A terrain recognition method and system for underwater dredging robots

Abstract

The present invention belongs to the field of image processing technology, and discloses a terrain recognition method and system for an underwater dredging robot, including: obtaining a light attenuation coefficient using shooting depth and plankton information, accurately correcting the influence of light on the image, and brightening the image based on the light attenuation coefficient to make the terrain features more prominent. Laplace decomposition is used to obtain a multi-layer terrain image to achieve multi-scale analysis; the accuracy of terrain analysis is improved by screening edge texture areas; defogging weights are determined in combination with the light attenuation coefficient, and defogging and restoration are performed using a dark primary color prior to effectively remove the fog effect and improve image visibility. The complete image is synthesized by inverse Laplace pyramid reconstruction to obtain an underwater terrain image with comprehensively improved quality, thereby obtaining underwater terrain information. The present invention not only improves image quality, but also achieves accurate analysis and recognition of underwater terrain, providing strong technical support for underwater terrain research.

Description

Technical Field

[0001] The present invention belongs to the technical field of image processing and relates to a terrain recognition method and system for an underwater dredging robot. Background Art

[0002] Underwater dredging robots typically use sensors (such as sonar, lasers, and cameras) and robotic arms to perform underwater terrain detection and clearing operations. Their primary task is to remove silt and sediment from the waterbed, particularly in reservoirs, ports, and rivers. When performing dredging, exploration, and construction operations in underwater environments, terrain recognition technology can help robots accurately understand the underwater terrain and obstacles, enabling more precise and efficient operations.

[0003] Underwater dredging robots are usually used in environments such as rivers, lakes, ports, and seabeds. When identifying the terrain during the dredging process based on underwater terrain images, the image clarity is poor due to poor underwater lighting conditions. Current dehazing algorithms based on dark color priors usually restore image clarity through global optimization, but this global optimization method may cause local distortion of the image, affecting the accurate identification of the terrain. Summary of the Invention

[0004] The purpose of the present invention is to solve the problem in the prior art that the defogging algorithm based on dark primary color prior may cause local distortion of the image and affect terrain recognition, and to provide a terrain recognition method and system for underwater dredging robots.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] A terrain recognition method for an underwater dredging robot, comprising:

[0007] Acquire underwater real-time terrain images based on underwater dredging robots and pre-process the acquired terrain images;

[0008] Obtain the light attenuation coefficient based on the shooting depth of the underwater dredging robot and the influence of plankton in the water;

[0009] Brighten the grayscale of different pixels in the pre-processed terrain image based on the light attenuation coefficient;

[0010] Perform Laplace decomposition on the adjusted image to obtain a multi-layer terrain image;

[0011] Based on the terrain image of any layer, obtain the connected domain corresponding to the terrain image and filter out the existing edge texture area;

[0012] Based on the proportion of edge texture areas in each layer of image, the importance weight of high-frequency information in each scale image is obtained, and the final dehazing weight is determined in combination with the light attenuation coefficient.

[0013] Based on the final dehazing weight, the image is dehazed and restored using the dark color prior, and the image information at multiple scales is synthesized into a complete image through inverse Laplacian pyramid reconstruction to obtain the underwater terrain information.

[0014] A further improvement of the present invention is:

[0015] Furthermore, the acquired terrain image is preprocessed, specifically, the acquired terrain image is gray-scale processed to obtain a gray-scale image.

[0016] Furthermore, based on the shooting depth of the underwater dredging robot and the influence of plankton in the water, the light attenuation coefficient is obtained, specifically:

[0017] Generally, the light attenuation in underwater areas is affected by the shooting depth and the plankton in the water. The deeper the shooting depth and the greater the plankton density, the greater the corresponding light attenuation. At the same time, the color wavelength of the water also affects the water's light absorption rate. The shorter the wavelength, the weaker the absorption capacity and the slower the light attenuation. In this case, the following situations exist:

[0018]

[0019] in, A terrain image The corresponding light attenuation coefficient; For terrain images The depth of the shot, The plankton density at the corresponding depth of this water column is, For this water body’s color wavelength value, This is the longest wavelength value of the water body; since the longer the wavelength, the stronger the corresponding absorption capacity, that is, The smaller it is, the greater the light attenuation of the water body is. Combined with the shooting depth and plankton density , then the larger the value, the larger the corresponding light attenuation coefficient, and it is normalized to the range of [0,1].

[0020] Furthermore, the grayscale of different pixels in the terrain image is brightened based on the light attenuation coefficient, specifically:

[0021] For each pixel of this terrain image, the weaker the light, the darker the corresponding pixel grayscale, that is, the grayscale value tends to 0. Therefore, based on the light attenuation coefficient, different pixel grayscales are brightened and adjusted. The greater the light attenuation, the greater the corresponding adjustment, and vice versa. Then, there exists:

[0022]

[0023] in, For terrain images Medium pixel The adjusted pixel value of Pixel The original pixel value, For terrain images The corresponding light attenuation coefficient.

[0024] Furthermore, based on the terrain image of any layer, the connected domain corresponding to the terrain image is obtained, and the existing edge texture areas are screened out, specifically:

[0025] Generally, the grayscale of the edge texture area in an image is significantly different from that of other areas, often showing high contrast. At the same time, the shape of the edge texture area usually appears in the form of linear or curved extension. For an image at a certain scale, the corresponding connected domain is obtained, and then the possible edge texture areas are screened out, which is expressed as:

[0026]

[0027] in, A connected region in this scale image is the confidence of the edge texture area; Connected domain The average gray value of The average gray value of this image; For this connected domain, the modulus of the largest principal component vector after PCA analysis is: The average modulus of all principal component vectors in this connected domain;

[0028] therefore Represents the connected domain in the image at this scale Grayscale difference, due to the strong contrast of grayscale in edge texture area, The larger the value, the more likely it is an edge texture area; Represents the morphological extension trend of this connected domain. Since the largest principal component vector in PCA Represents the most obvious extension degree, that is, if the extension trend of this connected domain is more obvious, the corresponding Relative to The larger the value is, the more likely it is an edge texture area. Finally, normalize the result to the range of [0, 1] and select The corresponding area is the high-frequency edge texture area in the image.

[0029] Furthermore, the proportion of edge texture area in each layer of image is:

[0030]

[0031] in, For the The edge texture area of ​​the layer image, For the The overall area of ​​the layer image;

[0032] The weight of the importance of high-frequency information in each scale image is obtained as follows:

[0033] Put the proportion of edge texture area in each layer of image into the sequence For images of any scale, there exists:

[0034]

[0035] in, For the Importance weight of high-frequency information in layer-scale images; For the The area ratio of high-frequency information in layer-scale images, For the The area ratio of high-frequency information in layer-scale images, is the number of decomposition layers; For the The relative value of the high-frequency information area of ​​the layer-scale image; the larger the value, the more high-frequency information the corresponding layer image has, and the greater the importance weight; It represents the relationship between the number of layers and high-frequency information. Since the lower the number of layers, the more important the corresponding high-frequency information is, and the corresponding importance weight is greater.

[0036] Furthermore, the final defogging weight is determined in combination with the light attenuation coefficient, specifically:

[0037]

[0038] in, For the Dehazing weights for layer-scale images; For the Importance weight of high-frequency information in layer-scale images, is the illumination attenuation coefficient of the original image; that is, if the illumination attenuation coefficient of this image is larger, the overexposure of the high-frequency information area after adjustment will be more serious, and a larger dehazing weight will be required to reduce this situation. Finally, the result is normalized to the range of [0,1].

[0039] Furthermore, based on the final defogging weight, the image is defogged and restored through the dark color prior, specifically:

[0040] Take the image at each scale as the input image and calculate the dark primary color of each pixel to generate the dark primary color image ;

[0041] By calculating the dark original image The maximum value position in the atmospheric illumination value is estimated and transmittance ;

[0042] By the estimated transmittance and atmospheric lighting And the defogging weights to restore a clear fog-free image:

[0043]

[0044] in, After defogging Layer-scale images, For the original Layer-scale foggy images, For the The defogging weight of the layer-scale image is optimized; that is, the dark color prior method is optimized according to the defogging weight of the different layer-scale images to improve the defogging effect.

[0045] Furthermore, the image at each scale is used as the input image, and the dark primary color of each pixel is calculated to generate a dark primary color image. , specifically:

[0046] Assume the input image is ,in, It is each pixel in the image; since some color channels of natural scene images are very dark, the minimum value of the RGB channels of each pixel is used to represent the "dark primary color";

[0047] For an RGB image, the dark primary is defined as:

[0048]

[0049] in, 、 、 are the pixel values ​​of the image in the red, green, and blue channels respectively;

[0050] By calculating the dark original color image The maximum value position in the atmospheric illumination value is estimated and transmittance , specifically:

[0051] In dark color image In the image, find the pixel position with the maximum value; the pixel position with the maximum value corresponds to the darkest part of the image; by obtaining the pixel value in the original image corresponding to the maximum value position in the dark original image, the atmospheric illumination value is inferred. Usually, the RGB pixel value corresponding to the maximum dark original value is selected as the atmospheric illumination value ;

[0052] Assume that the image The maximum position in Corresponding pixel value in the original image , then the atmospheric illumination value for:

[0053]

[0054] in, is an image The pixel value corresponding to the maximum dark primary color position in ;

[0055] Transmittance Reflects the haze level of each pixel in the image, transmittance It is usually done by estimating the dark primary color of each pixel in the image; the formula for calculating the transmittance is as follows:

[0056]

[0057] in, is a constant whose value range is [0,1].

[0058] A terrain recognition system for an underwater dredging robot, comprising:

[0059] a first acquisition module, wherein the first acquisition module acquires a real-time underwater terrain image based on the underwater dredging robot and preprocesses the acquired terrain image;

[0060] a second acquisition module, which acquires a light attenuation coefficient based on the shooting depth of the underwater dredging robot and the influence of plankton in the water;

[0061] An adjustment module, wherein the adjustment module brightens the grayscale of different pixels in the pre-processed terrain image based on the light attenuation coefficient;

[0062] a decomposition module, wherein the decomposition module performs Laplace decomposition on the adjusted image to obtain a multi-layer terrain image;

[0063] A screening module, which obtains a connected domain corresponding to a terrain image of any layer and screens out existing edge texture areas;

[0064] A determination module, which obtains the weight of the importance of high-frequency information in each scale image based on the proportion of edge texture areas in each layer of image, and determines the final defogging weight in combination with the light attenuation coefficient;

[0065] The reconstruction module performs dehazing and restoration on the image based on the final dehazing weights through a dark color prior, and synthesizes image information of multiple scales into a complete image through inverse Laplacian pyramid reconstruction, thereby obtaining underwater terrain information.

[0066] Compared with the prior art, the present invention has the following beneficial effects:

[0067] The present invention uses the shooting depth and plankton information to obtain the light attenuation coefficient, accurately corrects the impact of light on the image, and brightens the image based on the light attenuation coefficient, making the terrain features more prominent. Laplace decomposition is used to obtain a multi-layer terrain image, realizing multi-scale analysis and a comprehensive understanding of the underwater terrain structure. The accuracy of terrain analysis is improved by screening edge texture areas. The defogging weight is determined in combination with the light attenuation coefficient, and the dark primary color prior is used for defogging and restoration, effectively removing the fog effect and improving image visibility. Finally, the complete image is synthesized through inverse Laplace pyramid reconstruction to achieve efficient information fusion, obtain an underwater terrain image with comprehensively improved quality, and then obtain underwater terrain information. The present invention not only improves image quality, but also realizes accurate analysis and identification of underwater terrain, providing strong technical support for underwater terrain research. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without paying any creative work.

[0069] Figure 1 A schematic flow chart of a method for improving the clarity of underwater terrain images according to the present invention;

[0070] Figure 2 The figure is a schematic structural diagram of a system for improving the clarity of underwater terrain images according to the present invention. DETAILED DESCRIPTION

[0071] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings herein can be arranged and designed in various different configurations.

[0072] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention as claimed, but rather merely represents selected embodiments of the present invention. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without creative effort shall fall within the scope of protection of the present invention.

[0073] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings.

[0074] In the description of the embodiments of the present invention, it should be noted that if the terms "upper," "lower," "horizontal," "inner," etc. appear, the orientation or positional relationship indicated is based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship in which the inventive product is typically placed when in use. These terms are merely for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or component referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limitations on the present invention. In addition, the terms "first," "second," etc. are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0075] In addition, if the term "horizontal" appears, it does not mean that the component must be absolutely horizontal, but can be slightly tilted. For example, "horizontal" only means that its direction is more horizontal than "vertical", and does not mean that the structure must be completely horizontal, but can be slightly tilted.

[0076] In the description of the embodiments of the present invention, it should be noted that, unless otherwise expressly specified or limited, the terms "disposed," "installed," "connected," and "connected" should be understood in a broad sense. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to mechanical connections or electrical connections; they can refer to direct connections or indirect connections through an intermediate medium; and they can refer to internal connections between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0077] The present invention is described in further detail below with reference to the accompanying drawings:

[0078] See also Figure 1 The present invention discloses a terrain recognition method for an underwater dredging robot, comprising:

[0079] S101, acquiring a real-time underwater terrain image based on the underwater dredging robot, and preprocessing the acquired terrain image;

[0080] An underwater dredging robot is an automated robotic system designed specifically for dredging, salvaging, surveying, and repairing underwater environments. It is widely used for silt removal in reservoirs, rivers, ports, and waterways. During operation, the robot typically uses a high-definition camera to capture underwater images, allowing operators to observe the terrain and cleaning results in real time.

[0081] Because underwater dredging robots operate in turbid water and poor lighting, images are often unclear and heavily fogged, making it difficult to capture detailed information and thus affecting accurate terrain recognition. Current dehazing algorithms, such as dark-color priors, typically restore image clarity through global optimization. However, in underwater images, local features such as terrain texture, object edges, and water flow are crucial for terrain recognition. Global optimization methods may overlook this local texture information, resulting in suboptimal dehazing results.

[0082] After the underwater dredging robot obtains the real-time underwater terrain image captured by the high-definition camera, it pre-processes the captured terrain image and grayscale processes the acquired terrain image to obtain a grayscale image.

[0083] S102, obtaining a light attenuation coefficient based on the shooting depth of the underwater dredging robot and the influence of plankton in the water;

[0084] The dark-primary prior is a priori knowledge based on image color statistics. It was first used in image dehazing research. It can be used to estimate atmospheric light and transmittance in an image, thereby restoring image clarity. However, this global approach is not very effective in restoring underwater terrain details. To improve image quality, this step, based on the concept of multi-scale fusion, decomposes the image into representations of different scales using a Laplacian pyramid, and then processes and fuses the image information at each scale.

[0085] Since the underwater terrain in the dredging area generally has low light intensity, the image often appears blurry, dim, and loses details. These characteristics will cause the image to lose important high-frequency information after Laplacian pyramid decomposition. Therefore, it is necessary to first perform illumination compensation on the image to obtain clearer high-frequency detail processing.

[0086] Generally, the light attenuation in underwater areas is affected by the shooting depth and the plankton in the water. The deeper the shooting depth and the greater the plankton density, the greater the corresponding light attenuation. At the same time, the color wavelength of the water also affects the water's light absorption rate. The shorter the wavelength, the weaker the absorption capacity and the slower the light attenuation. In this case, the following situations exist:

[0087]

[0088] in, A terrain image The corresponding light attenuation coefficient; For terrain images The depth of the shot, The plankton density at the corresponding depth of this water column is, For this water body’s color wavelength value, This is the longest wavelength value of the water body; since the longer the wavelength, the stronger the corresponding absorption capacity, that is, The smaller it is, the greater the light attenuation of the water body is. Combined with the shooting depth and plankton density , then the larger the value, the larger the corresponding light attenuation coefficient, and it is normalized to the range of [0, 1].

[0089] S103, brightening and adjusting the grayscale of different pixels in the pre-processed terrain image based on the light attenuation coefficient;

[0090] For each pixel of this terrain image, the weaker the light, the darker the corresponding pixel grayscale, that is, the grayscale value tends to 0. Therefore, based on the light attenuation coefficient, different pixel grayscales are brightened and adjusted. The greater the light attenuation, the greater the corresponding adjustment, and vice versa. Then, there exists:

[0091]

[0092] in, For terrain images Medium pixel The adjusted pixel value of Pixel The original pixel value, For terrain images The corresponding light attenuation coefficient.

[0093] S104: Perform Laplace decomposition on the adjusted image to obtain a multi-layer terrain image.

[0094] The image resolutions of different layers are different, and the information they contain is different. Generally, the top layer contains more smooth low-frequency information, and as you go down, each layer gradually contains more high-frequency detail information, and finally the original image is restored at the bottom layer.

[0095] To accurately identify underwater terrain, we need to utilize a dark color prior to defogging underwater terrain images. High-frequency information corresponding to texture edges in images can better reflect the terrain's elevation. Therefore, we need to adaptively defog the high-frequency information of images at different scales and then fuse it. This preserves the image's edge texture during the defogging process, improving terrain recognition accuracy.

[0096] Because high-frequency information in decomposed multi-layer terrain images decreases from the bottom layer to the top layer, it represents a gradual decrease in the edge texture of the image. Furthermore, high-frequency information in the lower layers of the image represents more important terrain edge texture, which is more important for improving underwater terrain recognition accuracy. Therefore, it is necessary to determine the high-frequency information areas in images at different scales.

[0097] S105, based on the terrain image of any layer, obtaining the connected domain corresponding to the terrain image, and screening out the existing edge texture areas;

[0098] Generally, the grayscale of the edge texture area in an image is significantly different from that of other areas, often showing high contrast. At the same time, the shape of the edge texture area usually appears in the form of linear or curved extension. For an image at a certain scale, the corresponding connected domain is obtained, and then the possible edge texture areas are screened out, which is expressed as:

[0099]

[0100] in, A connected region in this scale image is the confidence of the edge texture area; Connected domain The average gray value of The average gray value of this image; For this connected domain, the modulus of the largest principal component vector after PCA analysis is: The average modulus of all principal component vectors in this connected domain;

[0101] therefore Represents the connected domain in the image at this scale Grayscale difference, due to the strong contrast of grayscale in edge texture area, The larger the value, the more likely it is an edge texture area; Represents the morphological extension trend of this connected domain. Since the largest principal component vector in PCA Represents the most obvious extension degree, that is, if the extension trend of this connected domain is more obvious, the corresponding Relative to The larger the value is, the more likely it is an edge texture area. Finally, normalize the result to the range of [0, 1] and select The corresponding area is the high-frequency edge texture area in the image.

[0102] After obtaining the high-frequency information corresponding to multi-scale images at different layers, dehazing is optimized using a dark-color prior, targeting these high-frequency regions. Since the dark-color prior estimates transmittance and atmospheric light to restore image clarity and contrast, dehazing is optimized by fusing multiple terrain images of different scales. Since the importance of high-frequency information varies across different scales, it is necessary to determine corresponding dehazing weights for each scale.

[0103] S106, based on the proportion of edge texture areas in each layer of image, obtain the weight of the importance of high-frequency information in each layer of scale image, and determine the final defogging weight in combination with the light attenuation coefficient;

[0104] For images of any scale, the higher the level of high-frequency information retained in the lower layers of the image, the more likely it is to represent the main information of the image. Therefore, a larger dehazing weight should be set for the lower layers of the image. At the same time, the greater the proportion of high-frequency information (edge ​​texture area) in the image, the more high-frequency information the image contains, the more significant dehazing optimization is needed, and the corresponding weight should be larger. The proportion of edge texture area in each layer of the image is specifically:

[0105]

[0106] in, For the The edge texture area of ​​the layer image, For the The overall area of ​​the layer image;

[0107] The weight of the importance of high-frequency information in each scale image is obtained as follows:

[0108] Put the proportion of edge texture area in each layer of image into the sequence For images of any scale, there exists:

[0109]

[0110] in, For the Importance weight of high-frequency information in layer-scale images; For the The area ratio of high-frequency information in layer-scale images, For the The area ratio of high-frequency information in layer-scale images, is the number of decomposition layers; For the The relative value of the high-frequency information area of ​​the layer-scale image; the larger the value, the more high-frequency information the corresponding layer image has, and the greater the importance weight; It represents the relationship between the number of layers and high-frequency information. Since the lower the number of layers, the more important the corresponding high-frequency information is, and the corresponding importance weight is greater.

[0111] The final defogging weight is determined in combination with the light attenuation coefficient, specifically:

[0112]

[0113] in, For the Dehazing weights for layer-scale images; For the Importance weight of high-frequency information in layer-scale images, is the illumination attenuation coefficient of the original image; that is, if the illumination attenuation coefficient of this image is larger, the overexposure of the high-frequency information area after adjustment will be more serious, and a larger dehazing weight will be required to reduce this situation. Finally, the result is normalized to the range of [0,1].

[0114] S107 , based on the final defogging weight, the image is defogged and restored using a dark color prior, and image information at multiple scales is synthesized into a complete image through inverse Laplacian pyramid reconstruction, thereby obtaining underwater terrain information.

[0115] Take the image at each scale as the input image and calculate the dark primary color of each pixel to generate the dark primary color image ;

[0116] Assume the input image is ,in, It is each pixel in the image; since some color channels of natural scene images are very dark, the minimum value of the RGB channels of each pixel is used to represent the "dark primary color";

[0117] For an RGB image, the dark primary is defined as:

[0118]

[0119] in, 、 、 are the pixel values ​​of the image in the red, green, and blue channels respectively;

[0120] By calculating the dark original image The maximum value position in the atmospheric illumination value is estimated and transmittance ;

[0121] In dark color image In the image, find the pixel position with the maximum value; the pixel with the maximum value corresponds to the darkest part of the image; by obtaining the pixel value in the original image corresponding to the maximum value position in the dark original image, the atmospheric illumination value is inferred. Usually, the RGB pixel value corresponding to the maximum dark original value is selected as the atmospheric illumination value. ;

[0122] Assume that the image The maximum position in Corresponding pixel value in the original image , then the atmospheric illumination value for:

[0123]

[0124] in, is an image The pixel value corresponding to the maximum dark primary color position in ;

[0125] Transmittance Reflects the haze level of each pixel in the image, transmittance It is usually done by estimating the dark primary color of each pixel in the image; the formula for calculating the transmittance is as follows:

[0126]

[0127] in, is a constant whose value range is [0,1].

[0128] By the estimated transmittance and atmospheric lighting And the defogging weights to restore a clear fog-free image:

[0129]

[0130] in, After defogging Layer-scale images, For the original Layer-scale foggy images, For the The defogging weight of the layer-scale image is optimized; that is, the dark color prior method is optimized according to the defogging weight of the different layer-scale images to improve the defogging effect.

[0131] After completing dehazing optimization for images at different scales, the images at different levels are fused and reconstructed. Here, the image information at multiple scales is synthesized back into a complete image through inverse Laplacian pyramid reconstruction, ensuring that the global brightness, contrast, and details of the image are well preserved. By obtaining a complete underwater image, underwater terrain information can be obtained.

[0132] By analyzing real-time, defogging underwater terrain images, underwater dredging robots can generate the shortest path or optimal removal strategy. For example, the robot can be set to focus on cleaning areas with heavy silt accumulation, while avoiding and marking areas with numerous obstacles to avoid wasted cleaning time. Furthermore, sensors such as sonar and lidar can be combined to create a three-dimensional model of the underwater environment, helping the robot understand complex terrain and further improving dredging efficiency.

[0133] See also Figure 2 The present invention discloses a terrain recognition system for an underwater dredging robot, comprising:

[0134] a first acquisition module, wherein the first acquisition module acquires a real-time underwater terrain image based on the underwater dredging robot and preprocesses the acquired terrain image;

[0135] a second acquisition module, which acquires a light attenuation coefficient based on the shooting depth of the underwater dredging robot and the influence of plankton in the water;

[0136] An adjustment module, wherein the adjustment module brightens the grayscale of different pixels in the pre-processed terrain image based on the light attenuation coefficient;

[0137] a decomposition module, wherein the decomposition module performs Laplace decomposition on the adjusted image to obtain a multi-layer terrain image;

[0138] A screening module, which obtains a connected domain corresponding to a terrain image of any layer and screens out existing edge texture areas;

[0139] A determination module, which obtains the weight of the importance of high-frequency information in each scale image based on the proportion of edge texture areas in each layer of image, and determines the final defogging weight in combination with the light attenuation coefficient;

[0140] The reconstruction module performs dehazing and restoration on the image based on the final dehazing weights through a dark color prior, and synthesizes image information of multiple scales into a complete image through inverse Laplacian pyramid reconstruction, thereby obtaining underwater terrain information.

[0141] An embodiment of the present invention provides a terminal device. The terminal device of this embodiment includes: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of each of the aforementioned method embodiments are implemented. Alternatively, when the processor executes the computer program, the functions of each module / unit in each of the aforementioned device embodiments are implemented.

[0142] The computer program may be divided into one or more modules / units, which are stored in the memory and executed by the processor to accomplish the present invention.

[0143] The terminal device may be a computing device such as a desktop computer, a notebook computer, a PDA, a cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.

[0144] The processor can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0145] The memory may be used to store the computer programs and / or modules, and the processor implements various functions of the terminal device by running or executing the computer programs and / or modules stored in the memory and calling the data stored in the memory.

[0146] If the module / unit integrated into the terminal device is implemented as a software functional unit and sold or used as a standalone product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention can also implement all or part of the process steps in the above-mentioned method embodiments by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When executed by a processor, the computer program can implement the steps of each of the above-mentioned method embodiments. The computer program includes computer program code, which can be in source code form, object code form, executable file, or some intermediate form. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, software distribution medium, etc.

[0147] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A terrain recognition method for an underwater dredging robot, characterized in that: include: Acquire underwater real-time terrain images based on underwater dredging robots and pre-process the acquired terrain images; Obtain the light attenuation coefficient based on the shooting depth of the underwater dredging robot and the influence of plankton in the water; Brighten the grayscale of different pixels in the pre-processed terrain image based on the light attenuation coefficient; Perform Laplace decomposition on the adjusted terrain image to obtain a multi-layer terrain image; Based on the terrain image of any layer, obtain the connected domain corresponding to the terrain image and filter out the existing edge texture area; Based on the proportion of edge texture areas in each layer of terrain image, the importance weight of high-frequency information in each layer of terrain image is obtained, and the final defogging weight is determined in combination with the light attenuation coefficient; The proportion of edge texture area in each layer of terrain image is: in, For the The edge texture area of ​​the terrain image layer, For the The overall area of ​​the layer topographic image; The weight of the importance of high-frequency information in each layer of terrain image is obtained as follows: Put the proportion of edge texture area in each layer of terrain image into the sequence For terrain images at any scale, there exists: in, For the Importance weight of high frequency information in layer terrain images; For the The area ratio of high-frequency information in the topographic image layer, For the The area ratio of high-frequency information in the topographic image layer, is the number of decomposition layers; For the The relative value of the high-frequency information area of ​​the topographic image of the layer; the larger the value, the more high-frequency information of the corresponding topographic image, and the greater the importance weight; Represents the relationship between the number of layers and high-frequency information. Since the lower the number of layers, the more important the corresponding high-frequency information is, the greater the corresponding importance weight; The final defogging weight is determined in combination with the light attenuation coefficient, specifically: in, For the Dehazing weights of the layer terrain image; For the Importance weight of high frequency information in layer terrain images, is the light attenuation coefficient of the original image; the original image is a real-time underwater terrain image obtained by the underwater dredging robot; Based on the final dehazing weight, the terrain image is dehazed and restored through the dark original color prior, and the terrain image information of multiple scales is synthesized into a complete terrain image through inverse Laplacian pyramid reconstruction to obtain the underwater terrain information.

2. The terrain recognition method for an underwater dredging robot according to claim 1, characterized in that: The preprocessing of the acquired terrain image specifically includes: performing grayscale processing on the acquired terrain image to obtain a grayscale image.

3. The terrain recognition method for an underwater dredging robot according to claim 2, characterized in that: The light attenuation coefficient is obtained based on the shooting depth of the underwater dredging robot and the influence of plankton in the water, specifically: Light attenuation in underwater areas is affected by the depth of the image and the presence of plankton in the water. The deeper the depth and the greater the density of plankton, the greater the corresponding light attenuation. The wavelength of the water also affects its light absorption rate. The shorter the wavelength, the weaker the absorption capacity and the slower the light attenuation. This results in: in, A terrain image The corresponding light attenuation coefficient; For terrain images The depth of the shot, The plankton density at the corresponding depth of this water column is, For this water body’s color wavelength value, This is the longest wavelength value of the water body; since the longer the wavelength, the stronger the corresponding absorption capacity, that is, The smaller it is, the greater the light attenuation of the water body is. Combined with the shooting depth and plankton density , then the larger the value, the larger the corresponding light attenuation coefficient, and it is normalized to the range of [0,1].

4. The terrain recognition method for an underwater dredging robot according to claim 3, characterized in that: The brightening adjustment of different pixel grayscales in the terrain image based on the light attenuation coefficient is specifically as follows: For each pixel in the terrain image, the weaker the light, the darker the corresponding pixel grayscale, that is, the grayscale value tends to 0. Therefore, based on the light attenuation coefficient, different pixel grayscales are brightened and adjusted. The greater the light attenuation, the greater the corresponding adjustment, and vice versa. Then, there exists: in, For terrain images Medium pixel The adjusted pixel value of Pixel The original pixel value, For terrain images The corresponding light attenuation coefficient.

5. The terrain recognition method for an underwater dredging robot according to claim 4, characterized in that: The method of obtaining a connected domain corresponding to a terrain image based on any layer and filtering out existing edge texture areas is as follows: Obtain the corresponding connected domain for the terrain image at a certain scale, and then filter out the possible edge texture areas from it, which can be expressed as: in, For a connected area in this terrain image is the confidence of the edge texture area; Connected domain The average gray value of The average gray value of this terrain image; For this connected domain, the modulus of the largest principal component vector after PCA analysis is: The average modulus of all principal component vectors in this connected domain; therefore Represents the connected domain in this terrain image Grayscale difference, due to the strong contrast of grayscale in edge texture area, The larger the value, the more likely it is an edge texture area; Represents the morphological extension trend of this connected domain. Since the largest principal component vector in PCA Represents the most obvious extension degree, that is, if the extension trend of this connected domain is more obvious, the corresponding Relative to The larger the value is, the more likely it is an edge texture area; finally, the result is normalized to the range of [0, 1], and the The corresponding area is the high-frequency edge texture area in the image.

6. The terrain recognition method for an underwater dredging robot according to claim 5, characterized in that: The defogging and restoration of the terrain image based on the final defogging weight is performed by using the dark color prior, specifically: The terrain image at each scale is used as the input image, and the dark primary color of each pixel is calculated to generate a dark primary color image , is each pixel in the image; By calculating the dark primary image The maximum value position in the atmospheric illumination value is estimated and transmittance ; By the estimated transmittance and atmospheric illumination values And the defogging weights to restore a clear fog-free image: in, After defogging Layer terrain image, For the original Layer-scale foggy images, For the The defogging weight of the layer terrain image is determined; that is, the dark color prior method is optimized according to the defogging weight of the different layer terrain images to improve the defogging effect.

7. The terrain recognition method for an underwater dredging robot according to claim 6, characterized in that: The image at each scale is used as the input image, and the dark primary color of each pixel is calculated to generate a dark primary color image. , specifically: Assume the input image is ,in, It is each pixel in the image; since some color channels of natural scene images are very dark, the minimum value of the RGB channels of each pixel is used to represent the "dark primary color"; For an RGB image, the dark primary is defined as: in, 、 、 are the pixel values ​​of the image in the red, green, and blue channels respectively; The dark primary image is calculated by The maximum value position in the atmospheric illumination value is estimated and transmittance , specifically: In dark color image Find the pixel position with the maximum value in the image; the pixel position with the maximum value corresponds to the darkest part of the image; by obtaining the pixel value in the original image corresponding to the maximum value position in the dark original image, infer the atmospheric illumination value, and select the RGB pixel value corresponding to the maximum dark original value as the atmospheric illumination value ; Assume that the image The maximum position in Corresponding pixel value in the original image , then the atmospheric illumination value for: in, is an image The pixel value corresponding to the maximum dark primary color position in ; Transmittance Reflects the haze level of each pixel in the image, transmittance It is calculated by estimating the dark primary color of each pixel of the image; The formula is as follows: in, is a constant whose value range is [0,1].

8. A terrain recognition system for an underwater dredging robot, characterized in that: include: a first acquisition module, wherein the first acquisition module acquires a real-time underwater terrain image based on the underwater dredging robot and preprocesses the acquired terrain image; a second acquisition module, which acquires a light attenuation coefficient based on the shooting depth of the underwater dredging robot and the influence of plankton in the water; An adjustment module, wherein the adjustment module brightens the grayscale of different pixels in the pre-processed terrain image based on the light attenuation coefficient; a decomposition module, wherein the decomposition module performs Laplace decomposition on the adjusted terrain image to obtain a multi-layer terrain image; A screening module, which obtains a connected domain corresponding to a terrain image of any layer and screens out existing edge texture areas; A determination module, which obtains the weight of the importance of high-frequency information in each layer of terrain image based on the proportion of edge texture areas in each layer of terrain image, and determines the final defogging weight in combination with the light attenuation coefficient; The proportion of edge texture area in each layer of terrain image is: in, For the The edge texture area of ​​the terrain image layer, For the The overall area of ​​the layer topographic image; The weight of the importance of high-frequency information in each layer of terrain image is obtained as follows: Put the proportion of edge texture area in each layer of terrain image into the sequence For terrain images at any scale, there exists: in, For the Importance weight of high frequency information in layer terrain images; For the The area ratio of high-frequency information in the topographic image layer, For the The area ratio of high-frequency information in the topographic image layer, is the number of decomposition layers; For the The relative value of the high-frequency information area of ​​the topographic image of the layer; the larger the value, the more high-frequency information of the corresponding topographic image, and the greater the importance weight; Represents the relationship between the number of layers and high-frequency information. Since the lower the number of layers, the more important the corresponding high-frequency information is, the greater the corresponding importance weight; The final defogging weight is determined in combination with the light attenuation coefficient, specifically: in, For the Dehazing weights of the layer terrain image; For the Importance weight of high frequency information in layer terrain images, is the light attenuation coefficient of the original image; the original image is a real-time underwater terrain image obtained by the underwater dredging robot; The reconstruction module defogs and restores the terrain image based on the final defogging weight through a dark color prior, and synthesizes terrain image information of multiple scales into a complete terrain image through inverse Laplacian pyramid reconstruction, thereby obtaining underwater terrain information.

Images