A method and system for processing image data collected from a steel bowl reference surface

Abstract

This application relates to the field of image processing technology, specifically to a method and system for processing image data acquired from a steel bowl reference surface. The method includes: calculating the probability that any pixel in the original image of the steel bowl reference surface belongs to a highly reflective region; dividing the steel bowl reference surface image into regions based on the probability that a pixel belongs to a highly reflective region; calculating the corresponding clipping value based on the grayscale distribution of any region; enhancing the original image of the steel bowl reference surface using a contrast-limited histogram equalization algorithm based on the clipping values ​​of each region, and then stitching the images together to output the final acquired image of the steel bowl reference surface. This application can achieve adaptive contrast optimization for different regions in the image.

Description

Technical Field

[0001] This application relates to the field of image processing technology, and in particular to a method and system for processing image data acquired from a steel bowl reference surface. Background Technology

[0002] As an important green mode of transportation, the manufacturing precision of key components of bicycles directly affects the performance, safety, and lifespan of the entire vehicle. Among them, the cup arm, as the core component connecting the frame and the fork, is crucial to the smoothness and stability of the steering system due to the geometric precision of its reference surfaces (such as flatness, perpendicularity, and surface roughness).

[0003] In recent years, with the development of machine vision and image processing technologies, non-contact optical inspection has gradually been introduced into the quality control of precision mechanical parts. High-resolution industrial cameras are used to acquire images of the reference surface of a steel bowl, and combined with image processing algorithms, the geometric parameters of the part can be evaluated.

[0004] In the manufacturing process of bicycle helmets, the reference surface is typically a polished metal surface (such as stainless steel or alloy steel). When illuminated by a light source, specular reflection of light on the reference surface creates strong highlight areas, leading to camera sensor saturation and overexposed areas in the image. Simultaneously, nonlinear scattering of the reflected light path introduces geometric distortion, blurring the edges of the reference surface and further compromising the accuracy of the helmet's reference surface evaluation. Existing image processing methods, such as histogram equalization algorithms, rely on enhancing the entire image, neglecting the influence of local reflection gradients.

[0005] A cup bearing reference surface image refers to the reference surface of the cup bearing (also known as the bearing housing or head cup) component in a bicycle head tube assembly, captured using optical equipment. Since the cup bearing is a critical precision component connecting the bicycle fork to the frame, its reference surface is typically a polished metal surface exhibiting typical specular reflection characteristics. During visual inspection and image acquisition, light produces strong specular reflections on its surface, resulting in excessively high pixel values ​​in certain areas of the image, forming "overexposed areas." These areas not only mask defects such as scratches and indentations but also blur edge contours, further affecting subsequent geometric parameter evaluation. Summary of the Invention

[0006] To address the issue of neglecting the influence of local reflection gradients when enhancing the global image in the aforementioned image processing, this application provides a method and system for processing image data acquired from a steel bowl reference surface. By calculating the probability that any pixel belongs to a highly reflective area, the image is divided into regions and the corresponding shear value is calculated, thereby performing local enhancement to complete the acquisition optimization of the steel bowl reference surface image.

[0007] In a first aspect, this application provides a method for processing image data acquired from a steel bowl reference surface, employing the following technical solution:

[0008] A method for processing image data acquired from a steel bowl reference surface, comprising the following steps:

[0009] For any pixel in the original image of the steel bowl reference surface, calculate the probability that the pixel belongs to a highly reflective area; divide the steel bowl reference surface image into regions based on the probability that the pixel belongs to a highly reflective area; calculate the corresponding shear value based on the grayscale distribution of any region; enhance the original image of the steel bowl reference surface using the limited contrast histogram equalization algorithm based on the shear value of each region, and stitch them together to output the final acquired image of the steel bowl reference surface;

[0010] The method for calculating the probability that a pixel belongs to a highly reflective area is as follows: obtain the gray value of each pixel in the original image, divide any pixel into a target pixel sequence, and calculate the probability that the pixel belongs to a highly reflective area based on the gray value difference between the pixel and all pixels in the corresponding target pixel sequence.

[0011] The step of calculating the corresponding clipping value based on the grayscale distribution of any region includes: using the difference between 1 and the mean grayscale value of all pixels in each region to obtain the clipping parameter corresponding to each region, pre-setting an empirical clipping value and a minimum clipping value, and taking the maximum value of the product of the clipping parameter and the empirical clipping value and the minimum clipping value as the clipping value corresponding to that region.

[0012] Furthermore, by arranging the light source perpendicular to the steel bowl reference plane, an industrial camera is used to acquire the original image of the steel bowl reference plane perpendicular to the steel bowl reference plane.

[0013] Furthermore, the step of dividing any pixel into a target pixel sequence includes: dividing a window of a preset size with any pixel as the center, and recording all pixels within the window as the target pixel sequence of that pixel.

[0014] Furthermore, the step of dividing the steel bowl reference surface image into regions based on the probability that a pixel belongs to a highly reflective region includes: obtaining the absolute value of the difference in the probability that any two pixels belong to a highly reflective region, inputting the absolute value of the difference into the superpixel segmentation algorithm, and completing the superpixel segmentation.

[0015] Furthermore, the range of the shear value is a closed interval between the minimum shear value and the empirical shear value.

[0016] Furthermore, the step of calculating the corresponding shear value based on the grayscale distribution of any region includes: matching a smaller shear value to highly reflective regions and a larger shear value to low-reflective regions to improve the contrast of different grayscale values.

[0017] Furthermore, it also includes: interpolating the splicing positions of each area to achieve smoothing of the splicing positions and splicing them according to the original spatial positions.

[0018] Secondly, this application provides a system for processing image data acquired from a steel bowl reference surface, employing the following technical solution:

[0019] A steel bowl reference surface acquisition image data processing system includes: a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a steel bowl reference surface acquisition image data processing method according to any one of the above claims is implemented.

[0020] This application has the following technical effects:

[0021] This application addresses the problem that the steel bowl's reference surface is made of polished metal, which is prone to strong specular reflection during image acquisition, leading to local overexposure. By calculating the probability that each pixel belongs to a highly reflective area, it can determine the different reflective areas in the image, avoiding the defects of traditional global enhancement causing detail loss or saturation distortion in highly reflective areas.

[0022] This application divides the image into multiple regions based on reflective properties and calculates the shear value independently for each region. Then, it uses contrast-limited histogram equalization to perform local enhancement, thereby achieving adaptive contrast optimization for different regions in the image. Attached Figure Description

[0023] The above and other objects, features, and advantages of the present invention will become readily apparent from the following detailed description of exemplary embodiments, accompanied by the accompanying drawings. In the drawings, several embodiments of the invention are illustrated by way of example and not limitation, and like or corresponding reference numerals denote like or corresponding parts, wherein:

[0024] Figure 1 This is a flowchart of a method for processing image data acquired from a steel bowl reference surface, provided in an embodiment of this application. Detailed Implementation

[0025] 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.

[0026] This application discloses a method for processing image data acquired from a steel bowl reference surface, referring to... Figure 1 The process includes steps S1 to S4, as detailed below:

[0027] S1: Calculate the probability that any pixel in the original image of the steel bowl reference surface belongs to a highly reflective area.

[0028] Specifically, in this embodiment, the light source is arranged perpendicular to the reference surface of the steel bowl, and an industrial camera is used perpendicular to the reference surface of the steel bowl to obtain the original image of the reference surface of the steel bowl.

[0029] It is important to note that because the steel bowl's reference surface is a polished metal surface, when the light source position is fixed, the incident angle of the light source varies at different positions on the steel bowl's reference surface, resulting in significant differences in reflection intensity in different areas. Traditional histogram equalization algorithms enhance the steel bowl's reference surface image based on the global grayscale distribution. However, since overexposed areas exist on the steel bowl's reference surface, directly enhancing the entire image would saturate the grayscale values ​​of these overexposed areas, making it impossible to recover details. Therefore, this invention analyzes the probability that any pixel belongs to a highly reflective area, facilitating further analysis based on the likelihood of a pixel belonging to a highly reflective area.

[0030] Specifically, the method for calculating the probability that a pixel belongs to a highly reflective area is as follows: obtain the gray value of each pixel in the original image, divide any pixel into a target pixel sequence, and calculate the probability that the pixel belongs to a highly reflective area based on the gray value difference between the pixel and all pixels in the corresponding target pixel sequence.

[0031] Specifically, dividing any pixel into a target pixel sequence includes: dividing a window of a preset size with any pixel as the center, and recording all pixels within the window as the target pixel sequence for that pixel.

[0032] Specifically, a window of a preset size is divided with any pixel in the image as the center. In this embodiment, the preset size is... A window is defined, and all pixels within the window are recorded as the target pixel sequence for that pixel. The probability that a pixel belongs to a highly reflective area is calculated based on the grayscale difference between that pixel and all pixels in its target pixel sequence, using the following expression:

[0033]

[0034] in, Indicates the first The possibility that a pixel belongs to a highly reflective area; Indicates the first The grayscale mean of all pixels in the target pixel sequence is the value of the pixel. The larger the value, the higher the probability that the pixel belongs to a highly reflective area. Indicates the first The grayscale value of each pixel; Indicates the first The target pixel sequence of the nth pixel The grayscale value of each pixel; This represents the standard normalization function; This indicates the difference in grayscale between the pixel and its target pixel sequence. The larger the value, the greater the difference in grayscale, indicating that the pixel may be a noise pixel. The lower the confidence that the target pixel sequence belongs to a highly reflective area. The greater the difference, the less likely the pixel is to belong to a highly reflective area. When the difference approaches 0, the probability that the pixel belongs to a highly reflective area approaches 1.

[0035] S2: Divide the steel bowl reference surface image into regions based on the probability that a pixel belongs to a highly reflective area.

[0036] Specifically, dividing the steel bowl reference surface image into regions based on the probability that a pixel belongs to a highly reflective area includes: obtaining the absolute value of the difference in the probability that any two pixels belong to a highly reflective area, inputting the absolute value of the difference into the superpixel segmentation algorithm, and completing the superpixel segmentation.

[0037] Specifically, in this embodiment, any pixel is used as a seed point, and the image is divided into blocks using a superpixel segmentation algorithm to obtain different regions.

[0038] It's important to note that traditional superpixel segmentation algorithms merge similar pixels into a single superpixel block by calculating the color and spatial distances between them. In this embodiment, however, merging is based on the probability that pixels belong to highly reflective regions. Specifically, the absolute value of the difference between any two pixels belonging to highly reflective regions is obtained, and this absolute value replaces the color distance used in traditional superpixel segmentation algorithms. This completes the superpixel segmentation, resulting in multiple superpixel blocks, i.e., different regions.

[0039] S3: Calculate the corresponding shear value based on the grayscale distribution of any region.

[0040] Specifically, calculating the corresponding clipping value based on the grayscale distribution of any region includes: obtaining the clipping parameter corresponding to each region by using the difference between 1 and the mean grayscale value of all pixels in each region after normalization; pre-setting an empirical clipping value and a minimum clipping value; taking the maximum value of the product of the clipping parameter and the empirical clipping value and the minimum clipping value as the clipping value corresponding to that region; and the range of the clipping value is a closed interval between the minimum clipping value and the empirical clipping value.

[0041] Specifically, based on the reflection state at different locations on the steel bowl's reference surface, the surface is divided into different regions. For any given region, a suitable clipping value is calculated by acquiring its reflection state. For highly reflective regions, since a large number of pixels are concentrated at high grayscale values, a smaller clipping value is needed to avoid over-enhancing these values ​​and to highlight details in the slightly lower grayscale areas. For low-reflective regions, a larger clipping value is needed to improve the contrast between different grayscale values, making observation easier.

[0042] Specifically, for any region, the clipping value corresponding to that region is calculated based on the grayscale values ​​of the pixels in that region. The calculation expression is as follows:

[0043]

[0044] in, Indicates the first Shear values ​​for each region; Indicates the first The average grayscale value of all pixels in a region; This indicates the maximum pixel value of the current image; Indicates will Perform normalization processing; This represents the preset empirical shear value; 0.05 is the preset minimum shear value. This represents the maximum value function.

[0045] in, This represents a preset empirical cropping value, which can be adjusted by the implementer according to the specific implementation situation. For any region, if the grayscale mean of the region is higher, the cropping value for that region will be smaller; conversely, if the grayscale mean is lower, the cropping value for that region will be larger. To avoid the cropping value being too small, which would lead to distortion of the enhanced image, a [presumably a specific parameter] is introduced. Let the shear value of any region be taken as and The maximum of the two, that is, the limit of the shear value to be within Within the range.

[0046] S4: Based on the shear values ​​corresponding to each region, the original image of the steel bowl reference surface is enhanced using a limited contrast histogram equalization algorithm, and then stitched together to output the final acquired image of the steel bowl reference surface.

[0047] Specifically, for any region, based on the shear value calculated in step S3, enhancement processing is performed using a limited contrast histogram equalization algorithm to obtain the enhanced image of that region. To avoid stitching artifacts at the boundaries of different regions due to enhancement of images from different regions, interpolation processing is performed on the stitching positions. In this embodiment, bilinear interpolation is used to process the stitching positions to make them smooth. After processing all regions, the enhancement results of all regions are stitched together according to their original spatial positions to obtain the final acquired image of the steel bowl reference surface.

[0048] This application also discloses a steel bowl reference surface acquisition image data processing system, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, a steel bowl reference surface acquisition image data processing method according to this application is implemented.

[0049] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

[0050] In this application, the aforementioned memory can be any tangible medium containing or storing a program that can be used by or in conjunction 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 (DRAM), dynamic random access memory (DRAM), static random access memory (SRAM), or any other medium that can be used to store required information and can be accessed by an application program, module, or both. Any such computer storage medium can be part of a device or accessible to or connected to a device.

[0051] The above are all preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.

Claims

1. A method for processing image data acquired from a steel bowl reference surface, characterized in that the steps include: include: Calculate the probability that any pixel in the original image of the steel bowl's reference surface belongs to a highly reflective area; The steel bowl reference surface image is divided into regions based on the probability that a pixel belongs to a highly reflective area; the corresponding shearing value is calculated based on the grayscale distribution of any region; the original image of the steel bowl reference surface is enhanced using a limited contrast histogram equalization algorithm based on the shearing value of each region, and then stitched together to output the final acquired image of the steel bowl reference surface; The method for calculating the probability that a pixel belongs to a highly reflective region includes: obtaining the grayscale value of each pixel in the original image, dividing any pixel into a target pixel sequence, and calculating the probability that the pixel belongs to a highly reflective region based on the grayscale difference between the pixel and all pixels in the corresponding target pixel sequence; the grayscale difference is inversely proportional to the probability that the corresponding pixel belongs to a highly reflective region. The step of calculating the corresponding shearing value based on the grayscale distribution of any region includes: using the difference between 1 and the mean grayscale value of all pixels in each region to obtain the shearing parameter corresponding to each region, pre-setting an empirical shearing value and a minimum shearing value, and taking the maximum value of the product of the shearing parameter and the empirical shearing value and the minimum shearing value as the shearing value corresponding to that region. Specifically, high-reflectivity areas are matched with smaller shear values ​​to suppress overexposure and restore details, while low-reflectivity areas are matched with larger shear values ​​to enhance contrast; bilinear interpolation is performed at the stitching positions of each area.

2. The method for processing image data acquired from a steel bowl reference surface according to claim 1, characterized in that, By arranging the light source perpendicular to the reference surface of the steel bowl, an industrial camera is used to acquire the original image of the steel bowl reference surface perpendicular to the reference surface.

3. The method for processing image data acquired from a steel bowl reference surface according to claim 1, characterized in that, The step of dividing any pixel into a target pixel sequence includes: dividing a window of a preset size with any pixel as the center, and recording all pixels within the window as the target pixel sequence of that pixel.

4. The method for processing image data acquired from a steel bowl reference surface according to claim 1, characterized in that, The step of dividing the steel bowl reference surface image into regions based on the probability that a pixel belongs to a highly reflective area includes: obtaining the absolute value of the difference in the probability that any two pixels belong to a highly reflective area, inputting the absolute value of the difference into the superpixel segmentation algorithm, and completing the superpixel segmentation.

5. The method for processing image data acquired from a steel bowl reference surface according to claim 1, characterized in that, The range of the shear value is a closed interval between the minimum shear value and the empirical shear value.

6. A system for acquiring and processing image data from a steel bowl reference surface, characterized in that, include: The processor and memory, wherein the memory stores computer program instructions that, when executed by the processor, implement a method for processing image data acquired from a steel bowl reference surface according to any one of claims 1-5.

Images