each image block using an image detection model, and then integrate the targets in all image blocks to obtain the detection result.

[0081] It should be noted that each module in this embodiment corresponds one-to-one with each step in Embodiment 1, and their specific implementation processes are the same, so they will not be repeated here.

[0082] Example 3

[0083] This embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of a long-range small target detection method as described in Embodiment 1 above.

[0084] Example 4

[0085] This embodiment provides a computer device, such as... Figure 2 As shown, the system includes a computer-readable storage medium 1003, a processor 1001, a communication interface 1002, and a computer program stored on the computer-readable storage medium 1003 and executable on the processor 1001. The processor 1001, communication interface 1002, and computer-readable storage medium 1003 can be connected via a bus or other means. The communication interface 1002 is used to receive and transmit data. When the processor 1001 executes the program, it implements the steps of the long-range small target detection method described in Embodiment 1 above.

[0086] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0087] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0088] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0089] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for detecting small targets at long distances, characterized in that, include: Obtain the sensor's diagonal size, aspect ratio, and resolution. Based on the sensor's diagonal size and aspect ratio, calculate the sensor's physical width. Based on the sensor's physical width and resolution, calculate the physical size of each pixel. Convert the pixel width occupied by the target in the image into the target's physical width. Obtain the actual width of the target and the distance range between the target and the sensor, and determine the focal length range by combining the physical width of the target; The image within the specified focal length range is acquired, and multiple image blocks are obtained through image segmentation. After the target in each image block is detected by an image detection model, the targets in all image blocks are integrated to obtain the detection result. The physical width of the sensor is: Wherein, scaling factor The resolution is P1×P2, the aspect ratio is Q1:Q2, and the diagonal size is H. The physical width of the target: The physical width of the sensor is The resolution is P1×P2. The pixel width occupied by the target in the image; The relationship between the focal length range and the distance range between the target and the sensor is as follows: In the formula, f is the focal length, and D is the distance between the lens and the target. The actual width of the target is [value], and the physical width of the target is [value]. .

2. The method for detecting small targets at long distances as described in claim 1, characterized in that, The image segmentation method is as follows: the image is divided into n1 parts along the long axis and n2 parts along the short axis, and finally the image is divided into n1 + n2 image blocks.

3. The method for detecting small targets at long distances as described in claim 1, characterized in that, The image detection model uses YOLOv8, and the backbone network of YOLOv8 is MobileNetV3.

4. The method for detecting small targets at long distances as described in claim 1, characterized in that, The dataset used to train the image detection model was obtained by annotating images acquired at various angles and heights after brightness enhancement, rotation, and flipping.

5. A long-range micro-target detection system, characterized in that, include: The target physical width calculation module is configured to: obtain the sensor diagonal size, aspect ratio and resolution; calculate the sensor physical width based on the sensor diagonal size and aspect ratio; calculate the physical size of each pixel based on the sensor physical width and resolution; and convert the pixel width occupied by the target in the image into the target physical width. The focal length range determination module is configured to: acquire the actual width of the target and the distance range between the target and the sensor, and determine the focal length range in combination with the physical width of the target; The target detection module is configured to: acquire images within the focal length range, obtain multiple image blocks through image segmentation methods, detect targets in each image block using an image detection model, and then integrate the targets in all image blocks to obtain the detection result; The physical width of the sensor is: Wherein, scaling factor The resolution is P1×P2, the aspect ratio is Q1:Q2, and the diagonal size is H. The physical width of the target: The physical width of the sensor is The resolution is P1×P2. The pixel width occupied by the target in the image; The relationship between the focal length range and the distance range between the target and the sensor is as follows: In the formula, f is the focal length, and D is the distance between the lens and the target. The actual width of the target is [value], and the physical width of the target is [value]. .

6. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps in the method for detecting small targets at a distance as described in any one of claims 1-4.

7. A computer device comprising a computer-readable storage medium, a processor, and a computer program stored on the computer-readable storage medium and executable on the processor, characterized in that, When the processor executes the program, it implements the steps in the method for detecting small targets at long distances as described in any one of claims 1-4.

Images