Image data processing method for a robot

Abstract

The application relates to the technical field of image processing, in particular to a robot image data processing method, which comprises the following steps: collecting continuous images and extracting gray scales and contours, analyzing pixel adjacency relations to divide connected regions, calculating edge density to screen priority fragments, executing path matching and sequence adjustment, and generating a robot image data processing sequence.In the application, by constructing a structure region with gray scale continuity and edge concentration, the spatial correlation of image internal pixels and the edge aggregation feature are strengthened, the accuracy of image fragment division and the effectiveness of density screening are improved, the path shunting and priority writing are realized in combination with the fragment sequence relationship, the organization structure and the circulation efficiency of the image content in the path are optimized, the continuity and the scheduling flexibility of the image sequence are improved through the complementary position and rearrangement of the differentiated path, and the extraction efficiency of the key image content and the response stability of the processing path are enhanced.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, and in particular to an image data processing method for a robot. Background Technology

[0002] Image processing technology encompasses the acquisition, analysis, and processing of image information, with core aspects including image acquisition, image enhancement, image segmentation, image recognition, and image compression. This field achieves the analysis, extraction, and transformation of image content through the digital conversion and processing of information contained in images, and is widely applied in various technological scenarios such as machine vision, medical imaging, remote sensing monitoring, industrial inspection, and autonomous driving. The implementation of image processing typically relies on specific algorithmic processes and data structures, such as filtering algorithms for image denoising, edge detection methods for image contour extraction, and feature extraction techniques for image content analysis. Traditional image data processing methods refer to processing image information from robot-acquired image data through image preprocessing, feature extraction, and analysis to meet the robot's application needs in tasks such as environmental perception, target recognition, or path planning. Image preprocessing typically employs techniques such as grayscale transformation, edge extraction, and geometric transformation, combined with algorithms such as template matching, threshold-based recognition methods, or principal component analysis to complete image feature recognition and judgment, thereby assisting the robot in achieving vision-based environmental understanding and operational execution.

[0003] Existing technologies generally adopt a unified processing strategy based on the overall image when processing robot image data, lacking a differentiated response mechanism for regional features in the image. High-value local information in the image is often submerged by low-density background interference, and edge structures and grayscale changes fail to form an effective combination relationship. During the processing of image content, there are problems such as information mixing and unclear processing priorities, resulting in insufficient weight of key content in the image analysis results. Important segments in the image path are easily weakened or omitted, which in turn leads to technical bottlenecks such as increased data redundancy, low path organization efficiency, and non-targeted image scheduling, affecting the execution coherence and path optimization effect in the robot image data processing process. Summary of the Invention

[0004] To achieve the above objectives, the present invention adopts the following technical solution: a robot image data processing method, comprising the following steps: S1: By acquiring continuous images of the robot's working area, grayscale extraction and contour scanning are performed on each frame of the image to identify the range of regions with continuous grayscale changes and edge distribution, and then the regions are reconstructed to build structural image fragments. S2: Analyze the spatial adjacency relationship between adjacent pixels through the structured image fragment, identify the continuous edge distribution area, divide it into connected pixel fragments, and reconstruct the connected region fragments to generate an image partition fragment set; S3: Calculate the distribution density of edge clusters within the image segments using the image segment set, filter image segments that meet the main processing requirements based on the density and preset conditions, remove sparse edge segments, and construct a priority image segment set; S4: Using the priority image fragment set, perform path matching judgment on the image fragments, and write the image fragments into the main image path, standard path and bypass cache path according to the position of the fragments in the processing order to generate a path image fragment sequence; S5: Using the path image segment sequence, the order of image segments from the main path, standard path and bypass cache path is adjusted, and image segments are filled in according to the path occupancy to generate the robot's image data processing sequence.

[0005] As a further embodiment of the present invention, the structural image segment includes a grayscale continuous region structure, an edge-concentrated distribution structure, and a pixel recombination mapping structure; the image partition segment set includes edge-connected segments, spatially independent segments, and continuous edge segment regions; the priority image segment set includes high-density edge segments, main processing condition segments, and filtered and retained segments; the path image segment sequence includes main image path segments, standard image path segments, and bypass cache path segments; and the robot's image data processing sequence includes image segment interpolation sequences, image segment rearrangement sequences, and path differentiation image sequences.

[0006] As a further aspect of the present invention, the step of selecting image segments that meet the main processing requirements based on density and preset conditions refers to calculating the distribution density of edge clusters within each segment in the image partition segment set, and selecting high-density image segments that meet the main path image processing requirements based on a preset density threshold.

[0007] As a further aspect of the present invention, the identification of continuously distributed edge regions refers to extracting pixel connected regions with continuous edge gray levels based on the spatial adjacency relationship of adjacent pixels in a structural image segment.

[0008] As a further aspect of the present invention, the specific steps of S1 are as follows: S101: Obtain continuous images of the robot's working area, perform grayscale value extraction on each frame, call the RGB values ​​of the pixels and perform weighted superposition according to the weighted grayscale conversion coefficient, convert the color image to a grayscale image, and obtain the image grayscale matrix set; S102: Based on the image grayscale matrix set, perform absolute value comparison on the grayscale difference between adjacent pixels in the image, mark pixels that are greater than the grayscale change threshold according to the preset grayscale change threshold, and aggregate the edge positions of all marked pixels into region edges to generate an edge distribution coordinate set. S103: Based on the edge distribution coordinate set, retrieve the target region with continuous gray level and concentrated edges, call the boundary coordinates to construct the position index matrix, perform structured recombination on the pixels inside the region according to the index position, and establish a structured image fragment.

[0009] As a further aspect of the present invention, the specific steps of S2 are as follows: S201: Based on the structured image fragment, monitor the spatial adjacency relationship between all pixels in the image, determine the adjacency relationship based on the pixel coordinate position, construct interconnected paths for related pixels, obtain pixel groups with spatial connectivity conditions, and obtain a pixel connectivity set. S202: Based on the pixel connectivity set, detect the edge coordinate distribution of each connectivity group and calculate the coordinate spacing between adjacent edge pixels. If the edge meets the preset connectivity requirements, it is classified as the same region, and all connected pixels are merged to generate a connectivity segment group set. S203: Based on the connected segment grouping set, sequentially traverse the image coordinate range corresponding to each group of segments, and perform spatial partitioning on the pixel matrix of the corresponding region in the structured image segment to construct independent region data frames. Then, aggregate all independent regions in sequence to establish an image partition segment set.

[0010] As a further aspect of the present invention, the specific steps of S3 are as follows: S301: Based on the image partition segment set, extract the coordinate positions of edge pixels in each image segment, count the number of edge pixels per unit area, and calculate the edge aggregation degree in combination with the total number of pixels in the region to obtain the unit area index of edge distribution and obtain the edge density parameter set. S302: Call the edge density parameter set, compare the edge density value of the image segment with the preset main processing density threshold, filter the segments whose edge density is greater than the main processing threshold and mark them with an index, and put all marked segments into the same set to generate a target image segment list; S303: Based on the target image fragment list, the image data frames corresponding to all remaining high-density fragments are aggregated in order to construct a priority image fragment set.

[0011] As a further aspect of the present invention, the specific steps of S4 are as follows: S401: Based on the set of priority image segments, extract the pixel matrix of the image segments, monitor the edge continuity between adjacent image segments, determine the connection relationship between the two segments based on the spatial contact of adjacent edge pixels and the consistency of edge gradient direction, and obtain the path matching relationship set. S402: Call the path matching relationship set, mark the image segment index with continuous path characteristics according to the order of each image segment in the connection relationship, and write all image segments that meet the continuous connection conditions into the main image path sequence in index order to establish the main path image segment sequence; S403: Based on the main path image segment sequence, retrieve the index of the image segment that is not included, and combine it with the spatial position of the segment in the original set. Write the image segment that meets the preset auxiliary conditions into the standard path, and write the rest into the cache path. Then, aggregate the paths into a unified image output unit to generate a path image segment sequence.

[0012] As a further aspect of the present invention, the specific steps of S5 are as follows: S501: Based on the path image segment sequence, classify the image segments from the main image path, standard image path and bypass cache path, extract the current arrangement order of the image segments in each type of image path, record the position information according to the index order, and obtain the path image sequence set; S502: Call the path image sequence set, monitor the continuity of frame index in the main image path and the bypass cache path. If there is a frame gap, the frame data can be completed by copying image segments with similar order and obtaining transition image segments by linear interpolation, writing them into the gap position, and generating a list of supplementary image segments according to the main path index order. S503: Based on the list of padded image segments, rearrange all image segments in the padded order and construct a continuous frame data stream. After frame order encoding of each image, write it into the image data structure in sequence to establish the robot's image data processing sequence.

[0013] As a further aspect of the present invention, the frame gap refers to the position in the path image segment sequence where the frame index is discontinuous and the image segment is missing when detected in the order of the main path index.

[0014] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, by constructing structural regions with grayscale continuity and edge concentration, the spatial correlation and edge aggregation features of pixels within the image are strengthened, improving the accuracy of image segment division and the effectiveness of density screening. By combining the segment order relationship, path splitting and priority writing are realized, optimizing the organization structure and flow efficiency of image content in the path. By supplementing and rearranging differentiated paths, the continuity and scheduling flexibility of image sequences are improved, enhancing the extraction efficiency of key image content and the response stability of processing paths. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a schematic diagram of the steps of the present invention; Figure 2 This is a detailed schematic diagram of S1 of the present invention; Figure 3 This is a detailed schematic diagram of S2 of the present invention; Figure 4 This is a detailed schematic diagram of S3 of the present invention; Figure 5 This is a detailed schematic diagram of S4 of the present invention; Figure 6 This is a detailed schematic diagram of S5 of the present invention. Detailed Implementation

[0017] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0018] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0019] Please see Figure 1 This invention provides a method for processing image data of a robot, comprising the following steps: S1: By continuously acquiring continuous images of the robot's working area, performing grayscale extraction and contour scanning frame by frame on the continuous images, identifying the range of regions with continuous grayscale changes and edge distribution within the image, and reorganizing the pixels within the image according to the region boundaries, constructing a structured image fragment from the reorganized image content. S2: Through structural image fragments, perform connectivity analysis on the spatial adjacency relationship between adjacent pixels within the structural image fragments, divide connected pixels into independent image fragments according to the continuous distribution of edges, and perform fragment division operation on all connected regions in the structural image fragments in sequence to generate a set of image partition fragments. S3: Through the image partition fragment set, perform density calculation on the edge cluster distribution within each image fragment, and perform hierarchical screening of image fragments based on the density distribution results. Image fragments that meet the preset main processing conditions are included in the priority processing range, while image fragments with sparse edge distribution are removed, and a priority image fragment set is constructed. S4: Using the priority image fragment set, perform path matching judgment on the image fragments in the priority image fragment set, and write the corresponding image fragments into the main image path according to the positional relationship of the image fragments in the processing order. Write the image fragments that are not selected into the main image path into the standard image path and the bypass cache path respectively, and generate a path image fragment sequence. S5: Through the path image fragment sequence, perform order adjustment on the image fragments from the differentiated image paths in the path image fragment sequence, and perform padding and sequential rearrangement of the image fragments according to the current occupancy of the main image path and the bypass cache path to generate the robot's image data processing sequence.

[0020] The structural image segments include grayscale continuous region structures, edge-concentrated distribution structures, and pixel recombination mapping structures. The image partition segment set includes edge-connected segments, spatially independent segments, and continuous edge segment regions. The priority image segment set includes high-density edge segments, main processing condition segments, and filtered and retained segments. The path image segment sequence includes main image path segments, standard image path segments, and bypass buffer path segments. The robot's image data processing sequence includes image segment completion sequences, image segment rearrangement sequences, and path differentiation image sequences.

[0021] Please see Figure 2 The specific steps of S1 are as follows: S101: Obtain continuous images of the robot's working area, perform grayscale value extraction on each frame, call the RGB values ​​of the pixels and perform weighted superposition according to the weighted grayscale conversion coefficient, convert the color image to a grayscale image, and obtain the image grayscale matrix set; Using a high-resolution vision sensor mounted on the robot's front end, such as an industrial-grade CMOS camera with 4K resolution and a sampling rate of 60 frames per second, optical signals from the target work area are continuously acquired during operation. These optical signals are then converted into digital image signals, forming a sequence of original color image segments. Data analysis is performed on each frame of the sequence, extracting the brightness values ​​of each pixel in the red, green, and blue channels. Weighted grayscale conversion coefficients, pre-stored in a configuration file, are retrieved, including weights for the red, green, and blue components. Based on the differences in human eye sensitivity to different colors of light, the red component weight is set to 0.299, the green component weight to 0.587, and the blue component weight to 0.114. For each pixel, the extracted red channel value is multiplied by the red component weight to obtain the red light contribution value; the green channel value is multiplied by the green component weight to obtain the green light contribution value; and the blue channel value is multiplied by the blue component weight to obtain the blue light contribution value. Then, the red, green, and blue light contribution values ​​are summed, and the sum is the grayscale value of that pixel. For example, if a pixel's RGB values ​​are 200 for red, 100 for green, and 50 for blue, the grayscale value in this process is: 200 × 0.299 + 100 × 0.587 + 50 × 0.114 = 59.8 + 58.7 + 5.7 = 124.2. This calculation process is repeated for all pixels in the image, replacing the original RGB data of each pixel with the calculated grayscale value, thus constructing a single-channel grayscale image. All pixel data of this grayscale image are then stored in matrix form, creating an image grayscale matrix set.

[0022] S102: Based on the image grayscale matrix set, perform absolute value comparison on the grayscale difference between adjacent pixels in the image, mark pixels that are greater than the grayscale change threshold according to the preset grayscale change threshold, and aggregate the edge positions of all marked pixels into region edges to generate an edge distribution coordinate set. Following the row and column order of the matrix, each pixel within the image is traversed. For the currently traversed center pixel, its eight directly adjacent neighboring pixels in the horizontal, vertical, and diagonal directions are determined. The grayscale value of the center pixel and the grayscale value of each neighboring pixel are obtained, and a difference operation is performed: subtract the grayscale value of the neighboring pixels from the grayscale value of the center pixel, and then the absolute value of the difference is taken to obtain the absolute value of the grayscale difference. A preset grayscale abrupt change threshold is retrieved, which is used to determine whether there are significant edge features in the image. If one or more of the calculated absolute values ​​of grayscale differences are greater than the grayscale abrupt change threshold, the current center pixel is determined to be in a grayscale abrupt change region and is marked as an edge pixel; otherwise, if the absolute values ​​of grayscale differences in all directions are less than or equal to the grayscale abrupt change threshold, it is determined to be a pixel in a flat region and is not marked. For example, the grayscale abrupt change threshold is set to 30. If the grayscale value of the center pixel is 100, and the grayscale value of its right-hand neighboring pixel is 140. The process calculates the difference between the two values: 100 minus 140 equals -40, and taking the absolute value yields 40. Since 40 is greater than the preset threshold of 30, the center pixel is identified as an edge pixel and assigned a label state of 1, while non-edge pixels are assigned a label state of 0. After completing the full image scan, the row and column coordinates of all pixels with a label state of 1 are extracted, and these coordinate data are aggregated to generate an edge distribution coordinate set.

[0023] S103: Based on the edge distribution coordinate set, retrieve the target region with continuous gray level and concentrated edges, call the boundary coordinates to construct the position index matrix, perform structured recombination on the pixels inside the region according to the index position, and establish a structured image fragment. Based on the coordinate data in the edge distribution coordinate set, the region of interest (ROI) is delineated in the grayscale image matrix. Specifically, this involves retrieving regions with dense edge pixel distribution, identifying closed or semi-closed geometric shapes enclosed by continuous edge pixels, and defining the pixel range covered by these shapes as the target region. The boundary pixel coordinates of this target region are extracted, and the bounding box of the region is determined based on the maximum and minimum values ​​of the boundary coordinates. A position index matrix is ​​constructed based on the row and column range of this bounding box. This index matrix records the relative position information of each pixel within the target region relative to the region's origin. Based on this position index matrix, the corresponding pixel data is extracted from the original grayscale matrix and rearranged and combined according to the spatial order indicated by the index. Interference data from background noise regions is removed, and the discrete target region pixels are reassembled into compact data blocks. For example, if the edge set indicates a target object located at the center of the image, its boundary coordinate range is from row 200 to row 500 and column 300 to column 600. The process constructs a 301-row by 301-column index matrix, and fills the corresponding pixel values ​​in the original image into the new matrix one by one, forming a structured data block with the external background removed, and finally establishing a structured image fragment.

[0024] Please see Figure 3 The specific steps of S2 are as follows: S201: Based on the structural image fragment, monitor the spatial adjacency relationship between all pixels in the image, determine the adjacency relationship based on the pixel coordinate position, construct interconnected paths for related pixels, obtain pixel groups with spatial connectivity conditions, and obtain the pixel connectivity set. Spatial adjacency monitoring is performed on every non-zero pixel in the structural image segment. This process employs an eight-neighborhood connectivity algorithm, which checks for any target pixel in its eight directions (up, down, left, right, upper left, upper right, lower left, and lower right) to identify pixels with the same non-zero value. If a non-zero pixel exists in any of these directions, a physical connection is established between the two pixels, creating a connection path. Using a disjoint-set data structure or depth-first search strategy, the algorithm extends along the established connection path, grouping all pixels reachable through this path into the same group. Each independent group represents an independent connected component in the image. The entire structural image segment is traversed until all non-zero pixels are assigned to specific groups, resulting in multiple pixel groups with spatial connectivity. The coordinates of all pixels in each group are recorded to obtain the pixel connectivity set. For example, a cluster of pixels exists in the upper left corner of the image, where pixel A is adjacent to pixel B, and pixel B is adjacent to pixel C. The process determines that A is connected to B and B is connected to C, and then deduces that A, B, and C belong to the same connected group, and packages and stores their coordinates.

[0025] S202: Based on the pixel connected set, detect the edge coordinate distribution of each connected group and calculate the coordinate spacing between adjacent edge pixels. If the edge meets the preset connectivity requirements, it is classified as the same region, and all connected pixels are merged to generate a connected segment group set. For each connected group in the pixel connected set, extract the coordinates of its outermost edge pixels. Calculate the spatial distance between edge pixels of different connected groups. Specifically, select all edge coordinates of the first connected group and all edge coordinates of the second connected group, select representative points (such as centroids or boundary midpoints) in each group's edge coordinates, and calculate the Euclidean distance between the two representative points as an approximation of the proximity distance between the two groups. Retrieve the preset connectivity requirement parameter, which is usually set as the maximum pixel distance threshold allowed for merging. If the calculated proximity distance is less than or equal to the connectivity requirement parameter, it is determined that the two connected groups spatially belong to the same segmented and broken part of a potential object and should be merged. Perform a merge operation, merging the pixel sets of the two groups into a new large group and updating its edge coordinates. Repeat the above detection and merging process until the distance between all remaining groups is greater than the connectivity requirement parameter. For example, set the connectivity requirement parameter to 5 pixels. If the nearest distance between the edge of group A and the edge of group B is 3 pixels, which is less than 5 pixels, then group A and group B are merged. Finally, generate a connected segment group set from all merged and organized groups.

[0026] S203: Based on the connected segment grouping set, traverse the image coordinate range corresponding to each group of segments in turn, and perform spatial partitioning on the pixel matrix of the corresponding region in the structured image segment to construct independent region data frames. Then, aggregate all independent regions in sequence to establish an image partition segment set. The image coordinate range covered by each group in the connected segment grouping set is read sequentially to determine the coordinates of the top-left and bottom-right corners of its smallest bounding rectangle. Based on these coordinate ranges, the corresponding local pixel matrix is ​​cut out from the original structural image segment. For each cut-out local matrix, independent storage space is dynamically allocated according to its size to construct an independent region data frame. This data frame contains not only pixel grayscale data but also the positional offset information of the region in the original image. All constructed independent region data frames are numbered and arranged according to their spatial distribution order from top to bottom and from left to right in the original image to establish an image partition segment set. For example, if three independent objects are identified in the image, located at the top left, center, and bottom right respectively, this process will cut out these three regions to form three independent data blocks, named segment 1, segment 2, and segment 3 respectively, and store them in the set for subsequent processing.

[0027] Please see Figure 4 The specific steps of S3 are as follows: S301: Based on the image segment set, extract the coordinate positions of edge pixels in each image segment, count the number of edge pixels per unit area, and calculate the degree of edge aggregation by combining the total number of pixels in the region, so as to obtain the unit area index of edge distribution and obtain the edge density parameter set. The process iterates through each image segment in the image segment set, scanning the pixel matrix within that segment again using the edge detection operator to extract the pixel coordinates of all identified edge points. The total number of edge pixels within the segment is counted, along with the total number of pixels in the segment (i.e., the segment's height multiplied by its width). A division operation is performed, dividing the total number of edge pixels by the total number of pixels to obtain the edge distribution density value of the segment. This density value reflects the complexity and texture richness of the image content. The calculated edge density value is associated with the segment's identifier ID and stored. For example, an image segment may be 100 pixels by 100 pixels, with a total of 10,000 pixels. The segment is found to contain 2,500 edge pixels. The unit area index for edge distribution is calculated as 2,500 divided by 10,000, resulting in 0.25. The density values ​​calculated for all segments are then aggregated to obtain the edge density parameter set.

[0028] Table 1. Data for calculating edge density of image segments

[0029] As shown in Table 1, the specific data for performing the above calculations on four different image fragments are listed. Frag_002 and Frag_004 show higher edge density, indicating that their internal textures are more complex.

[0030] S302: Call the edge density parameter set, compare the edge density value of the image segment with the preset main processing density threshold, filter the segments whose edge density is greater than the main processing threshold and mark the index, put all marked segments into the same set, and generate a list of target image segments; The edge density parameter set is retrieved, and a main processing density threshold is set. This threshold is an empirical value derived from historical data statistical analysis, used to distinguish high-value information areas from flat background areas. The edge distribution density value of each image segment is numerically compared with the main processing density threshold. If the density value of a segment is greater than the main processing density threshold, the segment is determined to contain key operational features or obstacle information, requiring priority processing, and its index number is recorded. If the density value is less than or equal to the threshold, it is temporarily marked as a secondary area. All segments identified as high-density are indexed into the same list. For example, the main processing density threshold is set to 0.20. According to the data in Table 1, the density of Frag_001 is 0.05, which is less than 0.20; the density of Frag_002 is 0.35, which is greater than 0.20. Therefore, this process marks Frag_002 and adds it to the list, while ignoring Frag_001. Finally, a list of target image segments containing the indices of Frag_002 and Frag_004 is generated.

[0031] S303: Based on the target image fragment list, and by sequentially aggregating the image data frames corresponding to all remaining high-density fragments, construct a priority image fragment set; Based on the target image fragment list, the system filters within the image partition fragment set. Density image fragments not listed are directly removed, freeing up their temporary storage space. The retained high-density fragments are sorted according to their original spatial scanning order or a preset priority weight. The image data frames corresponding to each sorted fragment are extracted and linked sequentially to form a compact queue of data to be processed. This queue discards redundant background information, containing only high-frequency texture regions that the robot needs to focus on. For example, the system deletes the data from Frag_001 and Frag_003, retaining only Frag_002 and Frag_004, and placing Frag_002 before Frag_004 to construct a priority image fragment set.

[0032] Please see Figure 5 The specific steps of S4 are as follows: S401: Based on the priority image segment set, extract the pixel matrix of the image segments, monitor the edge continuity between adjacent image segments, determine the connection relationship between the two segments based on the spatial contact of adjacent edge pixels and the consistency of edge gradient direction, and obtain the path matching relationship set. The pixel matrix of each image segment is extracted from the priority image segment set. For two spatially adjacent image segments, the edge pixel distribution at their boundary is monitored. The edge pixel set of the first segment on the boundary and the edge pixel set of the second segment on the corresponding boundary are obtained. The number of pixels with overlapping or adjacent spatial coordinates in the two sets are calculated as the number of contact points. At the same time, the direction vector of the edge gradient at the boundary is calculated, and the consistency of the gradient edge gradient direction of the two segments at the boundary is compared. If the number of contact points is greater than a preset contact number threshold, and the angle between the gradient directions is less than a preset angle threshold, then the two segments are determined to have continuity in texture and belong to different parts of the same object, and a connection relationship is established. For example, the right edge of segment A is adjacent to the left edge of segment B, the number of contact points is 15, which is greater than the threshold of 10; and the edge gradient directions are both horizontal to the right with an angle of 0 degrees. This process determines that A and B have a path matching relationship and obtains a path matching relationship set.

[0033] S402: Call the path matching relationship set, mark the image segment index with continuous path characteristics according to the order of each image segment in the connection relationship, and write all image segments that meet the continuous connection conditions into the main image path sequence in index order to establish the main path image segment sequence. The process involves calling a path matching relationship set to analyze the connection topology between segments. Starting with the initial segment as the root node, subsequent segments are sequentially located based on their connections, constructing one or more continuous image processing paths. Image segments within the same continuous path are numbered and labeled according to their connection order. The indices of all image segments marked as belonging to the same continuous path are written into the main image path sequence, strictly following their physical connection order. This sequence represents the complete visual extension path of the main objects in the image. For example, if the relationship set shows segment A connecting to segment B, and segment B connecting to segment C, this process will record the indices in the order of A, B, and C, establishing the main path image segment sequence.

[0034] S403: Based on the main path image segment sequence, retrieve the index of the image segment that is not included, and combine it with the spatial position of the segment in the original set. Write the image segments that meet the preset auxiliary conditions into the standard path, and write the rest into the cache path. Then, aggregate the paths into a unified image output unit to generate the path image segment sequence. The process iterates through all indices in the priority image fragment set, comparing them with indices already included in the main path image fragment sequence to filter out all remaining image fragments not included in the main path. For each remaining fragment, it checks whether its spatial location meets preset auxiliary conditions, such as whether it is located within a specific range around the main path or whether it has specific geometric features. If it meets the auxiliary conditions, it is written into the standard path sequence as a secondary processing object; if it does not meet the auxiliary conditions, it is written into the cached path sequence as backup data. The data units of the main path, standard path, and cached path are encapsulated and aggregated, with unified header information format, to generate the path image fragment sequence. For example, fragment D is not in the main path, but its position is close to the center of the main path, meeting the auxiliary conditions, and is written into the standard path; fragment E is located far away and is written into the cached path.

[0035] Please see Figure 6 The specific steps of S5 are as follows: S501: Based on the path image fragment sequence, classify the image fragments from the main image path, standard image path and bypass cache path, extract the current arrangement order of the image fragments in each type of image path, record the position information according to the index order, and obtain the path image sequence set; The path image segment sequence is parsed, and the path category label of each image segment is identified, classifying them into main image path frames, standard image path frames, and bypass buffer path frames. For each path category, the current arrangement index value of all image segments contained therein is read. A position information record table is created, and the original index, category label, and relative order in the current path of each frame are filled in sequentially. For example, the main path contains frames with indices 1, 2, and 4, and the standard path contains a frame with index 3. This process records the main path sequence as [1, 2, 4] and the standard path sequence as [3], thereby obtaining the path image sequence set.

[0036] S502: Call the path image sequence set, monitor the continuity of frame index in the main image path and the bypass buffer path. If there is a frame gap, it can be filled by copying image segments with similar order and using linear interpolation to obtain transition image segments, writing them into the gap position, filling the frame data according to the main path index order, and generating a list of supplementary image segments. The process calls the path image sequence set, focusing on checking the continuity of frame indices within the main image path and the bypass buffer path. It calculates the index difference between adjacent frames; if the difference equals 1, they are considered continuous; otherwise, a frame gap is identified. When a gap is detected, for example, index 2 is immediately followed by index 4, leaving index 3 missing, the process immediately searches for the image segment with index 3 in the standard path or the full buffer. If the frame is found, it is extracted and inserted between indices 2 and 4; if no precise index is found, image segments with similar spatial positions in the standard path or buffer path are selected as candidates for filling, prioritizing those in the original spatial region adjacent to the missing position to ensure the temporal and logical integrity of the data flow. This completion operation generates a list of seamless filler image segments.

[0037] S503: Based on the list of padded image segments, rearrange all image segments in the padded order and construct a continuous frame data stream. After frame order encoding of each frame image, write it into the image data structure in sequence to establish the robot's image data processing sequence. All image fragments are rearranged according to their final, completed order, forming a strictly monotonically increasing frame sequence. A continuous frame data stream is constructed, and frame order encoding is performed on each frame. This encoding process includes assigning a unique 64-bit timestamp and a 32-bit sequence number to each frame, and compressing the image pixel data into a standard transmission format. The encoded data packets are sequentially written into the robot's high-speed shared memory or video memory data structure, forming a linear stream of data that can be directly read by downstream navigation or recognition algorithms, thus establishing the robot's image data processing sequence.

[0038] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of protection of the technical solution.

Claims

1. A method for processing image data of a robot, characterized in that, Includes the following steps: S1: By acquiring continuous images of the robot's working area, grayscale extraction and contour scanning are performed on each frame of the image to identify the range of regions with continuous grayscale changes and edge distribution, and then the regions are reconstructed to build structural image fragments. S2: Analyze the spatial adjacency relationship between adjacent pixels through the structured image fragment, identify the continuous edge distribution area, divide it into connected pixel fragments, and reconstruct the connected region fragments to generate an image partition fragment set; S3: Calculate the distribution density of edge clusters within the image segments using the image segment set, filter image segments that meet the main processing requirements based on the density and preset conditions, remove sparse edge segments, and construct a priority image segment set; S4: Using the priority image fragment set, perform path matching judgment on the image fragments, and write the image fragments into the main image path, standard path and bypass cache path according to the position of the fragments in the processing order to generate a path image fragment sequence; S5: Using the path image segment sequence, the order of image segments from the main path, standard path and bypass cache path is adjusted, and image segments are filled in according to the path occupancy to generate the robot's image data processing sequence.

2. The image data processing method for a robot according to claim 1, characterized in that, The structural image fragments include grayscale continuous region structures, edge-concentrated distribution structures, and pixel recombination mapping structures. The image partition fragment set includes edge-connected fragments, spatially independent fragments, and continuous edge fragment regions. The priority image fragment set includes high-density edge fragments, main processing condition fragments, and filtered and retained fragments. The path image fragment sequence includes main image path fragments, standard image path fragments, and bypass buffer path fragments. The robot's image data processing sequence includes image fragment interpolation sequences, image fragment rearrangement sequences, and path differentiation image sequences.

3. The image data processing method for a robot according to claim 1, characterized in that, The process of selecting image segments that meet the main processing requirements based on density and preset conditions refers to calculating the distribution density of edge clusters within each segment in the image partition segment set, and selecting high-density image segments that meet the main path image processing requirements based on preset density thresholds.

4. The image data processing method for a robot according to claim 1, characterized in that, The identified edge-continuous distribution region refers to the pixel connected region with edge gray-level continuity extracted based on the spatial adjacency relationship of adjacent pixels in the structured image segment.

5. The image data processing method for a robot according to claim 1, characterized in that, The specific steps of S1 are as follows: S101: Obtain continuous images of the robot's working area, perform grayscale value extraction on each frame, call the RGB values ​​of the pixels and perform weighted superposition according to the weighted grayscale conversion coefficient, convert the color image to a grayscale image, and obtain the image grayscale matrix set; S102: Based on the image grayscale matrix set, perform absolute value comparison on the grayscale difference between adjacent pixels in the image, mark pixels that are greater than the grayscale change threshold according to the preset grayscale change threshold, and aggregate the edge positions of all marked pixels into region edges to generate an edge distribution coordinate set. S103: Based on the edge distribution coordinate set, retrieve the target region with continuous gray level and concentrated edges, call the boundary coordinates to construct the position index matrix, perform structured recombination on the pixels inside the region according to the index position, and establish a structured image fragment.

6. The image data processing method for a robot according to claim 1, characterized in that, The specific steps of S2 are as follows: S201: Based on the structured image fragment, monitor the spatial adjacency relationship between all pixels in the image, determine the adjacency relationship based on the pixel coordinate position, construct interconnected paths for related pixels, obtain pixel groups with spatial connectivity conditions, and obtain a pixel connectivity set. S202: Based on the pixel connectivity set, detect the edge coordinate distribution of each connectivity group and calculate the coordinate spacing between adjacent edge pixels. If the edge meets the preset connectivity requirements, it is classified as the same region, and all connected pixels are merged to generate a connectivity segment group set. S203: Based on the connected segment grouping set, sequentially traverse the image coordinate range corresponding to each group of segments, and perform spatial partitioning on the pixel matrix of the corresponding region in the structured image segment to construct independent region data frames. Then, aggregate all independent regions in sequence to establish an image partition segment set.

7. The image data processing method for a robot according to claim 1, characterized in that, The specific steps for S3 are as follows: S301: Based on the image partition segment set, extract the coordinate positions of edge pixels in each image segment, count the number of edge pixels per unit area, and calculate the edge aggregation degree in combination with the total number of pixels in the region to obtain the unit area index of edge distribution and obtain the edge density parameter set. S302: Call the edge density parameter set, compare the edge density value of the image segment with the preset main processing density threshold, filter the segments whose edge density is greater than the main processing threshold and mark them with an index, and put all marked segments into the same set to generate a target image segment list; S303: Based on the target image fragment list, the image data frames corresponding to all remaining high-density fragments are aggregated in order to construct a priority image fragment set.

8. The image data processing method for a robot according to claim 1, characterized in that, The specific steps of S4 are as follows: S401: Based on the set of priority image segments, extract the pixel matrix of the image segments, monitor the edge continuity between adjacent image segments, determine the connection relationship between the two segments based on the spatial contact of adjacent edge pixels and the consistency of edge gradient direction, and obtain the path matching relationship set. S402: Call the path matching relationship set, mark the image segment index with continuous path characteristics according to the order of each image segment in the connection relationship, and write all image segments that meet the continuous connection conditions into the main image path sequence in index order to establish the main path image segment sequence; S403: Based on the main path image segment sequence, retrieve the index of the image segment that is not included, and combine it with the spatial position of the segment in the original set. Write the image segment that meets the preset auxiliary conditions into the standard path, and write the rest into the cache path. Then, aggregate the paths into a unified image output unit to generate a path image segment sequence.

9. The image data processing method for a robot according to claim 1, characterized in that, The specific steps of S5 are as follows: S501: Based on the path image segment sequence, classify the image segments from the main image path, standard image path and bypass cache path, extract the current arrangement order of the image segments in each type of image path, record the position information according to the index order, and obtain the path image sequence set; S502: Call the path image sequence set, monitor the continuity of frame index in the main image path and the bypass cache path. If there is a frame gap, the frame data can be completed by copying image segments with similar order and obtaining transition image segments by linear interpolation, writing them into the gap position, and generating a list of supplementary image segments according to the main path index order. S503: Based on the list of padded image segments, rearrange all image segments in the padded order and construct a continuous frame data stream. After frame order encoding of each image, write it into the image data structure in sequence to establish the robot's image data processing sequence.

10. The image data processing method for a robot according to claim 9, characterized in that, The frame gap refers to the position in the path image segment sequence where the frame index is discontinuous and the image segment is missing when detected in the order of the main path index.

Images