Image-enhancement-based color masterbatch coloring performance detection method and system

Abstract

This invention belongs to the field of image processing technology, specifically relating to a method and system for detecting the coloring performance of masterbatch based on image enhancement. The method includes the following steps: S1, acquiring an initial grayscale image of the mixed sample, and performing a difference operation on the background light field image extracted from the initial grayscale image to obtain a preprocessed feature map; S2, constructing texture-aware adaptive gain weights by analyzing the energy activity of local regions in the preprocessed feature map, and determining the local contrast gain intensity of each pixel; S3, calculating the gradient distribution symmetry of each pixel in the preprocessed feature map in multiple directions to obtain the local isotropic shift index. This invention can physically remove processing flow mark interference, significantly improve the signal-to-noise ratio between defects and the background, and achieve a reliable assessment of the coloring performance and production stability of masterbatch.

Description

Technical Field

[0001] This invention belongs to the field of image processing technology, specifically relating to a method and system for detecting the color performance of masterbatches based on image enhancement. Background Technology

[0002] In the industrial production of high-performance rubber products, the coloring stability, dispersion uniformity, and batch consistency of masterbatches are core quality indicators that determine the product's appearance and service life, directly impacting its market competitiveness and long-term reliability. For high-concentration black masterbatches, after being mixed with rubber substrates through multiple processes such as intensive mixing, open milling, and extrusion, the finished product surface image typically exhibits a near-complete black state with extremely low contrast, making it difficult to distinguish differences visually or through conventional imaging. In this scenario, the grayscale difference between pigment agglomeration defects such as particles, spots, and streaks and the substrate background is extremely weak, easily masking signal characteristics and posing a significant challenge to automated online inspection.

[0003] Current quality inspection methods in the industry largely rely on traditional image enhancement algorithms, which have significant limitations and shortcomings in actual mass production environments. First, the physical flow marks generated during rubber processing, such as mixing, extrusion, and calendering, exhibit significant directional and periodic texture characteristics. Their gradient changes are often stronger than the actual defect signals, easily leading to algorithmic misjudgments and texture confusion. Second, while existing linear enhancement and conventional filtering methods increase the brightness difference of defects, they inevitably amplify interference from environmental noise and the microscopic inhomogeneities of the material itself, resulting in a decrease in image signal-to-noise ratio and quality degradation. Finally, existing solutions are mostly qualitative judgments, lacking objective evaluation indicators that can directly quantify the dispersion level of masterbatch and guide the adjustment of production process parameters, making it difficult to support closed-loop process optimization.

[0004] Therefore, how to accurately extract weak masterbatch dispersion defects under strong background texture interference, effectively eliminate non-defect interference such as flow marks and processing lines, and establish a stable and reliable quantitative evaluation system has become a key technical challenge to improve the detection accuracy of high-concentration black masterbatch and achieve high-quality and stable production of rubber products. Summary of the Invention

[0005] This invention provides a method and system for detecting the coloring performance of masterbatch based on image enhancement, in order to solve the technical problem of poor detection accuracy in high-concentration black masterbatch images due to the extremely small grayscale difference between defects and the background and interference from processing flow marks.

[0006] In a first aspect, the present invention provides a method for detecting the coloring performance of masterbatch based on image enhancement, comprising the following steps: S1, obtain the initial grayscale image of the mixed sample, and perform a difference operation on the background light field image extracted from the initial grayscale image to obtain a preprocessed feature map; S2, by analyzing the energy activity of local regions in the preprocessed feature map, construct texture-aware adaptive gain weights and determine the local contrast gain intensity of each pixel; S3, calculate the gradient distribution symmetry of each pixel in the preprocessed feature map in multiple directions to obtain the local isotropic offset index, and decouple the processing flow marks and scattered defects based on the local isotropic offset index. S4. Based on the adaptive gain weight and the local isotropic offset index, the preprocessed feature map is reconstructed to obtain the enhanced detection map after the forced fading of the flow mark interference, and the coloring performance of the masterbatch is determined based on the enhanced detection map.

[0007] Its effects are as follows: addressing the challenge of distinguishing between background grayscale and defects in high-concentration masterbatch compounded samples, which is also affected by processing flow marks, this method constructs texture-aware adaptive gain weights and local isotropic offset exponents to achieve active enhancement of weak defect signals and physical decoupling of strong directional flow marks. Compared with traditional global enhancement algorithms, this method can significantly improve the clarity of pigment agglomeration points under low contrast while effectively suppressing periodic texture false alarms caused by the extrusion process, thus significantly improving the accuracy of online detection.

[0008] Further, obtaining the background light field image includes: performing convolution processing on the initial grayscale image using a Gaussian smoothing operator, extracting the low-frequency background trend in the initial grayscale image, and obtaining the background light field image.

[0009] Its effects are as follows: by using the Gaussian smoothing operator to extract large-scale low-frequency background trends from the original image, it can accurately construct the illumination gradient caused by uneven ambient lighting or the curvature of the material surface; through convolution and extraction operations, it provides an accurate background benchmark for subsequent difference operations, ensuring that subsequent processing can be carried out on a pure data surface with uniform brightness, and solving the problem of false detection caused by light source angle or small undulations of the sample in industrial field.

[0010] Furthermore, the adaptive gain weights are calculated using the following formula:

[0011] In the formula, Representing coordinates Adaptive gain weights at the location, Represented by pixels The standard deviation of pixels in the neighborhood centered on the target. This represents the standard deviation of the total number of pixels in the image. This represents the system reference constant.

[0012] Its effects are as follows: by using a logarithmic gain formula that includes the ratio of local to global standard deviation, a dynamic enhancement mechanism that can sense texture activity is established; low gain is maintained in smooth background areas to avoid amplifying camera noise, while the gain intensity is automatically increased in suspected defect areas with energy abrupt changes, thus achieving targeted brightening of weak defect signals and fundamentally solving the drawback of conventional linear stretching methods causing noise explosion across the entire image when enhancing details.

[0013] Furthermore, the local isotropic migration index is calculated using the following formula:

[0014] In the formula, Indicates the local isotropic shift index. Represents pixels The maximum value among the gradient magnitudes in the horizontal, vertical, and diagonal directions. This represents the average value of the gradient magnitude in all directions. This represents the stability constant.

[0015] Its effect is that by calculating the difference ratio between the maximum gradient magnitude and the average magnitude, the gradient directionality of the pixel is quantified; the mathematical model can accurately distinguish the processing flow marks with a single dominant direction from the pigment agglomeration points that diffuse evenly in all directions at the numerical level, providing a solid mathematical basis for the subsequent removal of false defects, and enabling the algorithm to have the logical ability to identify true and false defects.

[0016] Furthermore, the enhanced detection map is reconstructed, including:

[0017] In the formula, This represents the pixel value of the enhanced detection map. This represents the pixel values ​​of the preprocessed feature map. Indicates the penalty order. Representing coordinates Adaptive gain weights at the location, This represents the local isotropic shift index.

[0018] Its effects are as follows: by adopting a nonlinear penalty reconstruction formula based on isotropic offset index, the feature decoupling results are directly applied to the pixel value reconstruction process; a high-order penalty is applied to suppress strongly directional pixels that are determined to be flow marks, while retaining high gain for cluster defects, thereby physically erasing the processing texture in the generated enhanced detection map, intuitively highlighting the real coloring defects, and greatly reducing the signal-to-noise ratio loss caused by flow marks.

[0019] Furthermore, the coloring performance of the masterbatch is determined based on the enhanced detection map, including: determining the defect response coefficient according to the area ratio of the feature points detected in the enhanced detection map to the entire map; By combining the pixel distribution of the enhanced detection map and the defect response coefficient, a production stability evaluation index is calculated to evaluate the coloring performance of the masterbatch.

[0020] Furthermore, the production stability evaluation index is calculated using the following formula:

[0021] In the formula, Indicates the production stability evaluation index. Indicates the defect response coefficient. This indicates the total number of pixels in the sample image. This represents the pixel value of the enhanced detection map.

[0022] Further, a preprocessed feature map is obtained, including: performing pixel-by-pixel difference between the initial grayscale image and the background light field image to eliminate the illumination gradient caused by uneven ambient illumination.

[0023] Furthermore, the physical logic of the local isotropic shift index is as follows: if there are processing flow marks in the local area, the maximum directional gradient is significantly greater than the average gradient, and the local isotropic shift index increases; if the local area is a spherical pigment agglomeration point, the gradients in each direction are similar, and the local isotropic shift index tends to zero.

[0024] Secondly, the present invention provides a color masterbatch coloring performance detection system based on image enhancement, including a memory and a processor. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned color masterbatch coloring performance detection method based on image enhancement is implemented.

[0025] The beneficial effects are: The core innovation of this invention lies in constructing an adaptive gain operator with clear physical meaning and a locally isotropic decoupling exponent, thus abandoning the black-box design and uncertainty of traditional image detection algorithms. Through in-depth mining and quantitative analysis of the symmetry of pixel multi-directional gradient distribution, it accurately captures the essential differences in characteristics between processing flow marks and coloring defects, achieving physical forced fading of processing flow marks in the mixed sample image, removing interference signals at their source, and significantly improving the detection accuracy and recognition accuracy of defects such as pigment agglomeration in low signal-to-noise ratio environments. Simultaneously, this invention innovatively transforms visual inspection results into a quantitative scoring index for production stability with exponential decay characteristics, making the evaluation of masterbatch dispersion quality more aligned with the actual perception and judgment logic of industrial production. Furthermore, the adjustment of each parameter in the algorithm corresponds to clear physical logic and process meaning, providing reliable data support and precise adjustment methods that are implementable and traceable for optimizing and adjusting on-site production processes, thus helping to achieve closed-loop quality control in masterbatch production. Attached Figure Description

[0026] Figure 1 This is a flowchart of the image enhancement-based color masterbatch coloring performance detection method of the present invention.

[0027] Figure 2 This is a comparison chart of the original image, the conventionally enhanced image, and the enhancement effect of the present invention.

[0028] Figure 3 This is a feature decoupling distribution diagram of the present invention.

[0029] Figure 4 This is a production stability evaluation curve for the present invention. Detailed Implementation

[0030] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0031] An embodiment of the image enhancement-based color masterbatch coloring performance detection method provided by the present invention: like Figure 1 As shown, the image enhancement-based method for detecting the color performance of masterbatch includes the following steps: S1. Obtain the initial grayscale image of the mixed sample, and perform a difference operation on the background light field image extracted from the initial grayscale image to obtain a preprocessed feature map.

[0032] In this step, an industrial vision system is used to take standardized photos of the mixed rubber samples, acquiring high-resolution initial grayscale images. To ensure the clarity and consistency of image acquisition, providing high-quality foundational data for subsequent image processing, this method employs a large-size 31×31 Gaussian smoothing operator on the initial grayscale image to eliminate interference from uneven lighting in industrial environments and the slight curvature of the sample surface. By performing global convolution processing and leveraging the Gaussian smoothing low-pass filtering properties, high-frequency details in the image are effectively filtered out, accurately extracting the low-frequency background trend dominated by illumination changes in the initial grayscale image. This results in the construction of a background light field image that perfectly matches the pixel dimensions of the original image. Then, a pixel-by-pixel difference operation is performed between the initial grayscale image and the background light field image, using the formula... The preprocessed feature map is calculated.

[0033] For example, if the left side of the sample image receives slightly stronger light, corresponding to a grayscale value of 50, while the right side receives weaker light, with a grayscale value of 45, after background light field extraction, the background light field values ​​on both sides will precisely match the illumination distribution of the original image, corresponding to 50 and 45 respectively. After pixel-by-pixel difference calculation, the grayscale values ​​of the background on both sides will both return to near 0, achieving complete cancellation of illumination differences. This spatial domain background difference processing can completely eliminate image substrate deviation caused by large-scale illumination gradients, allowing subsequent defect detection to be no longer interfered with by illumination factors, laying a uniform and pure image data foundation for contrast enhancement and feature extraction of weak defect signals.

[0034] S2 constructs texture-aware adaptive gain weights by analyzing the energy activity of local regions in the preprocessed feature map and determines the local contrast gain intensity of each pixel.

[0035] The calculation formula is as follows: (Based on the analysis of the dynamic distribution gain of local energy activity)

[0036] In the formula, Representing coordinates Adaptive gain weights at the location, Represented by pixels The standard deviation of pixels in the neighborhood centered on the target. This represents the standard deviation of the total number of pixels in the image. This represents the system reference constant.

[0037] The core of this formula is based on the strong correlation between local energy differences and defect features, balancing global stability and local specificity, and abandoning the irrationality of traditional fixed gain or black-box gain allocation. Its design principle revolves around the practical needs of industrial image inspection: First, based on... As a core indicator of local energy activity, the pixel grayscale fluctuations in defective areas are much greater than those in the background areas. The value will be significantly higher, which can be used as a basis for a preliminary distinction between defects and background; secondly, introduce As a global reference benchmark, it is used to balance the overall grayscale fluctuation differences between different images and avoid local gain imbalance caused by global brightness anomalies in a single image; The system reference constant can effectively suppress To prevent gain distortion caused by the denominator approaching zero when the value is too small, the algorithm's stability is ensured. Simultaneously, the natural logarithm function is employed. As a gain mapping relationship, it can both... It achieves moderate gain enhancement in defective areas while avoiding excessive gain that could lead to noise amplification, thus benefiting low-end regions. Gain suppression is applied to the background area to meet the core detection requirements of enhanced defects and stable background.

[0038] Assuming a normal background area for And suspected defects for Standard deviation of all pixels in the image for System reference constants for For background points, the gain weight is calculated as follows: For defects, the weight is... .

[0039] S3. Calculate the gradient distribution symmetry of each pixel in the preprocessed feature map in multiple directions to obtain the local isotropic offset index, and decouple the processing flow marks and scattered defects based on the local isotropic offset index.

[0040] The Sobel operator is used to calculate the gradient magnitudes in the horizontal, vertical, and diagonal directions. ; Calculate the offset index:

[0041] In the formula, Indicates the local isotropic shift index. Represents pixels The maximum value among the gradient magnitudes in the horizontal, vertical, and diagonal directions. This represents the average value of the gradient magnitude in all directions. This represents the stability constant.

[0042] The core of this formula is based on the essential physical difference in gradient direction between process flow marks and dispersed defects. Through mathematical quantification, it achieves precise decoupling between these two types of interference and defects, providing crucial support for subsequent defect detection by eliminating flow mark interference. Its design principle closely aligns with actual industrial inspection needs, progressively ensuring both discrimination accuracy and algorithm stability: First, it clarifies that process flow marks exhibit strong directionality, characterized by a significantly higher gradient amplitude in one direction and weaker gradients in other directions. In contrast, dispersed defects such as spherical agglomerates are isotropic, with uniform gradient amplitudes in all directions. Therefore, it selects... and The difference is used as the numerator to quantify the degree of unevenness of gradients in each direction. The larger the difference, the stronger the directionality of the region, and the more likely it is a processing flow mark; the smaller the difference, the more uniform the gradient in each direction, and the more likely it is a dispersed defect or a normal region. Secondly, the following is introduced... As a component of the denominator, it balances the differences in gradient magnitude across different grayscale ranges and noise environments, preventing distortion of the offset exponent due to an overall gradient that is too high or too low, and ensuring consistency of quantization standards across different scenarios. Finally, add... The stability constant's core function is to suppress... The exponential distortion caused by an excessively small denominator when the value approaches zero is mitigated. At the same time, the influence of slight noise on the gradient mean is reduced, and misjudgments caused by irrelevant noise are prevented. This ensures that the formula can still output effective quantization results stably in low signal-to-noise ratio environments. Ultimately, this exponent enables precise mathematical differentiation between processing flow marks and dispersion defects.

[0043] Assume a machining flow mark with a vertical gradient of... The gradients in all other directions are ,but , ,exist hour, If it is a spherical aggregation point, the gradient in all directions is... ,but , , .

[0044] S4. Based on the adaptive gain weight and the local isotropic offset index, the preprocessed feature map is reconstructed to obtain the enhanced detection map after the forced fading of the flow mark interference, and the coloring performance of the masterbatch is determined based on the enhanced detection map.

[0045] Based on the reconstruction formula: The final enhanced map is obtained, where, This represents the pixel value of the enhanced detection map. This represents the pixel values ​​of the preprocessed feature map. Indicates the penalty order. Representing coordinates Adaptive gain weights at the location, This represents the local isotropic shift index.

[0046] The core of this formula is to integrate the advantages of adaptive gain and isotropic offset exponent mentioned earlier. Through layered design, it achieves the triple goals of preserving defects, suppressing flow marks, and optimizing image quality, perfectly meeting the actual needs of defect detection in industrial rubber samples. The design of each factor corresponds to a clear physical logic and detection purpose. Its design principle is progressive and synergistic: First, it uses preprocessed feature maps... As a foundation, it is ensured that the image has eliminated the interference of uneven illumination, providing a clean image base for subsequent enhancement and compression; secondly, it introduces... The gain factor, continuing the gain allocation logic of local energy activity mentioned earlier, applies to... Areas with excessively high blemishes are targeted for enhancement, amplifying the brightness and contrast of minor imperfections, while maintaining low gain for background areas to avoid noise amplification. Finally, the design... The nonlinear penalty factor's core function is to precisely suppress processing flow marks. A high value indicates that this factor can achieve a non-linear decay of the brightness in the flow mark region through exponential penalty. The intensity of suppression can be flexibly adjusted as a penalty level; while flaws When the value approaches 0, the factor approaches 1, which can completely preserve the brightness of the defect point, realize the physical separation of flow marks and defects, and at the same time avoid excessive punishment that would cause the defect features to be distorted, thus ensuring the accuracy of subsequent detection.

[0047] When the penalty order At that time, regarding the aforementioned The flow mark points, the denominator term becomes The brightness at that point was significantly suppressed; for The defects, the denominator is The brightness is fully preserved.

[0048] Finally, the evaluation index was calculated: Defects In the formula, Indicates the production stability evaluation index. Indicates the defect response coefficient. This represents the total number of pixels in the sample image. Higher brightness and larger coverage area of ​​the defective pixels will result in a more pronounced difference. The lower.

[0049] Through a nonlinear penalty mechanism, flow mark interference can be physically eliminated, and the dispersion quality can be transformed into an intuitive percentage-based indicator.

[0050] Finally, the beneficial effects of this technical solution will be further explained with reference to the accompanying drawings.

[0051] like Figure 2The image shows a comparison of the original image, the conventionally enhanced image, and the enhancement effect of the present invention. The leftmost image is the original image, showing the initial morphology of the high-concentration black master sample. The background is dark, and flow marks and agglomeration points are mixed and extremely difficult to distinguish with the naked eye. The middle image is the prior art, showing the effect after conventional enhancement. It can be seen that the background is forced to be brightened, and the flow marks become coarse and black, which seriously misleads the detection. The rightmost image is the effect of the present invention, showing the effect after the flow marks are forcibly eliminated. The background remains deep black, the flow marks disappear, and only the pigment agglomeration defects appear as extremely bright white spots, with a sharp contrast.

[0052] like Figure 3 The image shown is a feature decoupling diagram: displayed in scatter plot form. The distribution clearly demonstrates the perfect mathematical division between flow marks and defects.

[0053] like Figure 4 The image shown is a stability rating graph, illustrating the continuous stability score. Each batch The curve clearly shows the boundary between the acceptable and risky zones, providing a direct reflection of the production stability of the masterbatch's coloring performance.

[0054] An embodiment of the color masterbatch coloring performance detection system based on image enhancement provided by the present invention: The image-enhanced masterbatch coloring performance detection system includes a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement the image-enhanced masterbatch coloring performance detection method described above.

[0055] The image-enhanced masterbatch coloring performance testing system also includes other components well known to those skilled in the art, such as communication interfaces. Their settings and functions are known in the art and will not be described in detail here.

[0056] In this invention, the aforementioned memory can be any tangible medium containing or storing a program that can be used or combined with an instruction execution system, apparatus, or device. For example, a computer-readable storage medium can be any suitable magnetic or magneto-optical storage medium, such as Resistive Random Access Memory (RRAM), Dynamic Random Access Memory (DRAM), Static Random Access Memory (SRAM), Enhanced Dynamic Random Access Memory (EDRAM), High-Bandwidth Memory (HBM), Hybrid Memory Cube (HMC), etc., or any other medium that can be used to store desired information and can be accessed by an application, module, or both. Any such computer storage medium can be part of a device or accessible to or connected to a device. Any application or module described in this invention can be implemented using computer-readable / executable instructions stored or otherwise maintained by such a computer-readable medium.

[0057] The above are all preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Therefore, all equivalent changes made in accordance with the structure, shape and principle of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A method for detecting the coloring performance of masterbatch based on image enhancement, characterized in that, Includes the following steps: S1, obtain the initial grayscale image of the mixed sample, and perform a difference operation on the background light field image extracted from the initial grayscale image to obtain a preprocessed feature map; S2, by analyzing the energy activity of local regions in the preprocessed feature map, construct texture-aware adaptive gain weights and determine the local contrast gain intensity of each pixel; S3, calculate the gradient distribution symmetry of each pixel in the preprocessed feature map in multiple directions to obtain the local isotropic offset index, and decouple the processing flow marks and scattered defects based on the local isotropic offset index. S4. Based on the adaptive gain weight and the local isotropic offset index, the preprocessed feature map is reconstructed to obtain the enhanced detection map after the forced fading of the flow mark interference, and the coloring performance of the masterbatch is determined based on the enhanced detection map.

2. The method for detecting the color performance of masterbatch based on image enhancement according to claim 1, characterized in that, Obtaining the background light field image includes: performing convolution processing on the initial grayscale image using a Gaussian smoothing operator, extracting the low-frequency background trend in the initial grayscale image, and obtaining the background light field image.

3. The method for detecting the color performance of masterbatch based on image enhancement according to claim 1, characterized in that, The adaptive gain weights are calculated using the following formula: In the formula, Representing coordinates Adaptive gain weights at the location, Represented by pixels The standard deviation of pixels in the neighborhood centered on the target. This represents the standard deviation of the total number of pixels in the image. This represents the system reference constant.

4. The method for detecting the color performance of masterbatch based on image enhancement according to claim 3, characterized in that, The local isotropic migration index is calculated using the following formula: In the formula, Indicates the local isotropic shift index. Represents pixels The maximum value among the gradient magnitudes in the horizontal, vertical, and diagonal directions. This represents the average value of the gradient magnitude in all directions. This represents the stability constant.

5. The method for detecting the color performance of masterbatch based on image enhancement according to claim 4, characterized in that, Reconstructing and enhancing the detection map, including: In the formula, This represents the pixel value of the enhanced detection map. This represents the pixel values ​​of the preprocessed feature map. Indicates the penalty order. Representing coordinates Adaptive gain weights at the location, This represents the local isotropic shift index.

6. The method for detecting the color performance of masterbatch based on image enhancement according to claim 5, characterized in that, Determining the coloring performance of masterbatch based on enhanced detection maps includes: determining the defect response coefficient based on the area ratio of the feature points detected in the enhanced detection map to the entire map. By combining the pixel distribution of the enhanced detection map and the defect response coefficient, a production stability evaluation index is calculated to evaluate the coloring performance of the masterbatch.

7. The method for detecting the color performance of masterbatch based on image enhancement according to claim 6, characterized in that, The production stability evaluation index is calculated using the following formula: In the formula, Indicates the production stability evaluation index. Indicates the defect response coefficient. This indicates the total number of pixels in the sample image. This represents the pixel value of the enhanced detection map.

8. The method for detecting the color performance of masterbatch based on image enhancement according to claim 1, characterized in that, Obtaining a preprocessed feature map includes: performing pixel-by-pixel difference between the initial grayscale image and the background light field image to eliminate the illumination gradient caused by uneven ambient lighting.

9. The method for detecting the color performance of masterbatch based on image enhancement according to claim 1, characterized in that, The physical logic of the local isotropic shift index is as follows: if there are processing flow marks in the local area, the maximum directional gradient is significantly greater than the average gradient, and the local isotropic shift index increases; if the local area is a spherical pigment agglomeration point, the gradients in each direction are similar, and the local isotropic shift index tends to zero.

10. A color masterbatch coloring performance detection system based on image enhancement, characterized in that, The method includes a memory and a processor, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the method for detecting the color performance of masterbatch based on image enhancement as described in any one of claims 1-9 is implemented.

Images