An image enhancement method for visual identification of winch cable status
Through the minimum error threshold optimization model and the segmented linear transformation method of the adaptive proportional coefficient k2, the problem of low quality and complex parameter adjustment of winch cable status images under complex lighting conditions is solved, and efficient cable feature extraction and accurate judgment are achieved. .
Patent Information
- Application Number
- CN202211032759.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-08-26
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2042-08-26
AI Technical Summary
The winch cable-coil state image is severely affected by light and environmental interference, resulting in low image quality and difficult to extract cable state characteristics. The existing image enhancement methods are poor in complex lighting conditions and complex parameter adjustments.
The minimum error threshold optimization model and the proportional coefficient k2 are used to adaptively solve the segmented linear transformation parameters, and the segmented linear transformation enhancement of the image is achieved by segmenting the image area and adjusting the stretching and shrinking ratios.
It effectively improves the image quality of the winch cable discharge status, simplifies the extraction of cable characteristics, improves the visual judgment accuracy of the winch cable discharge status, and is suitable for a variety of lighting conditions.
Smart Images

Figure CN115393590B_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to a method for visual identification image enhancement, in particular to a method for visual identification image enhancement for winch cable arrangement state. Background Art
[0002] Winches are widely used in many fields such as mines, ports, construction, and oceans. The cable-laying status of winches is directly related to whether engineering operations can operate normally and whether equipment can be safely retracted and released. The cable-laying status identification method based on a visual device obtains and analyzes the winch cable-laying status image, and mainly identifies the winch cable-laying status based on features such as cable center distance, number of inter-layer jumps, and number of cable edges. However, due to the complex and changeable working environment of the winch, the acquired images are seriously affected by lighting and on-site environment, resulting in low quality of the cable-laying status image, great difficulty in subsequent processing and analysis, and difficulty in extracting cable status features. Therefore, how to effectively enhance the quality of the cable-laying status image is crucial to accurately extract cable status features and timely identify the winch cable-laying status.
[0003] At present, there are generally linear transformation and nonlinear transformation methods for image enhancement of winch status images. The linear transformation method stretches the image grayscale according to a linear formula, and the transformation effect is single. Nonlinear transformation methods include GAMA transformation, which has a good transformation effect under single lighting conditions and is not suitable for complex lighting conditions; the MSR method achieves image enhancement by convolving the original image with Gaussian blur at three different scales, and maintains a good enhancement effect under complex lighting conditions, but Gaussian blur also increases the difficulty of subsequent cable feature extraction; the piecewise linear transformation method is more flexible and can enhance the image according to the cable characteristics, but in practical applications, the four parameters need to be manually adjusted based on prior data. Therefore, in view of the limitations of the above methods, the model and proportional coefficient are optimized through the minimum error threshold. k 2 Adaptively and quickly determine the piecewise linear transformation parameters to achieve grayscale image enhancement, facilitate cable feature extraction, and determine the winch cable arrangement status. Summary of the invention
[0004] In order to overcome the shortcomings of the existing visual discrimination image enhancement method, the present invention discloses an image enhancement method for visual discrimination of winch cable arrangement status.
[0005] The present invention adopts the following technical solution: an image enhancement method for visually distinguishing the winch cable arrangement state comprises the following steps.
[0006] Step 1: Minimum error threshold optimization method to solve the segmentation point x 2.
[0007] Select the gray level as L The cable row image is taken as the image to be enhanced, and the pixel points in the image to be enhanced ( i , j) has a gray value of f ( i , j ),and f ( i , j )∈{0, 1, …, L}, solve the minimum error threshold optimization model according to formula (1) J ( x ), get the split point x 2. Divide the image to be enhanced into region I and region II, corresponding to the background region and other regions respectively;
[0008] (1)
[0009] In the formula, P 1( x ) is the distribution probability of region I, , p i The gray value in the grayscale image is f ( i , j )The probability of a pixel appearing and satisfying ; P 2( x ) is the distribution probability in region II, ; is the variance of region I, ; μ 1( x ) is the mean value of region I, ; is the variance of region II, ; μ 2( x ) is the mean value of region II, .
[0010] Step 2: Minimum error threshold optimization method to solve the segmentation point x 1.
[0011] Select the other regions obtained in step 1 as the image to be enhanced, and solve the minimum error threshold optimization model twice according to formula (1): J ( x ), get the split point x 1( x 2> x 1) The image to be enhanced is divided into region I and region II, which correspond to the target region and the transition region respectively.
[0012] Step 3: Select the transformation scale factor k 2 Adaptive solution y 1. y 2.
[0013] Assume that the target area stretching ratio is k 1. The background area shrinkage coefficient is k 3, and satisfy the relationship
[0014] k 1 k 3=1, and k 1>1 (2)
[0015] The target area stretching ratio coefficient is obtained from the geometric relationship k 1 is as follows
[0016] (3)
[0017] In the formula, M f The image to be enhanced f ( i , j ), M g For the enhanced image g ( i , j )’s maximum grayscale value.
[0018] Combining equation (2) and equation (3), we get the equation
[0019] (4)
[0020] Solving equation (4) yields
[0021] (5)
[0022] In the formula, .
[0023] Select the transition area scale factor k 2. Obtained from geometric relationship
[0024] (6).
[0025] Step 4: Piecewise linear transformation enhancement.
[0026] Cable image f ( i , j ) According to formula (7), piecewise linear transformation image enhancement is performed to obtain the enhanced image g ( i , j )for
[0027] (7).
[0028] The beneficial effect of the present invention is that the present invention solves the segmentation point by using the minimum error threshold optimization method. x 1. x 2. Select the scale factor k 2 Adaptive solution of split points y 1. y 2. Perform piecewise linear transformation on different regions of the cable status image to enhance the image. Only one parameter needs to be adjusted. k 2. It is convenient, simple, fast and efficient. It effectively solves the problems of existing methods being affected by complex lighting and having many parameters. It highlights the edge details of the cable, suppresses background area interference, enhances the image quality of the cable arrangement state, and provides a basis for subsequent cable feature extraction. It improves the accuracy of visual judgment of the winch cable arrangement state and ensures the safety of winch operation. BRIEF DESCRIPTION OF THE DRAWINGS
[0029] Figure 1 It is a flow chart of the image enhancement method for visual identification of winch cable arrangement status of the present invention;
[0030] Figure 2 It is a piecewise linear transformation relationship diagram of the present invention;
[0031] Figure 3 This is an overexposed image of the winch cable arrangement;
[0032] Figure 4 This is the image of the winch cable arrangement under normal exposure;
[0033] Figure 5 This is an underexposed image of the winch cable arrangement;
[0034] Figure 6 This is the enhanced effect of the winch cable arrangement image under overexposure;
[0035] Figure 7 This is the enhanced image of the winch cable arrangement under normal exposure;
[0036] Figure 8 This is the enhanced effect of the underexposed winch cable image. Specific implementation plan
[0037] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.
[0038] See attached figure, Figure 1The present invention is a flow chart of an image enhancement method for visually distinguishing the cable-arranging state of a winch, a method for visually distinguishing the cable-arranging state of a winch, and includes the following four steps.
[0039] Step 1: Minimum error threshold optimization method to solve the segmentation point x 2.
[0040] Select Gray Scale L The cable row images are 255, and the cable row images under three different exposure conditions are selected as the images to be enhanced. Figure 3 This is an overexposed winch cable image. Figure 4 To expose the normal winch cable image, Figure 5 For the underexposed winch cable image, the minimum error threshold optimization model is solved according to formula (1) J ( x ), get the split point x 2, corresponding to 235, 147, 96, the selected image to be enhanced is divided into region I and region II, corresponding to the background area and other areas respectively.
[0041] Step 2: Minimum error threshold optimization model to solve segmentation points x 1.
[0042] Select from step one Figure 3 , Figure 4 , Figure 5 The other areas segmented are the images to be enhanced, and then the minimum error threshold optimization model is solved according to formula (1) J ( x ), get the split point x 1, corresponding to 163, 53, 2, the image to be enhanced is divided into region I and region II, corresponding to the target region and transition region respectively.
[0043] Step 3: Proportional coefficient k 2 Adaptive solution of split points y 1. y 2.
[0044] Get the maximum gray value of the image to be enhanced M f , the maximum gray value of the enhanced image M g Both are 255, take the transition area ratio coefficient k 2 is 0.6, combined Figure 3 The segmentation points between the target area, transition area and background area x 1 is 163, x 2 is 235, according to Figure 2 The geometric relationship is derived from equation (6) to obtain the segmentation point y 1 is 195,y 2 is 238; similarly, for Figure 4 The segmentation points between the target area, transition area and background area x 1 is 53, x 2 is 147, and the split point can be obtained y 1 is 163, y 2 is 220; for Figure 5 The segmentation points between the target area, transition area and background area x 1 is 2, x 2 is 96, which gives the split point y 1 is 197, y 2 is 253.
[0045] Step 4: Piecewise linear transformation enhancement.
[0046] According to formula (7), Figure 3 , Figure 4 , Figure 5 Perform piecewise linear enhancement to obtain the enhanced effect of the cable row image, as shown in the figure. Figure 6 , Figure 7 , Figure 8 As shown; In order to verify the superiority of the present invention, the GAMA transformation method is selected to compare the objective performance indicators with the present invention, and the image standard deviation reflecting the discrete degree of image grayscale distribution is used σ The larger the value, the more dispersed the image grayscale distribution, the more prominent the cable edge, and the better the method performance. The larger the value of Entropy, which reflects the amount of information contained in the image, the richer the information contained in the image, and the better the method performance. Figure 3 , Figure 4 , Figure 5 The objective performance indicators after the GAMA method and the present invention are enhanced are shown in Table 1. From the data in Table 1, it can be seen that:
[0047] Table 1 Objective performance indicators of different methods
[0048]
[0049] 1) Under overexposure conditions, Figure 3 , Figure 4 , Figure 5 After enhancement by the present invention, the standard deviations are better, which are 55.5198, 50.3382, and 54.3443, respectively, while after enhancement by the GAMA method, the standard deviations are 26.9030, 24.7354, and 26.2007, respectively; Figure 3 , Figure 4 , Figure 5The entropies after enhancement by the present invention are better, which are 6.5213, 6.5939, and 6.8337, respectively, while the entropies after enhancement by the GAMA method are 5.7022, 5.7466, and 5.9528, respectively;
[0050] 2) Under normal exposure conditions, Figure 3 , Figure 4 , Figure 5 After enhancement by the present invention, the standard deviations are better, which are 52.6175, 58.5896, and 54.9576, respectively, while after enhancement by the GAMA method, the standard deviations are 36.8395, 45.0385, and 38.2223, respectively. Figure 3 , Figure 4 , Figure 5 The entropies after enhancement by the present invention are better, which are 7.2842, 7.2762, and 7.4248, respectively, while the entropies after enhancement by the GAMA method are 7.0563, 7.1798, and 7.1487, respectively;
[0051] 3) Under underexposure conditions, Figure 3 , Figure 4 , Figure 5 After enhancement by the present invention, the standard deviations are better, which are 97.1552, 96.0100, and 97.0752, respectively, while after enhancement by the GAMA method, the standard deviations are 61.7307, 63.0881, and 62.4791, respectively. Figure 3 , Figure 4 , Figure 5 The entropies after enhancement by the present invention are better, which are 5.3195, 5.4274, and 5.4035, respectively, while the entropies after enhancement by the GAMA method are 5.1261, 5.1759, and 5.2054, respectively;
[0052] Therefore, the objective performance indicators of the present invention under three lighting conditions are σ and Entropy are higher than those of the GAMA method, indicating that the image enhancement effect of the present invention is better than that of the GAMA method, which can effectively enhance the quality of cable row images, provide a good foundation for subsequent feature extraction, and achieve good results under various lighting conditions.
[0053] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims
1. A method for visually distinguishing the winch cable status image enhancement, characterized in that: The minimum error threshold optimization method is used to determine the grayscale interval segmentation points of the image to be enhanced. x 1. x 2. Select the transformation ratio k 2 Adaptively determine the grayscale interval segmentation points of the enhanced image y 1. y 2. Enhance the image quality of marine winch cable arrangement by piecewise linear transformation, and proceed in the following four steps: Step 1: Use the minimum error threshold optimization method to solve x 2. Determine the segmentation point between the gray background area and other areas of the image to be enhanced; Step 2: Use the minimum error threshold optimization method to solve x 1. Determine the segmentation point between the grayscale target area and the transition area of the image to be enhanced; Step 3: Select the transformation ratio coefficient k 2 Adaptive solution y 1. y 2. Determine the grayscale interval segmentation points corresponding to the enhanced image; Step 4: Piecewise linear transformation image enhancement; Step 1: Minimum error threshold optimization method to solve the segmentation point x 2 Select the gray level as L The cable row image is taken as the image to be enhanced, and the pixel points in the image to be enhanced ( i , j ) has a gray value of f ( i , j ),and f ( i , j )∈{0, 1, …, L }, solve the minimum error threshold optimization model according to formula (1) J ( x ), get the split point x 2. Divide the image to be enhanced into region I and region II, corresponding to the background region and other regions respectively; (1) In the formula, P 1( x ) is the distribution probability of region I , p i The gray value in the grayscale image is f ( i , j )The probability of a pixel appearing and satisfying ; P 2( x ) is the distribution probability in region II, ; is the variance of the grayscale value in region I, ; μ 1( x ) is the mean gray value of region I, ; is the variance of the grayscale value in region II, μ 2( x ) is the mean gray value of region II, ; Step 2: Minimum error threshold optimization method to solve the segmentation point x 1 Select the other regions obtained in step 1 as the image to be enhanced, and solve the minimum error threshold optimization model according to formula (1): J ( x ), get the split point x 1( x 2> x 1) Divide the image to be enhanced into region I and region II, which correspond to the target region and transition region respectively; Step 3: Select the transformation scale factor k 2 Adaptive solution y 1. y 2 Assume that the target area stretching ratio is k 1. The background area shrinkage coefficient is k 3, and satisfy the relationship k 1 k 3=1, and k 1>1 (2) The target area stretching ratio coefficient is obtained from the geometric relationship k 1 is as follows (3) In the formula, M f The image to be enhanced f ( i , j ), M g For the enhanced image g ( i , j )’s maximum grayscale value; Combining equation (2) and equation (3), we get the equation (4) Solving equation (4) yields (5) In the formula, ; Select the transition area scale factor k 2. Obtained from geometric relationship (6) Step 4: Piecewise linear transformation enhancement Cable image f ( i , j ) According to formula (7), piecewise linear transformation image enhancement is performed to obtain the enhanced image g ( i , j )for (7)。
