Uav backhaul picture parameter adjustment method and system

Abstract

The application relates to a UAV back transmission picture parameter adjustment method and system; the method comprises the following steps: acquiring a current frame picture collected and transmitted by a UAV; picture content recognition is performed on the current frame picture to determine the picture main body content type corresponding to the current frame picture; overall color information of the current frame picture is extracted; a basic picture parameter adjustment value is determined according to the picture main body content type; a picture parameter correction value is determined according to the overall color information, the basic picture parameter adjustment value is corrected according to the picture parameter correction value, and a group of picture parameter adjustment values corresponding to the current frame picture is generated; the whole frame picture parameter of the current frame picture is synchronously adjusted according to the group of picture parameter adjustment values, and an adjusted back transmission picture is obtained. The scheme can match the picture parameter adjustment of the UAV back transmission picture with the main body content type and the main color distribution of the current frame picture, and keep the whole frame adjustment consistency.

Description

Technical Field

[0001] This invention relates to the field of drone image processing technology, and in particular to a method and system for adjusting parameters of drone-transmitted images. Background Technology

[0002] In scenarios such as aerial photography, inspection, security, rescue, surveying, and routine image acquisition, drones typically need to transmit images captured by their cameras back to ground-based equipment, mobile terminals, remote control displays, or servers in real-time or near real-time. Operators can determine the drone's current position, the subject being filmed, and the content of the footage based on the transmitted images. Therefore, the display status of the transmitted images affects the operator's observation and judgment of the content.

[0003] In existing technologies, when displaying images transmitted from drones, fixed image parameters are typically used, or users manually adjust image parameters such as color, contrast, and saturation. Fixed image parameters are difficult to adapt to images with different subjects, such as water surfaces, buildings, and people, which differ in color distribution, texture density, and visual focus. Manual adjustment requires repeated intervention from the operator during flight or viewing, which is not conducive to ensuring that the display state of the transmitted image corresponds to the current image content in a timely manner.

[0004] Furthermore, while some image processing methods can adjust the image based on overall color or brightness, adjusting parameters solely based on the overall color state without considering the type of main content in the current frame can lead to a lack of simplistic adjustment criteria. For example, images of water, architecture, and people may have similar overall brightness or color proportions, but their suitable color, contrast, and saturation parameters may differ. Using fixed adjustment values ​​or adjusting only based on the primary color can result in insufficient matching between the adjusted returned image and the current image content.

[0005] Therefore, the main technical problem with the existing drone-transmitted image parameter adjustment method is that when the main content and overall color state of the drone-transmitted image change, it is difficult to generate image parameter adjustment values ​​that match the entire frame based on the main content type and main color distribution of the current frame, resulting in insufficient matching between the image parameter adjustment of the transmitted image and the current image content. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to provide a method and system for adjusting the parameters of drone-transmitted images, so as to at least solve the problem that existing drone-transmitted image parameter adjustment methods are difficult to generate image parameter adjustment values ​​that match the whole frame based on the main content type and main color distribution of the current frame when the main content and overall color state of the drone-transmitted image changes, resulting in insufficient matching between the image parameter adjustment of the transmitted image and the current image content.

[0007] To solve the above-mentioned technical problems, the present invention provides a method for adjusting parameters of UAV-transmitted images, comprising: Acquire the current frame image captured and transmitted back by the drone; Perform image content recognition on the current frame to determine the main content type of the current frame; Extract the overall color information of the current frame, wherein the overall color information includes color statistical information used to characterize the main color distribution of the current frame; Determine the basic image parameter adjustment values ​​based on the type of the main content of the image; Based on the overall color information, determine the image parameter correction value, and adjust the basic image parameter adjustment value according to the image parameter correction value to generate a set of image parameter adjustment values ​​corresponding to the current frame image; According to the set of image parameter adjustment values, the entire frame image parameters of the current frame image are synchronously adjusted to obtain the adjusted return image. The parameters of the entire frame include at least one of color parameters, contrast parameters, and saturation parameters.

[0008] The present invention also provides a drone-transmitted image parameter adjustment system, comprising: The image acquisition module is used to acquire the current frame image captured and transmitted back by the drone; The content recognition module is used to recognize the content of the current frame and determine the main content type of the current frame. The color extraction module is used to extract the overall color information of the current frame, wherein the overall color information includes color statistical information that characterizes the main color distribution of the current frame. The basic parameter determination module is used to determine the basic image parameter adjustment value according to the main content type of the image. The adjustment value generation module is used to determine the image parameter correction value based on the overall color information, and to correct the basic image parameter adjustment value based on the image parameter correction value, thereby generating a set of image parameter adjustment values ​​corresponding to the current frame image. The image adjustment module is used to synchronously adjust the entire frame image parameters of the current frame image according to the set of image parameter adjustment values ​​to obtain the adjusted return image. The parameters of the entire frame include at least one of color parameters, contrast parameters, and saturation parameters.

[0009] Compared with the prior art, the present invention has at least the following beneficial effects: This invention, after acquiring and transmitting the current frame image from a drone, first identifies the content of the current frame to determine the type of main content. Then, it extracts the overall color information of the current frame, including color statistics representing the main color distribution. Therefore, image parameter adjustment no longer relies solely on fixed parameters or a single image factor, but simultaneously uses the main content type and main color distribution of the current frame as the basis for parameter generation. This allows for the generation of an image parameter adjustment basis that better matches the current image content, taking into account changes in the main content and color state in the drone's transmitted image.

[0010] This invention determines basic image parameter adjustment values ​​based on the main content type of the image, and determines image parameter correction values ​​based on overall color information. The basic image parameter adjustment values ​​are then corrected using the correction values ​​to generate a set of image parameter adjustment values ​​corresponding to the current frame. This process ensures a clear sequential logic in the generation of image parameter adjustment values: the main content type determines the basic adjustment direction that matches the current image content, and overall color information corrects the basic adjustment values ​​based on the main color distribution of the current frame. Therefore, compared to methods that only use fixed adjustment values ​​or only adjust based on color information, this invention can reduce the possibility of mismatches between image parameter adjustment values ​​and the actual content of the current frame.

[0011] This invention synchronously adjusts the entire frame's image parameters according to the aforementioned set of image parameter adjustment values ​​to obtain the adjusted transmitted image. Since these image parameter adjustment values ​​are jointly determined by the main content type and overall color information of the image, and serve as the basis for synchronous adjustment of the entire frame's image parameters, it can maintain the consistency of the entire transmitted image adjustment while ensuring that the adjusted transmitted image matches the main content type and primary color distribution of the current frame's image. This solves the problem of insufficient matching between existing UAV transmitted image parameter adjustments and the current image content. Attached Figure Description

[0012] Figure 1 This is a schematic diagram illustrating the application scenario of the drone image transmission parameter adjustment system in an embodiment of the present invention.

[0013] Figure 2This is a flowchart illustrating the method for adjusting the parameters of the UAV transmitted image in an embodiment of the present invention.

[0014] Figure 3 This is a flowchart illustrating the process of determining the main content type of the screen in an embodiment of the present invention.

[0015] Figure 4 This is a schematic diagram of the process for determining candidate screen regions in an embodiment of the present invention.

[0016] Figure 5 This is a schematic diagram illustrating the process of forming the main screen area in an embodiment of the present invention.

[0017] Figure 6 This is a schematic diagram of the overall color information acquisition process in an embodiment of the present invention.

[0018] Figure 7 This is a schematic diagram illustrating the process of generating basic screen parameter adjustment values ​​in an embodiment of the present invention.

[0019] Figure 8 This is a schematic diagram illustrating the process of generating a set of image parameter adjustment values ​​in an embodiment of the present invention.

[0020] Figure 9 This is a flowchart illustrating the parameter adjustment system for drone-transmitted images in an embodiment of the present invention. Detailed Implementation

[0021] The technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings and embodiments. The described embodiments are used to illustrate the implementation of the present invention and are not intended to limit the scope of protection of the present invention. Where there is no conflict, technical features in different embodiments can be combined with each other.

[0022] refer to Figure 1In one embodiment, the UAV image transmission parameter adjustment system 100 can be applied to image transmission scenarios between the UAV 200 and the processing device 300. The UAV 200 may include an image acquisition device 201 and a communication device 202. The image acquisition device 201 is used to acquire images corresponding to the flight process of the UAV, and the communication device 202 is used to transmit the acquired images to the processing device 300. The processing device 300 may be a ground terminal device, a mobile terminal, a remote controller display terminal, a server, or a processing device installed on the UAV 200 itself. The processing device 300 may include a processor 301, a memory 302, and a display device 303. The memory 302 may store a program for executing the method of this embodiment, a preset set of main content types, the correspondence between main content types and basic image parameter adjustment values, a reference color range, and a preset adjustment range. After reading the current frame, the processor 301 performs image content recognition, overall color information extraction, determination of basic image parameter adjustment values, determination of image parameter correction values, and synchronous adjustment of the entire frame's image parameters. The adjusted return image is then output to the display device 303, or to other display, storage, or transmission modules.

[0023] In this embodiment, the current frame refers to a single frame transmitted back by the UAV 200 and used as the current processing object. This current frame can originate from any frame in a continuous video stream, or from image frames acquired and transmitted back by the UAV 200 at preset time intervals. The overall frame parameters refer to the display parameters acting on the current frame as a whole, and may include at least one of color parameters, contrast parameters, and saturation parameters. Color parameters may correspond to hue, color temperature, color gain, or color channel adjustment values; contrast parameters may correspond to the adjustment amount of brightness difference; and saturation parameters may correspond to the adjustment amount of color vividness. These parameters can be represented by absolute values, relative increments, proportional coefficients, or table-derived values; this embodiment does not limit their specific data format.

[0024] refer to Figure 2 In one embodiment, the method for adjusting the parameters of the drone's transmitted images includes steps S10 to S60.

[0025] S10: Acquire the current frame image captured and transmitted back by the drone.

[0026] In step S10, the processing device 300 receives the transmitted images sent by the drone 200 through the communication device 202, and reads the current frame image that needs to be adjusted from the transmitted images. The current frame image can be the original transmitted frame without image parameter adjustment, or it can be an image frame that has been decoded, scaled, or format-converted. If the drone 200 sends the images in the form of a video stream, the processing device 300 can read the image frames one by one according to the video decoding order, and take the currently arrived or currently to be displayed frame as the current frame image; if the drone 200 sends the images in the form of an image frame sequence, the processing device 300 can read the current frame image according to the timestamp, frame sequence number, or buffer queue.

[0027] The data format of the current frame can include RGB, YUV, HSV, or other image formats that can represent pixel color information. If subsequent color statistics need to be based on a specific color space, the processing device 300 can perform color space conversion on the current frame after step S10. For example, when the current frame is in YUV format and color categories need to be counted later, it can be converted to RGB or HSV format; when the current frame already contains color values ​​that can be used for statistics, it can directly proceed to subsequent processing. The above format conversion is a preprocessing of the current frame and does not change the source attribute of the current frame as the image transmitted back by the UAV.

[0028] In practical applications, the current frame may correspond to different drone scenarios. For example, when a drone is flying over a lake, the current frame may primarily feature the water surface; when a drone is conducting city patrols, the current frame may primarily feature buildings; and when a drone is filming or recording activities, the current frame may primarily feature people. Different types of main content in the frame correspond to different color distributions and visual attention requirements. Therefore, this embodiment does not directly use fixed parameters for display after acquiring the current frame. Instead, it uses the current frame as a common input for subsequent content recognition and color statistics to generate a set of image parameter adjustment values ​​that match the current frame.

[0029] S20, perform image content recognition on the current frame to determine the main content type of the current frame.

[0030] In step S20, the processing device 300 performs image content recognition on the current frame to determine the main content type of the current frame. The main content type refers to the category of image content in the current frame that has a major impact on image parameter adjustment; it can be determined based on the main area, main object, or main visual content in the image. The main content type may include water surfaces, buildings, people, or other types such as roads, vegetation, sky, and vehicles. For the embodiments listed in the specification, the preset main content type set includes at least one of water surfaces, buildings, and people.

[0031] Content recognition can be achieved through image classification, region detection, semantic segmentation, or a combination of these methods. For example, the processing device 300 can first divide the current frame into multiple candidate frame regions, and then identify the candidate content type of each candidate frame region; alternatively, it can first perform semantic recognition on the entire frame to obtain the content tags corresponding to each pixel or region, and then determine the main content type of the image based on the region proportion and positional relationship. This embodiment does not limit the specific name of the recognition algorithm; the key is that the recognition result needs to be able to determine the main content type of the current frame and use this main content type as the data basis for subsequently determining the adjustment values ​​of basic image parameters.

[0032] In one embodiment, image content recognition can be implemented by a content recognition model or program executed by a processing device. The content recognition model may include an input unit, a feature extraction unit, and a classification output unit. The input unit receives the current frame or candidate image regions; the feature extraction unit extracts color distribution, texture, edge, contour, or semantic features from the input image; the classification output unit outputs candidate content types based on the extracted features. Candidate content types may include water surfaces, buildings, people, roads, vegetation, sky, or vehicles. This content recognition model can be trained using image samples labeled with content types or implemented using pre-configured image classification rules. This embodiment does not limit the specific model structure or algorithm name, as long as it can output candidate content types for subsequent determination of the main image region.

[0033] In drone-transmitted footage, the main content may change due to variations in flight perspective, camera rotation, or subject movement. If only a rough classification of the entire frame is performed, it may be difficult to determine which content type should be used for parameter adjustment when multiple content types appear simultaneously. Therefore, this embodiment combines the proportion and positional relationship of candidate image regions within the current frame to determine the main content. Candidate image regions with larger proportions typically have a more significant impact on the overall frame display; candidate image regions closer to the center are usually closer to the content the operator is focused on. By considering both proportion and positional relationship, when drone-transmitted footage contains multiple content types, the most suitable main content type for parameter adjustment can be determined.

[0034] In one embodiment, reference Figure 3 Step S20 includes the following sub-steps.

[0035] S210, the current frame is divided into regions to obtain at least one candidate frame region.

[0036] In step S210, the processing device 300 divides the current frame into at least one candidate frame region. A candidate frame region refers to a frame region used for content type recognition and subject determination. Candidate frame regions can be obtained through regular grid partitioning, such as dividing the current frame into several rectangular regions horizontally and vertically; they can also be obtained through image segmentation, edge detection, connected component analysis, object detection boxes, or semantic region extraction. If regular grid partitioning is used, the boundary of each candidate frame region can be determined by pixel coordinates; if semantic region extraction is used, candidate frame regions can correspond to consecutive pixel regions of the same content type.

[0037] The purpose of region segmentation is to separate parts of the current frame that may have different content types, so that subsequent recognition processes can obtain candidate content types for local areas. For example, in a drone city inspection video, the current frame may contain the sky at the top, buildings in the middle, and roads at the bottom; if the entire frame is classified directly, it may not accurately reflect the main position of the buildings in the image. Through candidate image region segmentation, the processing device 300 can identify the candidate regions corresponding to the sky, buildings, and roads respectively, providing a basis for determining the main image region in the subsequent process.

[0038] S220, Identify the candidate content type corresponding to each of the candidate screen regions.

[0039] In step S220, the processing device 300 performs content recognition on each candidate screen region to obtain the corresponding candidate content type. The candidate content type refers to the category of the main screen content in the candidate screen region. For a candidate screen region, one candidate content type or multiple candidate content types and their confidence levels can be output. When outputting multiple candidate content types, the type with the highest confidence level can be selected as the candidate content type corresponding to that candidate screen region, or multiple types can be retained for subsequent rule determination.

[0040] For example, candidate image regions containing lake surface textures and large areas of blue-green pixels can be identified as water surfaces; candidate image regions containing regular edges, walls, windows, or roof features can be identified as buildings; and candidate image regions containing human outlines, heads, torsos, or limbs can be identified as people. These identifications can be achieved through classification models, detection models, segmentation models, or rules based on color and shape. If model processing is used, the input can be the image patch corresponding to the candidate image region, and the output can be the candidate content type and its corresponding confidence score. If rule processing is used, the processing device 300 can determine the candidate content type based on features such as color range, texture direction, edge density, or region shape.

[0041] S230, based on the candidate content type corresponding to each candidate screen area, determine the candidate screen area that meets the preset subject type condition from each candidate screen area.

[0042] In step S230, the processing device 300 determines whether each candidate image region meets the preset subject type conditions based on the candidate content type. The preset subject type conditions are used to filter candidate areas related to the adjustment of image parameters in this application from the candidate image regions. The preset subject type conditions may include candidate content types belonging to a preset subject content type set. The preset subject content type set can be pre-configured according to common application scenarios of drone-transmitted images, such as including at least one of water surface, building, and human figures.

[0043] The purpose of setting preset subject type conditions is to prioritize content types that are strongly related to the adjustment of image parameters. For example, water-related images typically have large areas of blue, green, or gray, so adjustments may focus more on color shift and saturation; architectural images usually contain many lines, edges, and neutral colors, so adjustments may focus more on contrast and color naturalness; and images of people usually involve skin tones, so adjustments may focus more on the appropriateness of color and saturation. By setting preset subject type conditions, the processing device 300 can first filter out candidate image areas that are more likely to affect the image display adjustment strategy.

[0044] S240, based on the proportion and positional relationship of candidate image regions that meet the preset subject type conditions in the current frame, determine the subject image region, and determine the candidate content type corresponding to the subject image region as the subject content type of the image.

[0045] In step S240, the processing device 300 further determines the candidate image regions that meet the preset subject type conditions. The region proportion can represent the area ratio of the candidate image region in the current frame, and the positional relationship can represent the relationship between the candidate image region and the center region, boundary, or other reference position of the current frame. In this embodiment, the candidate image regions with a greater impact on the entire frame are preferentially determined based on the region proportion; when the region proportion cannot clearly determine the subject image region, a supplementary determination is made based on the positional relationship.

[0046] For example, if the building area occupies 60% of the current frame, the water area occupies 20%, and the people area occupies 5%, then the processing device 300 can identify the building area as the main image area and classify the building type as the main content type of the image. As another example, if the water area and building area in the current frame are relatively close in size and neither meets the preset proportion condition, then if the water area covers the central area of ​​the image, the water area can be identified as the main image area. Therefore, the main content type of the image is not determined by any single recognition result, but by a combination of candidate content types, area proportions, and positional relationships.

[0047] In one embodiment, reference Figure 4 Step S230 includes the following sub-steps.

[0048] S231, determine whether the candidate content type belongs to the preset main content type set.

[0049] In step S231, the processing device 300 reads a preset set of main content types and matches the candidate content type corresponding to each candidate screen area with the preset set of main content types. The preset set of main content types can be stored in the memory 302, or it can be issued by the server or configured by the user. This set can include at least one or more of the following: water surface, building, and people. If a candidate content type belongs to this set, it means that the candidate screen area meets the preset main content type condition; if a candidate content type does not belong to this set, it means that the candidate screen area is not considered as a priority candidate main content area for the time being.

[0050] S232, if there are candidate screen areas whose candidate content type belongs to the preset main content type set, the candidate screen areas whose candidate content type belongs to the preset main content type set are determined as candidate screen areas that satisfy the preset main type conditions.

[0051] In step S232, when the processing device 300 detects that the candidate content type of at least one candidate image region belongs to a preset main content type set, these candidate image regions are taken as candidate image regions that meet the preset main content type conditions. At this time, the subsequent main image region determination process is only performed in these candidate image regions that meet the conditions, which can make the main content determination focus on the preset common main content of drones.

[0052] S233, if there are no candidate screen areas whose candidate content type belongs to the preset main content type set, each of the identified candidate screen areas is determined as a candidate screen area that satisfies the preset main content type condition.

[0053] In step S233, when no candidate image region belonging to the preset main content type set is identified in the current frame, the processing device 300 does not directly terminate the image parameter adjustment process. Instead, it determines all identified candidate image regions as candidate image regions that meet the preset main content type conditions. This processing method is used to form a fallback mechanism. For example, if the current frame mainly consists of roads, farmland, or sky, and the preset main content type set does not yet include these types, the system can still determine the main image region based on the area proportion and positional relationship, and proceed to the subsequent basic image parameter adjustment value determination process. If no corresponding record is found subsequently, the preset default image parameter adjustment value can be used to ensure the continuous execution of the adjustment process.

[0054] In one embodiment, reference Figure 5 Step S240 includes the following sub-steps.

[0055] S241, calculate the area ratio of each candidate screen region that meets the preset subject type condition in the current frame.

[0056] In step S241, the processing device 300 calculates the area ratio of each candidate image region that meets the preset subject type conditions in the current frame. The area ratio can be calculated based on the number of pixels in the candidate image region and the total number of pixels in the current frame. If the candidate image region is a rectangle, the area can be calculated based on the width and height of the rectangle; if the candidate image region is an irregular region, the number of pixels contained in the region can be counted.

[0057] Here is an example of how to calculate the area percentage: r_i = A_i / A_total Where r_i represents the area proportion of the i-th candidate image region, A_i represents the pixel area of ​​the i-th candidate image region, and A_total represents the total pixel area of ​​the current frame. This formula is used to illustrate one method of calculating the area proportion and does not limit the candidate image region to be represented in a specific shape.

[0058] S242, determine whether the candidate image area with the largest area ratio meets the preset ratio condition.

[0059] In step S242, the processing device 300 determines the candidate image region with the largest area ratio from the candidate image regions that meet the preset subject type conditions, and determines whether the candidate image region meets the preset ratio condition. The preset ratio condition can be that the area ratio is greater than or equal to a preset area threshold, or that the area ratio has a significant advantage relative to other candidate image regions. For example, the preset ratio condition can be set to an area ratio of not less than 30%, or the difference between the candidate image region with the largest area ratio and the second largest candidate image region is not less than a preset difference.

[0060] By using preset proportion conditions for judgment, it is possible to avoid determining the main image area based solely on minor differences in area when multiple areas are similar in size. For example, if the water area accounts for 32%, the building area accounts for 31%, and the character area accounts for 8%, directly selecting the largest water area might overlook the fact that the building area is located in the center of the image and is more of a focus for the operator. Therefore, when the candidate image area with the largest area proportion does not meet the preset proportion conditions, this embodiment switches to determining the position based on the central area.

[0061] S243, if the candidate image area with the largest area ratio satisfies the preset ratio condition, the candidate image area with the largest area ratio is determined as the main image area.

[0062] In step S243, when the candidate image area with the largest area proportion meets the preset proportion condition, it indicates that the candidate image area has a high area proportion in the current frame and is sufficient to represent the main content of the current frame. The processing device 300 determines the candidate image area as the main image area and its corresponding candidate content type as the main content type of the image. This method is suitable for scenarios where the main subject of the image is relatively clear, such as when a drone takes an aerial view of a lake, the water surface area occupies most of the area of ​​the entire frame; when a drone inspects the exterior walls of a building, the building area occupies most of the area of ​​the entire frame.

[0063] S244, if the candidate image area with the largest area ratio does not meet the preset ratio condition, the main image area is determined according to the positional relationship between each candidate image area that meets the preset main image type condition and the center area of ​​the current frame image.

[0064] In step S244, when the candidate image region with the largest area does not meet the preset proportion condition, the processing device 300 further determines the main image region based on the positional relationship between each candidate image region and the central region of the current frame. The central region can be a rectangular region, a circular region, or other preset shaped region centered on the center point of the current frame. When observing the transmitted image, the drone operator usually pays more attention to the content near the center of the image. Therefore, when the area proportion is insufficient to uniquely determine the main subject, the central region can serve as a supplementary determination criterion.

[0065] In one embodiment, step S244 includes the following sub-steps.

[0066] K1 determines the central region of the current frame.

[0067] In step K1, the processing device 300 determines the central region based on the size of the current frame. For example, if the width of the current frame is W and the height is H, the center point can be determined as (W / 2, H / 2), and then a rectangular central region with a width of W_c and a height of H_c can be determined with this center point as the center. W_c and H_c can be a preset ratio of the width and height of the current frame, such as 20% to 40%, or can be set according to the size of the display device or the application scenario.

[0068] K2 determines whether there is a candidate image region that overlaps with the central region.

[0069] In step K2, the processing device 300 determines whether each candidate image region that meets the preset subject type conditions overlaps with the central region. Overlap can be determined by the number of overlapping pixels, the overlapping area, or the boundary relationship between the candidate image region and the central region. If the overlapping area is greater than zero, or greater than the preset minimum overlapping area, then the candidate image region can be considered to overlap with the central region.

[0070] K3, in the case of candidate image areas that overlap with the central area, the candidate image area with the largest overlap area with the central area is determined as the main image area.

[0071] In step K3, if one or more candidate image regions overlap with the central region, the processing device 300 calculates the overlap area between each overlapping candidate image region and the central region, and selects the candidate image region with the largest overlap area as the main image region. This rule can select the region with a stronger association with the central region when multiple candidate regions are close to the center at the same time. For example, if the areas of the person region and the building region in the current frame are similar, and the overlap area between the person region and the central region is larger, then the processing device 300 can determine the person region as the main image region.

[0072] K4, in the absence of a candidate image area that overlaps with the central area, the candidate image area that is closest to the central area is determined as the main image area.

[0073] In step K4, if no candidate frame area overlaps with the central area, the processing device 300 calculates the distance between each candidate frame area and the central area, and determines the candidate frame area with the closest distance as the main frame area. The distance can be calculated based on the Euclidean distance between the center point of the candidate frame area and the center point of the current frame, or it can be calculated based on the minimum distance between the boundary of the candidate frame area and the boundary of the central area.

[0074] An example of how distance is calculated is as follows: d_i=sqrt((cx_i-cx_0)^2+(cy_i-cy_0)^2) Where d_i represents the distance between the i-th candidate frame region and the center point of the current frame, cx_i and cy_i represent the x and y coordinates of the center point of the i-th candidate frame region, and cx_0 and cy_0 represent the x and y coordinates of the center point of the current frame. This formula illustrates one method of distance calculation; in practical applications, other distance calculation methods that reflect the proximity of positions can also be used.

[0075] S30, extract the overall color information of the current frame, the overall color information including color statistics information used to characterize the main color distribution of the current frame.

[0076] In step S30, the processing device 300 extracts overall color information from the current frame. Overall color information refers to the color distribution information obtained from the perspective of the entire current frame, and is not limited to a certain local area. Color statistical information is used to characterize the main color distribution of the current frame, and may include the pixel proportion of each color category, the main colors, the average value of color channels, the color channel distribution range, or other statistical results that can reflect the color distribution of the image.

[0077] The purpose of extracting overall color information is to correct the adjustment values ​​of basic image parameters. The type of the main content of an image reflects its content category, but the same content type may have different color states under different environments. For example, in images of water surfaces, a lake may appear blue-green on a sunny day and grayish on a cloudy day; in images of buildings, the main colors of glass curtain wall buildings and brick wall buildings may differ significantly; and in images of people, lighting conditions and background color will affect the overall color distribution. Therefore, this embodiment, after determining the type of the main content of the image, also extracts overall color information to correct the adjustment values ​​of basic image parameters based on the actual color state of the current frame.

[0078] In one embodiment, reference Figure 6 Step S30 includes the following sub-steps.

[0079] S310, Obtain the color values ​​of multiple pixels in the current frame.

[0080] In step S310, the processing device 300 obtains the color values ​​of multiple pixels from the current frame. The pixels can be all pixels in the current frame, or a subset of pixels selected according to a preset sampling rule. Using all pixels yields more complete color statistics; using sampled pixels reduces computational load and is suitable for real-time processing scenarios. Sampling methods can include equal-interval sampling, random sampling, row-column sampling, or region-based sampling.

[0081] Color values ​​can be RGB, HSV, YUV, or other color spaces. If the current frame is in RGB format, the color value can include the R, G, and B channel values; if the current frame is in HSV format, the color value can include hue, saturation, and lightness. The choice of color space can be determined based on subsequent color classification and parameter correction rules.

[0082] S320, classify the multiple pixels according to the color range to which the color value belongs to obtain at least one color category.

[0083] In step S320, the processing device 300 categorizes multiple pixels according to the color range to which their color values ​​belong. The color range can be preset, for example, dividing colors into categories such as blue, green, yellow, red, gray, white, and black. For RGB color values, categorization can be based on the magnitude relationship between channel values ​​and threshold ranges; for HSV color values, categorization can be based on hue range, saturation range, and brightness range.

[0084] For example, when a pixel's hue falls within the blue range and its saturation reaches a certain threshold, it can be categorized as blue; when a pixel's RGB channel values ​​are close and its brightness is in the middle range, it can be categorized as gray. Color category settings do not need to cover a detailed classification of all natural colors; they only need to reflect the main color distribution of the current frame and support subsequent image parameter adjustments.

[0085] S330, Calculate the pixel percentage of each color category in the current frame to obtain the color statistics information.

[0086] In step S330, the processing device 300 counts the number of pixels contained in each color category and calculates the pixel percentage of each color category in the current frame. The pixel percentage can represent the distribution degree of a certain color category in the entire frame and is a form of color statistical information.

[0087] One example of how to calculate pixel percentage is as follows: p_k = N_k / N_total Where p_k represents the percentage of pixels belonging to the k-th color category, N_k represents the number of pixels belonging to the k-th color category, and N_total represents the total number of pixels included in the statistics. If all pixels are counted, N_total is the total number of pixels in the current frame; if sampling is used, N_total is the number of sampled pixels.

[0088] S340, determine the main color of the current frame based on the color category with the highest pixel percentage, and use the color statistics information and the main color as the overall color information.

[0089] In step S340, the processing device 300 compares the pixel proportion of each color category and determines the color corresponding to the color category with the highest pixel proportion as the dominant color of the current frame. The dominant color represents the dominant color category in the current frame. For example, if the blue category has the highest pixel proportion, the dominant color can be blue; if the gray category has the highest pixel proportion, the dominant color can be gray. The processing device 300 outputs the color statistics information and the dominant color together as overall color information to subsequent steps.

[0090] In some cases, the pixel proportions of multiple color categories may be similar. The processing device 300 can further determine the primary color by combining the main content type of the image. For example, under the main content type of people, priority can be given to the color distribution related to skin tone or the area of ​​the person; under the main content type of water, the focus can be on the proportion changes in the blue, green, and gray ranges. The above processing is an optional implementation method, the purpose of which is to make the overall color information better serve the subsequent image parameter correction.

[0091] S40, determine the basic image parameter adjustment value according to the type of the main content of the image.

[0092] In step S40, the processing device 300 determines basic image parameter adjustment values ​​based on the image subject content type determined in step S20. Basic image parameter adjustment values ​​refer to a set of initial adjustment values ​​determined based on the image subject content type before correction is made according to the actual color state of the current frame. Basic image parameter adjustment values ​​may include at least one of color parameters, contrast parameters, and saturation parameters.

[0093] In one embodiment, the correspondence between the main content type and the basic image parameter adjustment values ​​can be pre-set according to the common display requirements of the drone's transmitted images. For example, water surface images can correspond to a lower saturation increase value and a medium color parameter adjustment value to avoid the water surface color being overly vibrant; building images can correspond to a medium contrast adjustment value and a lower color parameter adjustment value to maintain the discernibility of building edges and textures; and people images can correspond to a lower saturation adjustment value and a color parameter adjustment value related to skin tone range to reduce the possibility of color distortion in the people area. The above correspondence can be stored using parameter tables, level tables, or mapping rules, and its specific values ​​can be determined according to the display device, application scenario, and user configuration.

[0094] The reason why different types of subject matter in a scene correspond to different basic adjustment values ​​is that different subjects have different visual focuses. For example, water scenes usually need to maintain a natural water color, avoiding oversaturation or color cast; architectural scenes usually need to maintain clear edges and textures, which can be addressed with appropriate contrast adjustments; and scenes of people usually need to maintain natural skin tones and subject colors, which can be addressed with more cautious color and saturation adjustments. Basic image parameter adjustment values ​​can be determined through preset correspondences, giving the system a stable initial adjustment basis when dealing with different types of subject matter.

[0095] In one embodiment, reference Figure 7 Step S40 includes the following sub-steps.

[0096] S410: Obtain the correspondence between the preset main content type and the basic screen parameter adjustment value.

[0097] In step S410, the processing device 300 reads from the memory 302 the preset correspondence between main content types and basic image parameter adjustment values. This correspondence can be a table, configuration file, mapping function, or rule set. The correspondence can record one or more of the basic adjustment values ​​for color parameters, contrast parameters, and saturation parameters corresponding to different main content types.

[0098] For example, water surfaces can correspond to the first set of basic adjustment values, buildings can correspond to the second set, and people can correspond to the third set. Basic adjustment values ​​can be represented as incremental values, such as increasing color parameter 'a', increasing contrast parameter 'b', and decreasing saturation parameter 'c'; they can also be represented as proportional coefficients or level values, such as color parameter level C1, contrast parameter level K2, and saturation parameter level S1. This embodiment does not require the basic image parameter adjustment values ​​to use specific units, as long as they can be recognized by subsequent image adjustment modules and applied to the current frame.

[0099] S420, based on the type of the main content of the screen, query the corresponding record in the correspondence relationship.

[0100] In step S420, the processing device 300 uses the image main content type obtained in step S20 to query the corresponding record in the correspondence relationship. The corresponding record may contain one or more sets of basic image parameter adjustment values ​​corresponding to the image main content type. If the image main content type is water surface, then the corresponding record for water surface is queried; if the image main content type is architecture, then the corresponding record for architecture is queried; if the image main content type is people, then the corresponding record for people is queried.

[0101] S430, if a corresponding record is found, the basic screen parameter adjustment value in the corresponding record is used as the basic screen parameter adjustment value.

[0102] In step S430, when the processing device 300 finds a corresponding record that matches the type of the main content of the image, it reads the basic image parameter adjustment value from the corresponding record and uses it as the basic image parameter adjustment value for the current frame. If the corresponding record includes multiple parameter items, all parameter items can be read, or only the currently enabled parameter items can be read. For example, if the system only enables the contrast and saturation parameters, only the basic contrast adjustment value and the basic saturation adjustment value from the corresponding record can be read.

[0103] S440, if no corresponding record is found, the preset default screen parameter adjustment value is used as the basic screen parameter adjustment value.

[0104] In step S440, when there is no corresponding record in the correspondence that matches the main content type of the current screen, the processing device 300 reads the preset default screen parameter adjustment value and uses it as the basic screen parameter adjustment value. This default value can be a zero adjustment value that does not change the screen, or it can be a general adjustment value applicable to general returned screens. Setting the preset default screen parameter adjustment value can prevent the adjustment process from being interrupted due to the main content type not being configured, ensuring that the UAV's returned screens can still complete subsequent parameter generation and display output.

[0105] S50, determine the image parameter correction value based on the overall color information, and correct the basic image parameter adjustment value based on the image parameter correction value to generate a set of image parameter adjustment values ​​corresponding to the current frame.

[0106] In step S50, the processing device 300 determines the image parameter correction value based on the overall color information obtained in step S30, and uses the image parameter correction value to correct the basic image parameter adjustment value obtained in step S40, generating a set of image parameter adjustment values ​​corresponding to the current frame. The image parameter correction value is used to reflect the adjustment requirements of the basic adjustment value for the actual color state of the current frame. For example, if the basic saturation adjustment value corresponding to a certain type of subject is high, but the main color of the current frame is already biased towards highly saturated colors, then the image parameter correction value can reduce the saturation adjustment range; if the current frame is generally grayish, then the image parameter correction value can appropriately correct the color or contrast adjustment value.

[0107] In one embodiment, reference Figure 8 Step S50 includes the following sub-steps.

[0108] S510, determine the reference color range corresponding to the main content type of the image based on the main content type of the image.

[0109] In step S510, the processing device 300 determines the corresponding reference color range based on the type of the main content of the image. The reference color range refers to a pre-defined range of color distributions corresponding to a specific type of main content in the image. The reference color range can be represented by color category, color channel range, hue range, saturation range, or brightness range. For example, water surfaces can correspond to a reference color range related to blue, green, or gray; buildings can correspond to a reference color range related to gray, white, brown, or other neutral colors; and people can correspond to a reference color range related to skin tone.

[0110] The reference color gamut does not require precisely defining all color variations in a natural scene; rather, it serves as a benchmark for the dominant color of the current frame. By selecting the reference color gamut based on the type of content in the image, it avoids comparing all image content to the same reference color. For example, a predominantly bluish hue in a water scene might be within the normal range, while a predominantly bluish hue in a people scene might indicate the need for color correction.

[0111] In one embodiment, the reference color range can be set according to color category or color space interval. For example, the reference color range corresponding to water surface can include a blue range, a green range, or a gray range; the reference color range corresponding to building can include a gray range, a white range, a brown range, or other neutral color ranges; and the reference color range corresponding to people can include a preset skin tone range. The reference color range can be represented by a hue interval, a color channel interval, or a set of color categories. After comparing the main color of the current frame with the corresponding reference color range, the processing device can determine whether the main color matches the type of the main content of the current frame and determine the image parameter correction value accordingly.

[0112] S520, compare the main color of the current frame with the reference color range to determine the color offset information of the main color relative to the reference color range.

[0113] In step S520, the processing device 300 compares the primary color determined in step S340 with the reference color range determined in step S510 to determine color offset information. The color offset information indicates whether the primary color is within the reference color range, the direction of deviation, and the degree of deviation. If the primary color is within the reference color range, it can be considered that the primary color of the current frame basically matches the type of the main content of the image; if the primary color is outside the reference color range, the degree of deviation can be determined based on its distance from the boundary of the reference color range.

[0114] Taking hue value as an example, if the hue value of the primary color is H_main and the reference color range is [H_low, H_high], then an example rule for color offset can be expressed as: Delta_H=0, when H_low<=H_main and H_main<=H_high Delta_H=H_main-H_high, when H_main>H_high Delta_H=H_main-H_low, when H_main <H_low Delta_H represents the hue offset of the primary color relative to the reference color range. This rule is only used to illustrate one method of generating color offset information. When using the RGB color space or other color spaces, color offset information can also be determined based on color channel differences, color category distances, or differences in color distribution.

[0115] S530, determine the corresponding image parameter correction value based on the color offset information.

[0116] In step S530, the processing device 300 determines the image parameter correction value based on the color offset information. The image parameter correction value may include at least one of color parameter correction value, contrast parameter correction value, and saturation parameter correction value. The image parameter correction value can be determined by a mapping table or by proportional calculation. For example, when the color offset is large, a larger color parameter correction value can be generated; when the color offset is small, a smaller correction value can be generated; when the color offset information is zero offset information, the image parameter correction value can be zero or close to zero.

[0117] An example calculation method is as follows: M_j=k_jDelta_j Where M_j represents the correction value of the j-th image parameter, Delta_j represents the color offset related to the j-th image parameter, and k_j represents the preset correction coefficient. This formula is used to illustrate the correspondence between the image parameter correction value and the color offset information. In practical applications, segmented mapping, lookup table mapping, or bit-level rules can also be used to determine the image parameter correction value.

[0118] When the primary color is within the reference color range, the color offset information is determined to be zero offset information or offset information less than a preset offset threshold. In this case, the processing device 300 can generate a zero image parameter correction value or a small image parameter correction value to avoid over-correcting the current frame image that is already within the reference range.

[0119] S540, the image parameter correction value and the basic image parameter adjustment value are combined to obtain the candidate image parameter adjustment value.

[0120] In step S540, the processing device 300 combines the image parameter correction value with the basic image parameter adjustment value. The combination method can be additive combination, proportional combination, level combination, or lookup table combination. For example, when both the basic image parameter adjustment value and the image parameter correction value are expressed as incremental values, they can be added together to obtain the candidate image parameter adjustment value.

[0121] An example of a synthesis method is as follows: V_j=B_j+M_j Where V_j represents the adjustment value of the j-th candidate image parameter, B_j represents the adjustment value of the j-th base image parameter, and M_j represents the correction value of the j-th image parameter. This formula can be applied to one or more parameters among color, contrast, and saturation parameters. If a scaling factor or level value is used, the base value and correction value can also be combined into a candidate image parameter adjustment value according to a preset compositing rule.

[0122] S550, determine whether the adjustment value of the candidate image parameter is within the preset adjustment range.

[0123] In step S550, the processing device 300 determines whether the candidate image parameter adjustment values ​​are within a preset adjustment range. The preset adjustment range is used to limit the boundary of the final output set of image parameter adjustment values ​​to avoid abnormal image display due to excessive correction. The preset adjustment range can be set separately for different parameters. For example, color parameters can correspond to a first range, contrast parameters can correspond to a second range, and saturation parameters can correspond to a third range.

[0124] The judgment rule can be expressed as: V_min_j <= V_j and V_j <= V_max_j Where V_min_j represents the lower limit of adjustment for the j-th frame parameter, and V_max_j represents the upper limit of adjustment for the j-th frame parameter. When the adjustment value of a candidate frame parameter meets this condition, it means that the candidate value is within the allowable range; otherwise, amplitude limiting processing is required.

[0125] S560, if the candidate image parameter adjustment value is within the preset adjustment range, the candidate image parameter adjustment value is determined as a set of image parameter adjustment values ​​corresponding to the current frame image.

[0126] In step S560, if the candidate image parameter adjustment value is within a preset adjustment range, the processing device 300 directly determines the candidate image parameter adjustment value as a set of image parameter adjustment values ​​corresponding to the current frame image. A set of image parameter adjustment values ​​may include one parameter value or multiple parameter values. For example, when only the contrast parameter is adjusted, the set of image parameter adjustment values ​​may only include the contrast adjustment value; when the color parameter, contrast parameter, and saturation parameter are adjusted simultaneously, the set of image parameter adjustment values ​​may include three corresponding adjustment values.

[0127] S570, if the candidate image parameter adjustment value exceeds the preset adjustment range, the candidate image parameter adjustment value is subjected to amplitude limiting processing according to the preset adjustment range, and the amplitude-limited candidate image parameter adjustment value is determined as a set of image parameter adjustment values ​​corresponding to the current frame image.

[0128] In step S570, if the adjustment value of the candidate image parameter exceeds the preset adjustment range, the processing device 300 performs amplitude limiting processing on the adjustment value of the candidate image parameter. Amplitude limiting processing can set candidate values ​​below the lower limit as the lower limit value and candidate values ​​above the upper limit as the upper limit value.

[0129] An example of a limiting method is as follows: V_j_final=max(V_min_j,min(V_j,V_max_j)) Wherein, V_j_final represents the j-th frame parameter adjustment value after clipping, V_j represents the candidate frame parameter adjustment value before clipping, and V_min_j and V_max_j represent the lower and upper limits of the preset adjustment range, respectively. Through this process, the final set of frame parameter adjustment values ​​is within the parameter range allowed by the system, which facilitates its stable application to the current frame.

[0130] S60, according to the set of image parameter adjustment values, the entire frame image parameters of the current frame image are synchronously adjusted to obtain the adjusted return image.

[0131] In step S60, the processing device 300 synchronously adjusts the overall frame parameters of the current frame according to a set of frame parameter adjustment values ​​generated in step S50. Synchronous adjustment means using this set of frame parameter adjustment values ​​as common adjustment values ​​for the entire current frame, applying them to the overall frame parameters, rather than generating different adjustment values ​​for different candidate frame areas. The overall frame parameters may include at least one of color parameters, contrast parameters, and saturation parameters.

[0132] For example, when a set of image parameter adjustment values ​​includes a color parameter adjustment value C, a contrast parameter adjustment value K, and a saturation parameter adjustment value S, the processing device 300 can process all pixels or the entire frame of image data in the current frame according to the same set of parameters. For the color parameter, the color channel gain or hue can be adjusted; for the contrast parameter, the difference between pixel brightness and average brightness can be adjusted; for the saturation parameter, the vividness of the color can be adjusted. The specific image transformation method can be implemented according to the display device or image processing software, but the input is always the same set of image parameter adjustment values.

[0133] In one embodiment, contrast adjustment may employ the following example rule: Y_prime = Y_mean + K(Y - Y_mean) Where Y represents the pixel brightness value before adjustment, Y_prime represents the pixel brightness value after adjustment, Y_mean represents the average brightness value of the current frame or a preset area, and K represents the contrast parameter adjustment value or the contrast coefficient calculated from it. This formula is only used to illustrate one method of adjusting the contrast of the entire frame and does not limit this application to using this formula. Color parameters and saturation parameters can also be adjusted by color space transformation, channel gain, or lookup table methods.

[0134] In step S60, the processing device 300 obtains the adjusted return image and can output it to the display device 303 for display, or store, forward, or use it for subsequent image processing. Since the set of image parameter adjustment values ​​are generated based on the main content type and overall color information of the image, the adjusted return image can reflect the main content and main color distribution of the current image while maintaining consistency throughout the entire frame.

[0135] Through steps S10 to S60 described above, after acquiring the current frame image captured and transmitted back by the UAV, this embodiment uses image content recognition to determine the type of main content corresponding to the current frame image and extracts overall color information including color statistics. The type of main content reflects the content category in the current frame image that has a major impact on image parameter adjustment, and the overall color information reflects the main color distribution of the current frame image. Both serve as the basis for subsequent parameter generation, enabling image parameter adjustment to no longer rely solely on fixed parameters or single image factors, but to establish a correspondence with the main content and color state of the current frame image.

[0136] Based on this, this embodiment determines basic image parameter adjustment values ​​according to the type of the main content of the image, determines image parameter correction values ​​according to the overall color information, and uses the image parameter correction values ​​to correct the basic image parameter adjustment values, generating a set of image parameter adjustment values ​​corresponding to the current frame. This set of image parameter adjustment values ​​serves as the basis for synchronous adjustment of the entire frame's image parameters, acting on the entire frame's image parameters. This ensures that the adjusted transmitted image maintains consistency across the entire frame while adapting to the main content type and primary color distribution of the current frame, thereby improving the problem of insufficient matching between the drone's transmitted image parameter adjustment and the current image content.

[0137] refer to Figure 9 In one embodiment, the UAV image transmission parameter adjustment system may include an image acquisition module 110, a content recognition module 120, a color extraction module 130, a basic parameter determination module 140, an adjustment value generation module 150, and an image adjustment module 160.

[0138] The image acquisition module 110 is used to acquire the current frame image captured and transmitted back by the UAV. The image acquisition module 110 may include a communication interface, a decoding unit, or an image buffer unit, or it may be implemented by a software program in the processing device 300. The image acquisition module 110 receives image data sent by the UAV 200 and outputs the currently unprocessed frame image as the current frame image.

[0139] The content recognition module 120 is used to recognize the content of the current frame and determine the main content type of the current frame. The content recognition module 120 can perform processing such as region division, candidate content type recognition, preset main type condition filtering, region proportion calculation and center region position relationship determination, and output the determined main content type of the frame to the basic parameter determination module 140 and the adjustment value generation module 150.

[0140] The color extraction module 130 is used to extract the overall color information of the current frame, which includes color statistics information representing the main color distribution of the current frame. The color extraction module 130 can obtain the color values ​​of multiple pixels in the current frame, classify them according to color range to obtain at least one color category, calculate the pixel percentage of each color category, and determine the main color of the current frame.

[0141] The basic parameter determination module 140 is used to determine basic image parameter adjustment values ​​based on the main content type of the image. The basic parameter determination module 140 can read the preset correspondence between the main content type and the basic image parameter adjustment values, and query the corresponding record according to the main content type of the image; if no corresponding record is found, the basic parameter determination module 140 can read the preset default image parameter adjustment value.

[0142] The adjustment value generation module 150 is used to determine the image parameter correction value based on the overall color information, and to correct the basic image parameter adjustment value based on the image parameter correction value, thereby generating a set of image parameter adjustment values ​​corresponding to the current frame. The adjustment value generation module 150 can determine a reference color range based on the main content type of the image, compare the main color of the current frame with the reference color range to determine color offset information, and determine the image parameter correction value based on the color offset information; then, it combines the image parameter correction value with the basic image parameter adjustment value, and performs amplitude limiting processing when necessary.

[0143] The image adjustment module 160 is used to synchronously adjust the entire frame image parameters of the current frame image according to the set of image parameter adjustment values ​​to obtain the adjusted return image. The image adjustment module 160 can use the set of image parameter adjustment values ​​as common parameters for the entire frame, apply them to at least one of the color parameters, contrast parameters, and saturation parameters of the current frame image, and output the adjusted return image.

[0144] In one embodiment, the image acquisition module 110, content recognition module 120, color extraction module 130, basic parameter determination module 140, adjustment value generation module 150, and image adjustment module 160 can be implemented by the processor 301 executing the program stored in the memory 302; alternatively, they can be implemented by one or more of multiple processors, image processing chips, field-programmable gate arrays, or application-specific integrated circuits. Each module can be deployed in the same processing device 300, or separately in the UAV 200, ground terminal equipment, mobile terminal, or server, as long as the aforementioned data transmission and processing can be completed.

[0145] The above-disclosed embodiments are merely preferred embodiments of this application and should not be construed as limiting the scope of this application. Therefore, any equivalent variations made in accordance with the description of this application are still within the scope of this application.

Claims

1. A method for adjusting a picture parameter of a backhaul of a UAV, characterized in that, include: Acquire the current frame image captured and transmitted back by the drone; Perform image content recognition on the current frame to determine the main content type of the current frame; Extract the overall color information of the current frame, wherein the overall color information includes color statistical information used to characterize the main color distribution of the current frame; Determine the basic image parameter adjustment values ​​based on the type of the main content of the image; Based on the overall color information, determine the image parameter correction value, and adjust the basic image parameter adjustment value according to the image parameter correction value to generate a set of image parameter adjustment values ​​corresponding to the current frame image; According to the set of image parameter adjustment values, the entire frame image parameters of the current frame image are synchronously adjusted to obtain the adjusted return image. The parameters of the entire frame include at least one of color parameters, contrast parameters, and saturation parameters.

2. The method for adjusting parameters of UAV transmitted images according to claim 1, characterized in that, The step of identifying the content of the current frame and determining the main content type of the current frame includes: The current frame is divided into regions to obtain at least one candidate frame region; Identify the candidate content type corresponding to each of the candidate image regions; Based on the candidate content type corresponding to each candidate screen area, determine the candidate screen area that meets the preset subject type condition from each candidate screen area; Based on the proportion and positional relationship of candidate image regions that meet the preset subject type conditions in the current frame, the subject image region is determined, and the candidate content type corresponding to the subject image region is determined as the subject content type of the image.

3. The method for adjusting parameters of UAV transmitted images according to claim 2, characterized in that, The preset subject type condition includes that the candidate content type belongs to a preset subject content type set, and the preset subject content type set includes at least one of water surface, building and human figures; The step of determining candidate screen regions that meet preset subject type conditions from each candidate screen region based on the candidate content type corresponding to each candidate screen region includes: If there are candidate screen areas whose candidate content type belongs to the preset main content type set, the candidate screen areas whose candidate content type belongs to the preset main content type set are determined as candidate screen areas that meet the preset main content type conditions. If there are no candidate screen areas whose candidate content type belongs to the preset main content type set, each of the identified candidate screen areas will be determined as candidate screen areas that meet the preset main content type conditions.

4. The method for adjusting parameters of UAV transmitted images according to claim 2, characterized in that, The step of determining the main image region based on the region proportion and positional relationship of candidate image regions that meet the preset main image type conditions in the current frame includes: Calculate the area ratio of each candidate image region that meets the preset subject type condition in the current frame; Determine whether the candidate image area with the largest area ratio meets the preset ratio condition; If the candidate image area with the largest area ratio meets the preset ratio condition, the candidate image area with the largest area ratio is determined as the main image area. If the candidate image area with the largest area ratio does not meet the preset ratio condition, the main image area is determined according to the positional relationship between each candidate image area that meets the preset main image type condition and the center area of ​​the current frame.

5. The method for adjusting parameters of UAV transmitted images according to claim 4, characterized in that, The step of determining the main image region based on the positional relationship between each candidate image region that satisfies the preset main image type condition and the center region of the current frame image includes: Determine the central region of the current frame; Determine whether there are candidate image regions that overlap with the central region; In the case where there are candidate image regions that overlap with the central region, the candidate image region with the largest overlap area with the central region is determined as the main image region. If there is no candidate image area that overlaps with the central area, the candidate image area that is closest to the central area is determined as the main image area.

6. The method for adjusting parameters of UAV transmitted images according to claim 1, characterized in that, The step of extracting the overall color information of the current frame includes: Obtain the color values ​​of multiple pixels in the current frame; The multiple pixels are categorized according to the color range to which their color values ​​belong, resulting in at least one color category; The color statistics information is obtained by calculating the pixel percentage of each color category in the current frame. The dominant color of the current frame is determined based on the color category with the highest pixel percentage, and the color statistics and the dominant color are used as the overall color information.

7. The method for adjusting parameters of UAV transmitted images according to claim 1, characterized in that, The step of determining the basic image parameter adjustment value based on the main content type of the image includes: Obtain the correspondence between preset main content types and basic image parameter adjustment values; Based on the type of the main content of the image, query the corresponding record in the correspondence relationship; If a corresponding record is found, the basic screen parameter adjustment value in that corresponding record shall be used as the basic screen parameter adjustment value. If no corresponding record is found, the preset default screen parameter adjustment value will be used as the basic screen parameter adjustment value. Different content types correspond to different basic adjustment values ​​for at least one of the following: color parameters, contrast parameters, and saturation parameters.

8. The method for adjusting parameters of UAV image transmission according to claim 6, characterized in that, The step of determining the image parameter correction value based on the overall color information includes: Based on the type of the main content of the image, determine the reference color range corresponding to the type of the main content of the image; The main color of the current frame is compared with the reference color range to determine the color offset information of the main color relative to the reference color range; Determine the corresponding image parameter correction value based on the color offset information; Wherein, when the primary color is within the range of the reference color, the color offset information is determined to be zero offset information or offset information less than a preset offset threshold.

9. The method for adjusting parameters of UAV image transmission according to claim 8, characterized in that, The step of correcting the basic image parameter adjustment value based on the image parameter correction value to generate a set of image parameter adjustment values ​​corresponding to the current frame includes: The image parameter correction value is combined with the basic image parameter adjustment value to obtain the candidate image parameter adjustment value; Determine whether the adjustment value of the candidate image parameter is within the preset adjustment range; If the candidate image parameter adjustment value is within the preset adjustment range, the candidate image parameter adjustment value is determined as a set of image parameter adjustment values ​​corresponding to the current frame image; If the adjustment value of the candidate image parameter exceeds the preset adjustment range, the adjustment value of the candidate image parameter is limited according to the preset adjustment range, and the adjusted value of the candidate image parameter after the limit processing is determined as a set of image parameter adjustment values ​​corresponding to the current frame image.

10. A system for adjusting parameters of UAV-transmitted images, characterized in that, include: The image acquisition module is used to acquire the current frame image captured and transmitted back by the drone; The content recognition module is used to recognize the content of the current frame and determine the main content type of the current frame. The color extraction module is used to extract the overall color information of the current frame, wherein the overall color information includes color statistical information that characterizes the main color distribution of the current frame. The basic parameter determination module is used to determine the basic image parameter adjustment value according to the main content type of the image. The adjustment value generation module is used to determine the image parameter correction value based on the overall color information, and to correct the basic image parameter adjustment value based on the image parameter correction value, thereby generating a set of image parameter adjustment values ​​corresponding to the current frame image. The image adjustment module is used to synchronously adjust the entire frame image parameters of the current frame image according to the set of image parameter adjustment values ​​to obtain the adjusted return image. The parameters of the entire frame include at least one of color parameters, contrast parameters, and saturation parameters.

Images