Unmanned aerial vehicle image optimization method and device, terminal equipment and storage medium

Abstract

The invention discloses an unmanned aerial vehicle image optimization method and apparatus, a terminal device and a storage medium. The method comprises the steps of obtaining an image acquisition instruction, physical characteristic parameters of a to-be-acquired object and environmental parameters; carrying out detection calibration on a to-be-collected object according to the image collection instruction, the physical characteristic parameters and the environmental parameters to obtain a calibration distance with the to-be-collected object; according to the calibration distance and the image acquisition instruction, performing image acquisition on a to-be-acquired object to obtain an initial acquisition image; performing image edge detection according to the initial acquisition image to obtain an edge detection image; performing gradient refining processing according to each pixel point of the edge detection image to obtain a target optimization image; and transmitting the target optimized image to a preset image processing center so as to carry out image processing according to the target optimized image. According to the invention, the image acquisition precision of the unmanned aerial vehicle is improved.

Description

Technical Field

[0001] The present invention relates to the field of image processing technology, and in particular to a method, device, terminal equipment and storage medium for optimizing unmanned aerial vehicle images. Background Art

[0002] Drone technology has developed rapidly in recent years. With its high-resolution camera, it can perform high-sensitivity imaging of the ground from high altitudes. This feature makes drones show great application potential in many fields such as forest fire prevention and earthquake relief. Drones are not only small in size, easy to carry and deploy, but also have a fast flight speed and can quickly reach the target area for image acquisition, greatly improving the efficiency of monitoring and emergency response. However, in actual applications, drones are easily disturbed by wind when flying in the air, causing the fuselage and camera to shake, making it difficult for the drone to focus and the collected images blurred. The blur of the image not only affects the visual quality of the image, but also leads to inaccurate results in subsequent image processing. Summary of the invention

[0003] The embodiments of the present invention provide a method, apparatus, terminal device and storage medium for optimizing drone images, which can effectively solve the problem that drones in the prior art are easily disturbed by wind when flying in the air, causing the fuselage and camera to shake, thereby making it difficult for the drone to focus and the collected images to be blurred.

[0004] An embodiment of the present invention provides a method for optimizing a drone image, comprising:

[0005] Obtaining image acquisition instructions, physical characteristic parameters of the object to be acquired, and environmental parameters;

[0006] Perform detection calibration on the object to be collected according to the image collection instruction, the physical characteristic parameters and the environmental parameters to obtain a calibration distance to the object to be collected;

[0007] According to the calibration distance and the image acquisition instruction, an image of the object to be acquired is acquired to obtain an initial acquired image;

[0008] Performing image edge detection according to the initial collected image to obtain an edge detection image;

[0009] Performing gradient refinement processing on each pixel of the edge detection image to obtain a target optimized image;

[0010] The target optimized image is transmitted to a preset image processing center so that image processing is performed according to the target optimized image.

[0011] Furthermore, the environmental parameters include: background radiation intensity; the physical characteristic parameters include: the self-radiation intensity of the object to be collected;

[0012] Performing detection calibration on the object to be collected according to the image collection instruction, the physical characteristic parameters and the environmental parameters to obtain a calibration distance to the object to be collected, including:

[0013] According to the image acquisition instruction, calculating the difference between the self-radiation intensity and the background radiation intensity to obtain a radiation intensity difference;

[0014] A calibration distance to the object to be collected is obtained by multiplying the radiation intensity difference, the physical characteristic parameter and the environmental parameter.

[0015] Furthermore, the physical characteristic parameters also include: normalized detection rate, detector bandwidth, detection micro-element area and noise equivalent power;

[0016] The determination of the normalized detection rate includes:

[0017] The normalized detection rate is obtained by performing a product calculation based on the detection micro-element area, the detector bandwidth and the noise equivalent power.

[0018] Further, performing image edge detection according to the initial collected image to obtain an edge detected image includes:

[0019] According to the initial acquisition image and the preset detection window, the initial acquisition image is divided to obtain a plurality of sub-region images;

[0020] Noise points are removed according to the center point of each sub-region image and a preset correction factor to obtain a first target image after noise reduction;

[0021] Calculate a horizontal partial derivative and a vertical partial derivative according to the first target image and a preset spatial scale coefficient;

[0022] Calculating according to the first target image, the horizontal partial derivative, and the vertical partial derivative to obtain a gradient direction and a gradient value of each pixel point of the first target image;

[0023] The edge features of the first target image are extracted according to the gradient direction and the gradient value to obtain an edge detection image.

[0024] Further, the gradient direction and gradient value of each pixel point of the first target image are obtained by calculation according to the first target image, the horizontal partial derivative, and the vertical partial derivative, including:

[0025] Performing a convolution calculation according to the first target image and the horizontal partial derivative to obtain a first convolution result;

[0026] Performing convolution calculation according to the first target image and the vertical partial derivative to obtain a second convolution result;

[0027] Performing an inverse tangent function calculation according to the first convolution result and the second convolution result to obtain a gradient direction of each pixel point of the first target image;

[0028] A square sum calculation is performed according to the first convolution result and the second convolution result to obtain a gradient value of each pixel point of the first target image.

[0029] Furthermore, the edge feature includes any one of the following items and combinations thereof: image color feature, image texture feature, image shape feature and image spatial relationship feature.

[0030] Further, performing gradient refinement processing on each pixel of the edge detection image to obtain a target optimized image includes:

[0031] Determine the gradient direction and gradient value of each pixel on the edge detection image according to the edge detection image;

[0032] Taking each pixel point on the edge detection image as the current pixel point, respectively, and comparing the gradient value of the current pixel point with a preset gradient threshold; wherein the preset gradient threshold includes: a first gradient threshold value for representing a high threshold value and a second gradient threshold value for representing a low threshold value; the first gradient threshold value is greater than the second gradient threshold value;

[0033] When the gradient value of the current pixel is greater than or equal to the first gradient threshold, the current pixel is determined to be an image edge point, and a first thinning result is obtained;

[0034] When the gradient value of the current pixel point is less than or equal to the second gradient threshold, determining that the current pixel point is not an image edge point, and obtaining a second thinning result;

[0035] When the gradient value of the current pixel is less than the first gradient threshold and greater than the second gradient threshold, the pixel corresponding to the maximum gradient value is taken as the image edge point according to the gradient value of the current pixel and the gradient values ​​of the pixels adjacent to the corresponding gradient direction, to obtain a third thinning result;

[0036] Pixel points are spliced ​​according to the first refinement result, the second refinement result, and the third refinement result to obtain a target optimized image.

[0037] As an improvement of the above solution, another embodiment of the present invention provides a drone image optimization device, including:

[0038] A data acquisition module is used to obtain image acquisition instructions, physical characteristic parameters of the object to be acquired, and environmental parameters;

[0039] A detection calibration module, used to perform detection calibration on the object to be captured according to the image capture instruction, the physical characteristic parameters and the environmental parameters, and obtain a calibration distance to the object to be captured;

[0040] An image acquisition module, used to acquire an image of the object to be acquired according to the calibration distance and the image acquisition instruction, to obtain an initial acquired image;

[0041] An edge detection module, used to perform image edge detection based on the initial collected image to obtain an edge detection image;

[0042] An image optimization module, used for performing gradient refinement processing on each pixel point of the edge detection image to obtain a target optimized image;

[0043] The image sending module is used to transmit the target optimized image to a preset image processing center so that image processing is performed according to the target optimized image.

[0044] Another embodiment of the present invention provides a terminal device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and when the processor executes the computer program, a drone image optimization method as described in the above embodiment is implemented.

[0045] Another embodiment of the present invention provides a computer-readable storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute a drone image optimization method described in the above embodiment.

[0046] By implementing the present invention, at least the following beneficial effects are achieved:

[0047] The present invention provides a method, device, terminal equipment and storage medium for optimizing unmanned aerial vehicle images. The method can obtain image acquisition instructions, physical characteristic parameters of an object to be acquired and environmental parameters; perform detection calibration on the object to be acquired according to the image acquisition instructions, the physical characteristic parameters and the environmental parameters to obtain a calibration distance with the object to be acquired; perform image acquisition on the object to be acquired according to the calibration distance and the image acquisition instructions to obtain an initial acquisition image; perform image edge detection according to the initial acquisition image to obtain an edge detection image; perform gradient refinement processing on each pixel point of the edge detection image to obtain a target optimized image; transmit the target optimized image to a preset image processing center so that image processing is performed according to the target optimized image. By acquiring the physical characteristics and environmental parameters of the object to be collected, and combining the image acquisition instructions for detection calibration, the drone can accurately adjust the distance and angle between the camera and the object to be collected, thereby offsetting the impact of wind interference on the camera stability to a certain extent, reducing the focusing difficulty caused by the shaking of the fuselage and camera, and improving the stability of image acquisition; after the initial collected image is processed by edge detection and gradient refinement, it can highlight the edge features in the image, enhance the image clarity and contrast, reduce the impact of image blur caused by wind interference, significantly improve the visual quality of the image, and improve the accuracy of drone image acquisition. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a flowchart of a method for optimizing drone images provided by one embodiment of the present invention;

[0049] Figure 2 It is a structural schematic diagram of a drone image optimization device provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0050] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0051] See also Figure 1 , is a flow chart of a method for optimizing a drone image provided by an embodiment of the present invention, comprising:

[0052] S1, obtaining image acquisition instructions, physical characteristic parameters of the object to be acquired, and environmental parameters;

[0053] Specifically, when the UAV is performing a flight mission and is in a flight state, it can receive image acquisition instructions, physical characteristic parameters of the object to be acquired, and environmental parameters near the object to be acquired. The environmental parameters include: background radiation intensity, atmospheric transmittance, and effective incident aperture area; the physical characteristic parameters include: self-radiation intensity, frontal area, surface reflectivity, normalized detection rate, equivalent temperature difference, detection micro-element area, detector bandwidth, and minimum signal-to-noise ratio of the object to be acquired.

[0054] S2, performing detection calibration on the object to be collected according to the image collection instruction, the physical characteristic parameters and the environmental parameters to obtain a calibration distance to the object to be collected;

[0055] Specifically, the UAV image optimization method is applicable to a UAV, which is equipped with a detector for detection calibration and image acquisition. The farthest distance detected by the infrared imaging guidance system on the detector is limited by two factors: imaging area and signal energy, so the selected imaging area cannot be less than the preset area threshold, for example, it cannot be less than 10×10 pixels, otherwise, the detector will not be able to identify the target. Sufficiently strong signal energy requires a sufficiently large signal-to-noise ratio, and a suitable signal-to-noise ratio threshold is selected. Only when the signal-to-noise ratio is higher than the threshold can the system identify the target. After probabilistic statistical calculation and analysis, the effective range of the uncooled infrared detector search and tracking system is scientifically calculated based on the relationship between the signal-to-noise ratio and the target radiation intensity.

[0056] Preferably, the environmental parameters include: background radiation intensity; the physical characteristic parameters include: the self-radiation intensity of the object to be collected;

[0057] Performing detection calibration on the object to be collected according to the image collection instruction, the physical characteristic parameters and the environmental parameters to obtain a calibration distance to the object to be collected, including:

[0058] According to the image acquisition instruction, calculating the difference between the self-radiation intensity and the background radiation intensity to obtain a radiation intensity difference;

[0059] A calibration distance to the object to be collected is obtained by multiplying the radiation intensity difference, the physical characteristic parameter and the environmental parameter.

[0060] In a preferred embodiment of the present invention, the detector on the drone is an infrared detector. The background radiation intensity represents the radiation intensity of the environment in which the object to be collected is located; the self-radiation intensity represents the amount of energy radiated outward by the object to be collected per unit time. First, according to the image acquisition instruction, the self-radiation intensity L is calculated. 物体 And the background radiation intensity L 背景The difference is calculated by multiplying the radiation intensity difference, the physical characteristic parameter and the environmental parameter to obtain the calibration distance to the object to be collected. The calibration distance is calculated by the following formula:

[0061]

[0062] Where r is the calibration distance between the drone and the object to be collected, L 物体 is the self-radiation intensity of the object to be collected, L 背景 is the background radiation intensity of the target background, S is the front view area of ​​the object to be collected, λ(r) is the atmospheric transmittance of the radiation of the object to be collected at a distance r, η is the reflectivity of the surface of the object to be collected, and A enp is the effective incident aperture area of ​​the optical system, D′ is the normalized detectivity, N t is the system equivalent temperature difference, A d The detection element area, that is, the light window area of ​​the detector, is used to determine the measurement point or target position. By precisely controlling the position and size of the detection element, accurate positioning and measurement of the object to be collected can be achieved. Δf is the system detector bandwidth, δ is the preset correction factor, and SNR is the minimum signal-to-noise ratio value when the system is working.

[0063] In another preferred embodiment of the present invention, the physical characteristic parameters also include: normalized detection rate, detector bandwidth, detection micro-element area and noise equivalent power; the determination of the normalized detection rate includes: performing product calculation according to the detection micro-element area, the detector bandwidth and the noise equivalent power to obtain the normalized detection rate. The normalized detection rate D' is used to describe the detection capability of the drone detector material for weak light, that is, the signal-to-noise ratio of the detector per unit area, unit amplifier bandwidth and unit radiation power. The specific calculation formula is as follows:

[0064]

[0065] Where NEP is the noise equivalent power, F is the ratio of focal length to clear aperture d, NETD is the thermal sensitivity, is the rate of change of spectral emittance with respect to temperature at relative background temperature.

[0066] By calculating the difference between the radiation intensity of the object to be collected and the background radiation intensity, the interference of background radiation on image collection can be eliminated or reduced, thereby improving the accuracy of image collection; by multiplying the radiation intensity difference, physical characteristic parameters and environmental parameters, a more accurate calibration distance can be obtained, which helps to improve the accuracy in subsequent image processing and analysis. During the flight, the drone can manually or automatically select the object to be collected. In order to improve the shooting accuracy and reduce the scrap rate of the photos taken, the calibration distance is obtained by continuously focusing on the object to be collected, so that the image taken by the drone is clearer.

[0067] S3, performing image acquisition on the object to be acquired according to the calibration distance and the image acquisition instruction to obtain an initial acquired image;

[0068] Specifically, the calibration distance also helps to ensure that the size of the object to be captured in the image is consistent with the actual size; when using the calibration distance for image capture, image distortion caused by lens distortion can be corrected, thereby improving the overall quality of the image.

[0069] S4, performing image edge detection according to the initial collected image to obtain an edge detection image;

[0070] Preferably, performing image edge detection according to the initial collected image to obtain an edge detected image includes:

[0071] According to the initial acquisition image and the preset detection window, the initial acquisition image is divided to obtain a plurality of sub-region images;

[0072] Noise points are removed according to the center point of each sub-region image and a preset correction factor to obtain a first target image after noise reduction;

[0073] Calculate a horizontal partial derivative and a vertical partial derivative according to the first target image and a preset spatial scale coefficient;

[0074] Calculating according to the first target image, the horizontal partial derivative, and the vertical partial derivative to obtain a gradient direction and a gradient value of each pixel point of the first target image;

[0075] The edge features of the first target image are extracted according to the gradient direction and the gradient value to obtain an edge detection image.

[0076] Specifically, firstly, the initial acquisition image is divided according to the initial acquisition image and the preset detection window to obtain a number of sub-region images. For example, the preset detection window is set to 5*5, and then it is divided into 8 sub-regions, so that the initial acquisition image can be divided to obtain a number of sub-region images. Then, according to the center point of each sub-region image and the preset correction factor, the noise points are removed to obtain the first target image after noise reduction; each sub-region image is divided into 2 sub-regions by the center point (x, y), and the noise points in each window are judged. When the pixel point in the window is 0, the pixel point is automatically removed. The specific judgment formula is as follows:

[0077]

[0078] Wherein, f(x,y) represents the pixel point of the first target image, δ represents the preset correction factor, and when 0+δ≤f(x,y)≤255-δ, the pixel point in the window is 0 and the pixel point is removed; in other cases, the pixel point is retained.

[0079] Then, according to the first target image and the preset spatial scale coefficient, a horizontal partial derivative and a vertical partial derivative are calculated;

[0080] Specifically, after smoothing and denoising the first target image using a partition mean filtering algorithm, a smoothed and denoised image is obtained, and according to the smoothed and denoised image and a preset spatial scale coefficient, a horizontal partial derivative and a vertical partial derivative are calculated. The formula for smoothing and denoising is as follows:

[0081] Where x and y represent the coordinates of the pixel points, and σ is the preset spatial scale coefficient. After smoothing and denoising, the partial derivative functions in the x direction (horizontal direction) and y direction (vertical direction) are calculated to obtain the horizontal partial derivative G x and the vertical partial derivative G y , the specific calculation formula is as follows:

[0082]

[0083] In another preferred embodiment of the present invention, a suitable edge detection algorithm, such as a Sobel operator, a Canny operator, or a Prewitt operator, may be selected according to the characteristics and requirements of the first target image.

[0084] Specifically, the calculation is performed according to the first target image, the horizontal partial derivative, and the vertical partial derivative to obtain the gradient direction and gradient value of each pixel point of the first target image, including:

[0085] Performing a convolution calculation according to the first target image and the horizontal partial derivative to obtain a first convolution result;

[0086] Performing convolution calculation according to the first target image and the vertical partial derivative to obtain a second convolution result;

[0087] Performing an inverse tangent function calculation according to the first convolution result and the second convolution result to obtain a gradient direction of each pixel point of the first target image;

[0088] A square sum calculation is performed according to the first convolution result and the second convolution result to obtain a gradient value of each pixel point of the first target image.

[0089] In a preferred embodiment of the present invention, a convolution calculation is performed according to the first target image, and a convolution calculation is performed according to the horizontal partial derivative to obtain a first convolution result P(i, j); a convolution calculation is performed according to the first target image and the vertical partial derivative to obtain a second convolution result Q(i, j), and the calculation formula is as follows: Then, an inverse tangent function is calculated according to the first convolution result and the second convolution result to obtain the gradient direction θ(i, j) of each pixel point of the first target image; a square sum calculation is performed according to the first convolution result and the second convolution result to obtain the gradient value S(i, j) of each pixel point of the first target image, and the calculation formula is as follows:

[0090]

[0091] In another preferred embodiment of the present invention, the parameters of the edge detection algorithm, such as threshold, convolution kernel size, etc., are adjusted to obtain the best edge detection image. For example, for the Canny operator, the high and low thresholds can be adjusted to control the strength and continuity of the edge.

[0092] Preferably, the edge feature includes any one of the following items and combinations thereof: image color feature, image texture feature, image shape feature and image spatial relationship feature.

[0093] Through edge detection, the initial acquired image is simplified to an edge detection image that only contains important edge information, which greatly reduces the amount of data and removes a large number of redundant pixels. Since the edge detection image after edge detection is more concise, the amount of calculation can be significantly reduced in subsequent processing, reducing the calculation complexity, which is especially important for UAVs to transmit images in real-time flight.

[0094] S5, performing gradient refinement processing on each pixel point of the edge detection image to obtain a target optimized image;

[0095] Specifically, according to the edge detection image, determine the gradient direction and gradient value of each pixel point on the edge detection image;

[0096] Taking each pixel point on the edge detection image as the current pixel point, respectively, and comparing the gradient value of the current pixel point with a preset gradient threshold; wherein the preset gradient threshold includes: a first gradient threshold value for representing a high threshold value and a second gradient threshold value for representing a low threshold value; the first gradient threshold value is greater than the second gradient threshold value;

[0097] When the gradient value of the current pixel is greater than or equal to the first gradient threshold, the current pixel is determined to be an image edge point, and a first thinning result is obtained;

[0098] When the gradient value of the current pixel point is less than or equal to the second gradient threshold, determining that the current pixel point is not an image edge point, and obtaining a second thinning result;

[0099] When the gradient value of the current pixel is less than the first gradient threshold and greater than the second gradient threshold, the pixel corresponding to the maximum gradient value is taken as the image edge point according to the gradient value of the current pixel and the gradient values ​​of the pixels adjacent to the corresponding gradient direction, to obtain a third thinning result;

[0100] Pixel points are spliced ​​according to the first refinement result, the second refinement result, and the third refinement result to obtain a target optimized image.

[0101] In a preferred embodiment of the present invention, for the gradient direction and gradient value of each pixel on the edge detection image, each pixel on the edge detection image is taken as the current pixel, and the gradient value of the current pixel and the preset gradient threshold are compared; wherein the preset gradient threshold includes: a first gradient threshold for characterizing a high threshold and a second gradient threshold for characterizing a low threshold; the first gradient threshold is greater than the second gradient threshold. When the gradient value S(i,j) of the current pixel is greater than or equal to the first gradient threshold, the current pixel is determined to be an image edge point, and a first refinement result is obtained; when the gradient value S(i,j) of the current pixel is less than or equal to the second gradient threshold, the current pixel is determined not to be an image edge point, and a second refinement result is obtained; then, when the gradient value of the current pixel is less than the first gradient threshold and greater than the second gradient threshold, that is, the gradient value is between the first gradient threshold and the second gradient threshold, the gradient value of the current pixel and the gradient values ​​of the pixels adjacent to the corresponding gradient direction are compared, and the pixel corresponding to the maximum gradient value is taken as the image edge point, and a third refinement result is obtained. Finally, pixel splicing is performed according to the first refinement result, the second refinement result and the third refinement result to obtain the target optimized image. Specifically, morphological operations (such as dilation and erosion) or graph theory algorithms (such as minimum spanning tree) can be used to connect the disconnected edges to form continuous edges to obtain the target optimized image.

[0102] In another preferred embodiment of the present invention, in order to further improve the accuracy and robustness of gradient optimization, an intelligent optimization algorithm, such as the ant colony optimization algorithm (ACO), can be introduced to search for the optimal solution by simulating the behavior of nature, and perform well in image edge detection under complex backgrounds. The image edge detection method based on the mutation adaptive ant colony optimization algorithm (MAACO) is an effective choice. This method transforms the image edge detection problem into an optimization problem by defining a suitable objective function, and uses the MAACO algorithm to search for the optimal edge pixel point set; the optimized edge detection image is evaluated to check whether it meets the expected requirements. In practical applications, multiple iterations and adjustments may be required to obtain the best edge detection image.

[0103] By setting the first gradient threshold (high threshold) and the second gradient threshold (low threshold), edge points in the image can be screened out more accurately. The high threshold ensures that only pixels with larger gradient values ​​are determined as edge points, thereby reducing the possibility of false detection. When the gradient value of a pixel is between the two thresholds, the pixel is compared with its adjacent pixel gradient values, and the pixel corresponding to the maximum gradient value is selected as the edge point, further improving the accuracy of the captured image.

[0104] S6. Transmitting the target optimized image to a preset image processing center so as to perform image processing according to the target optimized image.

[0105] The target optimized image is transmitted to the preset image processing center for further processing.

[0106] Infrared detection calibration is performed according to the distance of the object to be collected. The initial collected image is preprocessed while improving the image collection accuracy. The optimal image of the target is obtained according to the gradient direction and gradient value of the edge detection image, precise focus is achieved, and the image collection accuracy of the drone is improved.

[0107] By implementing this embodiment, image acquisition instructions, physical characteristic parameters and environmental parameters of the object to be acquired are obtained; detection calibration is performed on the object to be acquired according to the image acquisition instructions, the physical characteristic parameters and the environmental parameters to obtain a calibration distance with the object to be acquired; image acquisition is performed on the object to be acquired according to the calibration distance and the image acquisition instructions to obtain an initial acquisition image; image edge detection is performed according to the initial acquisition image to obtain an edge detection image; gradient refinement is performed according to each pixel point of the edge detection image to obtain a target optimized image; the target optimized image is transmitted to a preset image processing center so that image processing is performed according to the target optimized image. By obtaining the physical characteristics and environmental parameters of the object to be acquired, and performing detection calibration in combination with the image acquisition instructions, the drone can accurately adjust the distance and angle between the camera and the object to be acquired, thereby offsetting the influence of wind interference on the stability of the camera to a certain extent, reducing the focusing difficulty caused by the shaking of the fuselage and the camera, and improving the stability of image acquisition; after edge detection and gradient refinement of the initial acquisition image, the edge features in the image can be highlighted, the clarity and contrast of the image can be enhanced, the influence of image blur caused by wind interference can be reduced, the visual quality of the image can be significantly improved, and the image acquisition accuracy of the drone can be improved.

[0108] See also Figure 2 , is a schematic diagram of the structure of a drone image optimization device provided by an embodiment of the present invention, comprising:

[0109] A data acquisition module is used to obtain image acquisition instructions, physical characteristic parameters of the object to be acquired, and environmental parameters;

[0110] A detection calibration module, used to perform detection calibration on the object to be captured according to the image capture instruction, the physical characteristic parameters and the environmental parameters, and obtain a calibration distance to the object to be captured;

[0111] An image acquisition module, used to acquire an image of the object to be acquired according to the calibration distance and the image acquisition instruction, to obtain an initial acquired image;

[0112] An edge detection module, used to perform image edge detection based on the initial collected image to obtain an edge detection image;

[0113] An image optimization module, used for performing gradient refinement processing on each pixel point of the edge detection image to obtain a target optimized image;

[0114] The image sending module is used to transmit the target optimized image to a preset image processing center so that image processing is performed according to the target optimized image.

[0115] The present invention provides an unmanned aerial vehicle image optimization device, which obtains image acquisition instructions, physical characteristic parameters of an object to be acquired, and environmental parameters according to a data acquisition module; in a detection calibration module, performs detection calibration on the object to be acquired according to the image acquisition instructions, the physical characteristic parameters, and the environmental parameters to obtain a calibration distance with the object to be acquired; in an image acquisition module, performs image acquisition on the object to be acquired according to the calibration distance and the image acquisition instructions to obtain an initial acquisition image; in an edge detection module, performs image edge detection according to the initial acquisition image to obtain an edge detection image; through an image optimization module, performs gradient refinement processing according to each pixel point of the edge detection image to obtain a target optimized image; finally, through an image sending module, transmits the target optimized image to a preset image processing center, so that image processing is performed according to the target optimized image. By acquiring the physical characteristics and environmental parameters of the object to be collected, and combining the image acquisition instructions for detection calibration, the drone can accurately adjust the distance and angle between the camera and the object to be collected, thereby offsetting the impact of wind interference on the camera stability to a certain extent, reducing the focusing difficulty caused by the shaking of the fuselage and camera, and improving the stability of image acquisition; after the initial collected image is processed by edge detection and gradient refinement, it can highlight the edge features in the image, enhance the image clarity and contrast, reduce the impact of image blur caused by wind interference, significantly improve the visual quality of the image, and improve the accuracy of drone image acquisition.

[0116] It should be noted that the device embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, in the accompanying drawings of the device embodiments provided by the present invention, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines. A person of ordinary skill in the art may understand and implement it without paying any creative effort.

[0117] Those skilled in the art can clearly understand that, for the sake of convenience and brevity, the specific working process of the device described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0118] Another embodiment of the present invention further provides a terminal device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor implements a drone image optimization method as described in the above embodiment when executing the computer program. The terminal device may be a computing device such as a desktop computer, a notebook, a PDA, and a cloud server. The terminal device may include, but is not limited to, a processor and a memory.

[0119] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the terminal device, and uses various interfaces and lines to connect various parts of the entire terminal device.

[0120] The memory can be used to store the computer program, and the processor realizes various functions of the terminal device by running or executing the computer program stored in the memory and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application required for a function, etc.; the data storage area can store data created according to the use of the mobile phone, etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device or other volatile solid-state storage device.

[0121] Another embodiment of the present invention provides a computer-readable storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute a drone image optimization method described in the above embodiment.

[0122] The storage medium is a computer-readable storage medium, and the computer program is stored in the computer-readable storage medium. When the computer program is executed by the processor, the steps of each of the above-mentioned method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium.

[0123] The above is a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the principle of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A drone image optimization method, characterized in that: include: Obtaining image acquisition instructions, physical characteristic parameters of the object to be acquired, and environmental parameters; Perform detection calibration on the object to be collected according to the image collection instruction, the physical characteristic parameters and the environmental parameters to obtain a calibration distance to the object to be collected; According to the calibration distance and the image acquisition instruction, an image of the object to be acquired is acquired to obtain an initial acquired image; Performing image edge detection according to the initial collected image to obtain an edge detection image; Performing gradient refinement processing on each pixel of the edge detection image to obtain a target optimized image; The target optimized image is transmitted to a preset image processing center so that image processing is performed according to the target optimized image.

2. A drone image optimization method as claimed in claim 1, characterized in that: The environmental parameters include: background radiation intensity; the physical characteristic parameters include: the self-radiation intensity of the object to be collected; Performing detection calibration on the object to be collected according to the image collection instruction, the physical characteristic parameters and the environmental parameters to obtain a calibration distance to the object to be collected, including: According to the image acquisition instruction, calculating the difference between the self-radiation intensity and the background radiation intensity to obtain a radiation intensity difference; A calibration distance to the object to be collected is obtained by multiplying the radiation intensity difference, the physical characteristic parameter and the environmental parameter.

3. A drone image optimization method as claimed in claim 2, characterized in that: The physical characteristic parameters also include: normalized detection rate, detector bandwidth, detection micro-element area and noise equivalent power; The determination of the normalized detection rate includes: The normalized detection rate is obtained by performing a product calculation based on the detection micro-element area, the detector bandwidth and the noise equivalent power.

4. A drone image optimization method as claimed in claim 3, characterized in that: Performing image edge detection according to the initial collected image to obtain an edge detected image includes: According to the initial acquisition image and the preset detection window, the initial acquisition image is divided to obtain a plurality of sub-region images; Noise points are removed according to the center point of each sub-region image and a preset correction factor to obtain a first target image after noise reduction; Calculate a horizontal partial derivative and a vertical partial derivative according to the first target image and a preset spatial scale coefficient; Calculating according to the first target image, the horizontal partial derivative, and the vertical partial derivative to obtain a gradient direction and a gradient value of each pixel point of the first target image; The edge features of the first target image are extracted according to the gradient direction and the gradient value to obtain an edge detection image.

5. A drone image optimization method as claimed in claim 4, characterized in that: Calculating according to the first target image, the horizontal partial derivative, and the vertical partial derivative to obtain a gradient direction and a gradient value of each pixel point of the first target image includes: Performing a convolution calculation according to the first target image and the horizontal partial derivative to obtain a first convolution result; Performing convolution calculation according to the first target image and the vertical partial derivative to obtain a second convolution result; Performing an inverse tangent function calculation according to the first convolution result and the second convolution result to obtain a gradient direction of each pixel point of the first target image; A square sum calculation is performed according to the first convolution result and the second convolution result to obtain a gradient value of each pixel point of the first target image.

6. A drone image optimization method as claimed in claim 5, characterized in that: The edge feature includes any one of the following items and their combination: image color feature, image texture feature, image shape feature and image space relationship feature.

7. A drone image optimization method as claimed in claim 6, characterized in that: Performing gradient refinement processing on each pixel of the edge detection image to obtain a target optimized image includes: Determine the gradient direction and gradient value of each pixel on the edge detection image according to the edge detection image; Taking each pixel point on the edge detection image as the current pixel point, respectively, and comparing the gradient value of the current pixel point with a preset gradient threshold; wherein the preset gradient threshold includes: a first gradient threshold value for representing a high threshold value and a second gradient threshold value for representing a low threshold value; the first gradient threshold value is greater than the second gradient threshold value; When the gradient value of the current pixel is greater than or equal to the first gradient threshold, the current pixel is determined to be an image edge point, and a first thinning result is obtained; When the gradient value of the current pixel point is less than or equal to the second gradient threshold, determining that the current pixel point is not an image edge point, and obtaining a second thinning result; When the gradient value of the current pixel is less than the first gradient threshold and greater than the second gradient threshold, the pixel corresponding to the maximum gradient value is taken as the image edge point according to the gradient value of the current pixel and the gradient values ​​of the pixels adjacent to the corresponding gradient direction, to obtain a third thinning result; Pixel points are spliced ​​according to the first refinement result, the second refinement result, and the third refinement result to obtain a target optimized image.

8. A drone image optimization device, characterized in that: include: A data acquisition module is used to obtain image acquisition instructions, physical characteristic parameters of the object to be acquired, and environmental parameters; A detection calibration module, used to perform detection calibration on the object to be captured according to the image capture instruction, the physical characteristic parameters and the environmental parameters, and obtain a calibration distance to the object to be captured; An image acquisition module, used to acquire an image of the object to be acquired according to the calibration distance and the image acquisition instruction, to obtain an initial acquired image; An edge detection module, used to perform image edge detection based on the initial collected image to obtain an edge detection image; An image optimization module, used for performing gradient refinement processing on each pixel point of the edge detection image to obtain a target optimized image; The image sending module is used to transmit the target optimized image to a preset image processing center so that image processing is performed according to the target optimized image.

9. A terminal device, characterized in that: The method comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, a method for optimizing unmanned aerial vehicle images as described in any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute a drone image optimization method as described in any one of claims 1 to 7.

Images