Optical crystal vickers hardness and transmittance measurement system based on image processing

Abstract

The application discloses an optical crystal Vickers hardness and transmittance measuring system based on image processing and particularly relates to the field of optical crystal measurement, and comprises an image cooperative acquisition module, an image preprocessing module, a feature extraction calculation module, an adaptive feedback module and a data output module. The optical crystal Vickers hardness and transmittance measuring system based on image processing realizes the physical separation and synchronous acquisition of the indentation image and the transmission light field image through the image cooperative acquisition module, eliminates the conflict between the exposure parameters and the focal length requirements of the same sensor from the hardware architecture, and avoids the parameter adjustment caused by the single sensor mode switching. The image preprocessing module reduces the repeated parameter adjustment caused by the poor image quality and enhances the measurement consistency. The feature extraction calculation module acquires the accurate optical crystal feature set and reduces the invalid parameter adjustment.

Description

Technical Field

[0001] This invention relates to the field of optical crystal measurement technology, and more specifically, to an optical crystal Vickers hardness and transmittance measurement system based on image processing. Background Technology

[0002] With the widespread application of optical crystals in high-end manufacturing, the accurate measurement of their mechanical and optical properties is particularly important. However, existing measurement technologies are insufficient in terms of automation, measurement accuracy, and multi-parameter synchronous detection capabilities, making it difficult to meet the demands of modern industry for efficient and intelligent measurement systems.

[0003] Currently, Vickers hardness measurement usually relies on the traditional indentation method, which involves observing the indentation morphology under a microscope and calculating the hardness value. Crystal transmittance measurement is mostly done using equipment such as spectrophotometers. Although these methods can provide relatively accurate data, they may present calibration difficulties or data fluctuations in complex environments. Furthermore, structural integration and automated control reduce human intervention and improve the testing efficiency and data consistency of single samples.

[0004] However, it still has some drawbacks in practical use. For example, the acquisition of indentation images and transmitted light spots depends on the same image sensor. Since hardness testing requires high-resolution local imaging while transmittance detection requires global light spot analysis, there are conflicts between the sensor's exposure parameters and focal length requirements in different modes, resulting in insufficient adaptive imaging configuration. This further leads to the need to frequently adjust the position and parameters of optical components, causing interruptions in the measurement process and affecting real-time performance and measurement consistency. Especially when continuously testing multiple samples, the repeated positioning error and parameter reset delay significantly reduce the accuracy and efficiency of batch testing. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art, the present invention provides an optical crystal Vickers hardness and transmittance measurement system based on image processing, which solves the problems mentioned in the background art through the following solutions.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] An image processing-based optical crystal Vickers hardness and transmittance measurement system includes:

[0008] Image collaborative acquisition module: used to synchronously distribute the light beams of each measurement point in the optical crystal to the first image sensor and the second image sensor through an optical beam splitter, so as to synchronously acquire the first imaging set of the measurement points, wherein the first imaging set includes at least a high-resolution indentation image acquired by the first image sensor and a transmitted light field distribution image acquired by the second image sensor;

[0009] Image preprocessing module: used to perform preprocessing operations on each image in the first imaging set to obtain a second imaging set with improved signal-to-noise ratio;

[0010] Feature extraction calculation module: used to perform feature extraction operations to obtain the optical crystal feature set corresponding to the second imaging set;

[0011] Adaptive feedback module: used to dynamically and collaboratively adjust the imaging control parameters of the first image sensor and the second image sensor based on the optical crystal feature set; wherein the imaging control parameters include at least a first exposure parameter for controlling the first image sensor and a second exposure parameter for controlling the second image sensor;

[0012] Data output module: used to associate, store and output the optical crystal feature set at the same measurement point.

[0013] Preferably, in the image collaborative acquisition module, the optical beam splitter is between 400 and 700. The splitting ratio error within the wavelength range is less than .

[0014] Preferably, the image preprocessing module acquires a second image set with improved signal-to-noise ratio, specifically including:

[0015] The first preprocessing substream performs noise suppression and edge enhancement on the high-resolution indentation image based on anisotropic diffusion filtering;

[0016] The second preprocessing substream performs light intensity distribution correction and effective region segmentation on the transmitted light field distribution image.

[0017] Preferably, the feature extraction calculation module performs feature extraction operations, specifically including:

[0018] The indentation analysis unit performs sub-pixel-level edge detection on the preprocessed high-resolution indentation image in the second imaging set, fits the geometric features of the indentation contour, and calculates the Vickers hardness value of the measurement point based on the geometric features of the indentation contour.

[0019] The light intensity analysis unit performs light intensity distribution analysis on the preprocessed transmitted light field distribution image in the second imaging set to obtain light intensity distribution characteristics, and calculates the optical transmittance of the measurement point based on the light intensity distribution characteristics.

[0020] Preferably, the feature extraction and calculation module calculates the Vickers hardness value of the measurement point in the indentation analysis unit, specifically including:

[0021] Extract the subpixel coordinates of the indentation edges in the image;

[0022] All extracted sub-pixel edge point sets are fitted into four straight line equations using the least squares method;

[0023] By solving the equations of adjacent lines simultaneously, the sub-pixel coordinates of the four indentation vertices are calculated.

[0024] Based on the vertex coordinates, the physical lengths of the first and second diagonals are calculated using the Euclidean distance formula.

[0025] The physical scale transmitted by the image preprocessing module is invoked to convert the pixel length into physical length;

[0026] The hardness value is automatically calculated based on the Vickers hardness calculation formula.

[0027] Preferably, the feature extraction and calculation module, in the light intensity analysis unit, calculates the optical transmittance of the measurement point, specifically including:

[0028] For all pixels within the segmented effective light spot area, calculate the arithmetic mean of their gray values, which is used as the transmitted light intensity measurement value at the measurement point. ;

[0029] The reference white field value transmitted by the image preprocessing module is called. The standard light intensity value at which the transmittance is 100% is used to calculate the optical transmittance at the measurement point. Specifically, it is expressed as:

[0030] .

[0031] Preferably, the feature extraction calculation module further includes a data verification unit, which executes the following after the indentation analysis unit and the light intensity analysis unit have completed their calculations:

[0032] Check the sum of squared residuals of the four straight lines fitted by the least squares method. If any residual is greater than a preset threshold, mark the Vickers hardness value of the measurement point with a low confidence flag.

[0033] Check whether the calculated diagonal length is within a reasonable physical range based on the test force; if it is outside the range, mark it as invalid data.

[0034] Check whether the transmitted light intensity measurement value is within the linear response range of the sensor; otherwise, mark its transmittance data as suspicious.

[0035] All data flags will be incorporated into the output of the optical crystal feature set.

[0036] Preferably, the adaptive feedback module dynamically and collaboratively adjusts the execution of imaging control parameters of the first image sensor and the second image sensor, specifically including:

[0037] A quality evaluator is used to generate quantitative metrics for evaluating image quality based on the optical crystal feature set or its intermediate data.

[0038] The strategy decision-maker is used to generate parameter adjustment instructions based on the deviation between the quantitative indicator and the preset target range, according to a preset control strategy.

[0039] A parameter mapper is used to convert the parameter adjustment instructions into executable imaging control parameters.

[0040] The technical effects and advantages of this invention are as follows:

[0041] 1. This invention achieves physical separation and synchronous acquisition of indentation images and transmitted light field images through an image collaborative acquisition module and an optical beam splitter. This eliminates the conflict between the same sensor in terms of exposure parameters and focal length requirements from the hardware architecture perspective, and avoids parameter adjustments caused by single sensor mode switching.

[0042] 2. This invention improves the signal-to-noise ratio of high-resolution indentation images through an image preprocessing module, enhances edge sharpness, improves the background uniformity of transmitted light field distribution images, reduces repeated parameter adjustments due to poor image quality, enhances measurement consistency, and provides high-quality input for feature extraction.

[0043] 3. This invention obtains a precise optical crystal feature set through a feature extraction and calculation module, reduces invalid parameter adjustments, further improves the real-time performance of measurements and the reliability of data, and optimizes the pertinence and effectiveness of adaptive adjustment. Attached Figure Description

[0044] Figure 1 This is a block diagram of an image processing-based optical crystal Vickers hardness and transmittance measurement system provided according to an embodiment of this application. Detailed Implementation

[0045] 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 embodiments of the present invention, and not all embodiments. 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.

[0046] The terminology used in the following embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. As used in the specification of this application, the singular expressions “a,” “an,” “the,” “the,” “the,” and “this” are intended to include the plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this application refers to and includes any or all possible combinations of one or more of the listed items.

[0047] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying or suggesting relative importance or implicitly indicating the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature, and in the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more.

[0048] As attached Figure 1 The image processing-based optical crystal Vickers hardness and transmittance measurement system shown includes an image collaborative acquisition module, an image preprocessing module, a feature extraction and calculation module, an adaptive feedback module, and a data output module.

[0049] Specifically, the image collaborative acquisition module is used to synchronously distribute the light beams of each measurement point in the optical crystal to the first image sensor and the second image sensor through an optical beam splitter, so as to synchronously acquire the first imaging set of the measurement points, wherein the first imaging set includes at least a high-resolution indentation image acquired by the first image sensor and a transmitted light field distribution image acquired by the second image sensor.

[0050] It should be noted that the image collaborative acquisition module mainly consists of an optical beam splitter, a first image sensor, a second image sensor, a synchronization control unit, and a data interface. Through innovative optical path design and sensor configuration, it achieves physical separation, synchronous acquisition, and standardized output of the indentation morphology and transmitted light field at the same measurement point.

[0051] In a preferred embodiment, the optical beam splitter is used to achieve physical beam separation. This embodiment employs a cubic prism beam splitter structure, consisting of two right-angle prisms optically bonded together along their hypotenuses. A broadband dielectric beam splitting film with a specific splitting ratio is deposited on the bonded surface. Simultaneously, the optical beam splitter is fixed to a six-dimensional adjustment frame to allow for micrometer-level translation and milliradian-level deflection adjustments, thereby ensuring that the centers of the two beams after splitting coincide with the centers of their respective sensor target surfaces, so as to achieve beam separation within the 400–700 nm range of the first image sensor. The splitting ratio error within the wavelength range is less than The flat spectral response characteristics.

[0052] Furthermore, the image collaborative acquisition module adopts a dual-image sensor architecture, and the original beam from the optical crystal is split into two sensors according to a preset optical path by the optical beam splitter. The first image sensor and the second image sensor are respectively located on two orthogonal outgoing light paths of the optical beam splitter; the first image sensor is a high-resolution CMOS sensor used for local macro imaging of the indentation area, and its pixel size is less than 2.5. It is also equipped with a macro imaging lens group, enabling it to operate at object dimensions smaller than 2. ×2 The second image sensor is a high dynamic range CCD sensor used to acquire the overall intensity distribution of the transmitted light field passing through the sample. It is equipped with a telecentric lens, which enables it to operate in a global field of view covering the entire aperture of the optical crystal.

[0053] In a preferred embodiment, the synchronization control unit is implemented by an FPGA, which generates a unified TTL level hardware trigger signal and distributes it directly to the external trigger interfaces of the first image sensor and the second image sensor via a coaxial cable. The FPGA integrates delay compensation logic, which adjusts the transmission time of the trigger signal at the microsecond level based on the pre-calibrated inherent electrical delay difference between the two sensors to ensure that the exposure start time difference between the two sensors is less than 1 millisecond.

[0054] In a preferred embodiment, when the data interface outputs, it assigns the same globally unique timestamp and measurement point ID to the synchronously acquired high-resolution indentation image and transmitted light field distribution image, and packages them into a data packet, namely the first imaging atlas, which is then transmitted to the downstream image preprocessing module through the data interface.

[0055] Specifically, the image preprocessing module is used to perform preprocessing operations on each image in the first imaging set to obtain a second imaging set with improved signal-to-noise ratio.

[0056] It should be noted that the image preprocessing module is triggered by the image collaborative acquisition module. It reads the first imaging atlas data package from the pre-allocated shared memory via direct memory access. The second imaging atlas includes at least the indentation image with noise suppression and enhanced edges, the spot image with corrected background uniformity and precise segmentation of the effective area, the calibrated physical scale, the coefficients used for correction, and the reference white field value. The physical scale is calibrated by capturing a standard micrometer-level scale image and calculating the number of pixels occupied by a known physical distance. The reference white field value is the average light intensity value calculated by the second preprocessing sub-stream after measuring the standard sample.

[0057] In one possible implementation, to generate a high-quality second imaging atlas, a dual-pipeline processing architecture is employed, which optimizes different characteristics of the first imaging atlas: a first preprocessing sub-stream performs noise suppression and edge enhancement on the high-resolution indentation image based on anisotropic diffusion filtering; and a second preprocessing sub-stream performs light intensity distribution correction and effective region segmentation on the transmitted light field distribution image.

[0058] Furthermore, the processing path of the first preprocessing substream includes at least processing the original indentation image using an anisotropic diffusion filtering algorithm based on partial differential equations. The diffusion coefficient is adaptively adjusted according to the local gradient of the image. Strong smoothing is performed in smooth regions with small gradients to filter out noise, while diffusion is suppressed in edge regions with large gradients. This effectively smooths the noise inside the image while preserving and sharpening the true edge information of the indentation to the greatest extent. After noise suppression, the residual noise is further smoothed to locate the positions in the image where the grayscale changes drastically, and the edges are identified by the zero crossover points of the second derivative.

[0059] Furthermore, the processing path of the second preprocessing substream includes at least performing flat field correction to compensate for various systematic, fixed-pattern noise and inhomogeneities, thereby obtaining a corrected image with a realistic light intensity distribution and a uniform background; further, it fills in the tiny holes inside the light spot that may be caused by noise and smooths its boundaries to extract the effective area where the light spot is located.

[0060] In a preferred embodiment, in the first preprocessed substream, the noise suppression algorithm based on anisotropic diffusion filtering has the following discrete iterative calculation formula:

[0061] ,

[0062] in, Represented as the number of iterations, Represented as the first In the next iteration, the grayscale value of the current pixel in the high-resolution indentation image. Represented as iteration step size, , These are respectively represented as the current pixel and its north and south neighbors at the th... Gray-scale difference in the next iteration.

[0063] It should be noted that the number of iterations... The range of values ​​is 1≤ ≤ , The iteration step size is dynamically set by the adaptive feedback module based on real-time noise estimation, with a typical value of 3 to 5. Used to control the allowable grayscale change range in a single iteration, with a value range of [value range missing]. In this embodiment, the iteration step size is... The adaptive feedback module adaptively adjusts the signal-to-noise ratio estimate, with a default value defined as 0.2; the discrete iterative calculation formula includes the current pixel and its four neighbors (north, east, south, and west) in the [number of iterations]. The gray-level difference in the next iteration, where The edge stopping function is defined as follows:

[0064] ,

[0065] in, The indentation contrast threshold is represented by the initial value obtained by calculating the gradient magnitude histogram of the high-resolution indentation image and taking the top 70% quantile.

[0066] In a preferred embodiment, in the second preprocessing substream, based on pre-acquired dark field images... Peace Field Image A flat-field correction operation is performed on the transmitted light field distribution image, and the calculation formula is as follows:

[0067] ,

[0068] in, Represented as pixel values ​​of the transmitted light field distribution image after flat-field correction. This is represented by the pixel grayscale value of the transmitted light field distribution image directly output by the second image sensor at the current measurement point; after performing flat-field correction, the maximum inter-class variance method is used to... The system automatically calculates the optimal grayscale threshold to distinguish the effective light spot area from the background noise and dark edges in order to perform effective region segmentation. The image is binarized into foreground and background. A 3×3 pixel circular structuring element is used to fill the holes in the binarized image, and the connected region with the largest area is selected as the effective light spot area.

[0069] It should be noted that the dark field image required in the light intensity distribution correction is... That is, the average grayscale image output by the second image sensor under completely dark conditions, obtained by acquiring at least 10 frames of images under completely dark conditions and averaging them; the required flat-field image That is, the average grayscale image output by the second image sensor when capturing a standard integrating sphere panel with uniform illumination.

[0070] In this embodiment, the physical scale, the reference white field value, and the default parameters in the discrete iterative calculation formula are all stored in the non-volatile memory within the adaptive feedback module after initialization and calibration. Before each measurement process begins, the adaptive feedback module sends the data to the corresponding configuration register via the system bus for its use. After processing by the first preprocessing substream, the signal-to-noise ratio of the high-resolution indentation image should be improved by at least 10 dB, and the edge sharpness should be improved by at least 50%. After processing by the second preprocessing substream, the background uniformity of the spot image should be better than 2%.

[0071] Specifically, the feature extraction calculation module is used to perform feature extraction operations to obtain the optical crystal feature set corresponding to the second imaging set.

[0072] It should be noted that the feature extraction calculation module is triggered by the interrupt signal of the image preprocessing module. The interrupt signal includes a pointer to a shared memory address, which stores the packaged data packet of the second imaging set. The optical crystal feature set includes at least the measurement point ID, the physical length of the first diagonal, the physical length of the second diagonal, the Vickers hardness value, the average light intensity value, and the transmittance value that are consistent in the first imaging set.

[0073] In one possible implementation, the feature extraction operation includes: an indentation analysis unit, which performs sub-pixel-level edge detection on the preprocessed high-resolution indentation image in the second imaging set, fits the geometric features of the indentation contour, and calculates the Vickers hardness value of the measurement point based on the geometric features of the indentation contour; and a light intensity analysis unit, which performs light intensity distribution analysis on the preprocessed transmitted light field distribution image in the second imaging set to obtain light intensity distribution features, and calculates the optical transmittance of the measurement point based on the light intensity distribution features.

[0074] Furthermore, in the indentation analysis unit, sub-pixel level edge detection is performed to extract the sub-pixel coordinate positions of the indentation edges in the image. Specifically, it is expressed as: , ,in, , These are respectively represented as the unit normal vectors at the edge points. Represented as integer pixel coordinates corresponding to the measurement point. This is represented as the sub-pixel offset from the extreme point along the normal direction; all extracted sub-pixel edge point sets are fitted into four straight line equations using the least squares method; the sub-pixel coordinates of the four indentation vertices are calculated by simultaneously solving the equations of adjacent straight lines; the physical lengths of the first and second diagonals are calculated using the Euclidean distance formula based on the vertex coordinates; the pixel lengths are converted into physical lengths by calling the physical scale transmitted by the image preprocessing module; and the hardness value is automatically calculated according to the Vickers hardness calculation formula.

[0075] It should be noted that the normal vector The acquisition is based on an approximate solution using the preprocessed high-resolution indentation image from the second imaging set, specifically expressed as follows:

[0076] ,

[0077] in, , These are represented as the gradient vectors in the x and y directions of the preprocessed high-resolution indentation image; the sub-pixel offset along the normal direction to the extreme point is... The calculation formula is specifically expressed as follows:

[0078] ,

[0079] in, , , They are , The second derivative; the gradient vector ( , The second derivative is obtained directly by the feature extraction calculation module calling the built-in Gaussian differential filter and convolving it with the preprocessed high-resolution indentation image.

[0080] Furthermore, in the light intensity analysis unit, for all pixels within the segmented effective light spot area, the arithmetic mean of their grayscale values ​​is calculated, and this value is used as the transmitted light intensity measurement value of the measurement point. That is, the light intensity distribution characteristics; then, the reference white field value transmitted by the image preprocessing module is invoked. The standard light intensity value at which the transmittance is 100% is used to calculate the optical transmittance at the measurement point. Specifically, it is expressed as:

[0081] .

[0082] In a preferred embodiment, the feature extraction calculation module further includes a data verification unit, which, after the indentation analysis unit and the light intensity analysis unit have completed their calculations, performs the following: checks the sum of squared residuals of the four straight lines fitted by the least squares method; if any residual is greater than a preset threshold, the Vickers hardness value of the measurement point is marked with a low confidence flag; checks whether the calculated diagonal length is within a reasonable physical range based on the test force; if it is outside the range, it is marked as invalid data; checks whether the transmitted light intensity measurement value is within the linear response range of the sensor; if not, its transmittance data is marked with a suspicious flag; all data flags are incorporated into the output of the optical crystal feature set.

[0083] It should be noted that the preset threshold of the sum of squared residuals of the indentation fitting is obtained by statistically analyzing a large number of known qualified indentation image samples after training and learning; the threshold of the standard deviation of the transmitted light intensity measurement value is set according to the noise model of the second image sensor and the experimental calibration of the photoelectric response linearity; and all the thresholds are stored in the adaptive feedback module and are loaded when the corresponding module is started.

[0084] Specifically, the adaptive feedback module is used to dynamically and collaboratively adjust the imaging control parameters of the first image sensor and the second image sensor based on the optical crystal feature set; wherein the imaging control parameters include at least a first exposure parameter for controlling the first image sensor and a second exposure parameter for controlling the second image sensor.

[0085] It should be noted that the working cycle of the adaptive feedback module is synchronized with the measurement cycle; after the feature extraction calculation is completed at each measurement point, it is triggered by a signal issued by the feature extraction calculation module; the generated imaging control parameters will be sent out and take effect through the drive interface of the image collaborative acquisition module before the acquisition of the next measurement point begins; the adaptive feedback module also receives a new sample start signal. When a new sample start signal is received, the historical cumulative value of the internal integrator of the strategy decision-maker is cleared to zero, and the imaging control parameters generated for the next measurement point are reset to a preset safe initial value to avoid parameter interference between different samples.

[0086] In one possible implementation, dynamically and collaboratively adjusting the imaging control parameters of the first image sensor and the second image sensor includes: a quality evaluator, used to generate a quantitative index for evaluating image quality based on the optical crystal feature set or its intermediate data; a strategy decision-maker, used to generate parameter adjustment instructions according to a preset control strategy based on the deviation of the quantitative index from a preset target range; and a parameter mapper, used to convert the parameter adjustment instructions into imaging control parameters executable by the first image sensor or the second image sensor.

[0087] It should be noted that the quality evaluator calculates the quantitative index for the indentation image path as the image contrast calculated using the Michelson contrast formula, specifically expressed as follows:

[0088] ,

[0089] in, This is expressed as a quantitative index of indentation image contrast. , These are the maximum and minimum gray values ​​within the effective indentation area of ​​the preprocessed indentation image, respectively, and the effective indentation area is the region within the indentation contour segmented by the feature extraction calculation module; for the light spot image path, the calculated quantization index is the average gray value obtained by calculating the arithmetic mean of all pixels within the effective light spot area segmented after preprocessing.

[0090] Furthermore, the strategy decision-maker uses a PI control algorithm to generate parameter adjustment instructions for the first image sensor. Its control error is the difference between the preset target contrast value and the currently measured contrast. The output is the adjustment amount of the exposure time of the first image sensor. In this embodiment, the proportional coefficient for the contrast adjustment of the first image sensor... The initial value is 0.5 to 2.0, and the integral coefficient is... The initial value is 0.1 to 0.5; for the second image sensor, its control error is the difference between the preset target average gray value and the current average gray value, and the output is the adjustment amount of the sensor gain of the second image sensor; in this embodiment, the gray-scale adjustment ratio of the second image sensor is... The initial value is 0.3 to 1.5, and the integral coefficient is... The initial value is 0.05 to 0.3; the proportionality coefficient and integral coefficient Preset based on historical data.

[0091] Furthermore, the parameter mapper maintains a lookup table recording the minimum and maximum values ​​of all adjustable parameters of the first and second image sensors. The parameter mapper receives the output of the strategy decision-maker to calculate a new exposure time value and determine whether the new exposure time value is within the allowable range. If it exceeds the allowable range, it is limited to the closest limit value. The same operation is performed on the gain parameter. In this embodiment, the minimum exposure time is not less than the sensor's inherent delay time, and the maximum value does not exceed 50% of the measurement period. The minimum gain parameter is 1x, and the maximum value does not exceed 80% of the sensor's linear gain range.

[0092] Specifically, the data output module is used to associate, store, and output the optical crystal feature set at the same measurement point.

[0093] Secondly: The accompanying drawings of the embodiments disclosed in this invention only involve the structures involved in the embodiments disclosed in this invention. Other structures can refer to the general design. In the absence of conflict, the same embodiment and different embodiments of this invention can be combined with each other.

[0094] In conclusion, 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 principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An optical crystal Vickers hardness and transmittance measurement system based on image processing, characterized in that, The method comprises the following steps: An image cooperative acquisition module is configured to synchronously distribute light beams of each measurement point in an optical crystal to a first image sensor and a second image sensor through an optical splitter to synchronously acquire a first imaging set of the measurement point, wherein the first imaging set comprises at least a high-resolution indentation image acquired by the first image sensor and a transmission light field distribution image acquired by the second image sensor; An image preprocessing module is configured to perform a preprocessing operation on each image in the first imaging set to acquire a second imaging set with improved signal-to-noise ratio; A feature extraction calculation module is configured to perform a feature extraction operation to acquire an optical crystal feature set corresponding to the second imaging set; An adaptive feedback module is configured to dynamically and cooperatively adjust imaging control parameters of the first image sensor and the second image sensor based on the optical crystal feature set, wherein the imaging control parameters comprise at least a first exposure parameter for controlling the first image sensor and a second exposure parameter for controlling the second image sensor; A data output module is configured to store and output the optical crystal feature set of the same measurement point in association; The adaptive feedback module dynamically and cooperatively adjusts the imaging control parameters of the first image sensor and the second image sensor, specifically comprising: A quality evaluator is configured to generate a quantitative index for evaluating image quality based on the optical crystal feature set or intermediate data thereof; A strategy decision maker is configured to generate a parameter adjustment instruction according to a preset control strategy based on a deviation of the quantitative index from a preset target range; A parameter mapper is configured to convert the parameter adjustment instruction into executable imaging control parameters; The strategy decision maker generates a parameter adjustment instruction using a PI control algorithm. For the first image sensor, the control error is the difference between the preset target contrast value and the currently measured contrast value, and the output is the adjustment amount of the exposure time of the first image sensor. For the second image sensor, the control error is the difference between the preset target average gray value and the current average gray value, and the output is the adjustment amount of the sensor gain of the second image sensor.

2. The image processing based optical crystalline Vickers hardness and transmittance measurement system of claim 1, wherein: The image cooperative acquisition module, the optical splitter has a splitting ratio error less than 0.5% in the 400-700 nm wavelength range. The image cooperative acquisition module, the optical splitter has a splitting ratio error less than 0.5% in the 400-700 nm wavelength range. The image cooperative acquisition module, the optical splitter has a splitting ratio error less than 0.5 3. The image processing based optical crystalline Vickers hardness and transmittance measurement system of claim 1, wherein: The image preprocessing module acquires the second imaging set with improved signal-to-noise ratio, specifically comprising: A first preprocessing sub-flow performs noise suppression and edge enhancement based on anisotropic diffusion filtering on the high-resolution indentation image; A second preprocessing sub-flow performs light intensity distribution correction and effective area segmentation on the transmission light field distribution image.

4. The image processing based optical crystalline Vickers hardness and transmittance measurement system of claim 1, wherein: The feature extraction calculation module performs the feature extraction operation, specifically comprising: An indentation analysis unit performs sub-pixel level edge detection on the preprocessed high-resolution indentation image in the second imaging set, fits the geometric features of the indentation profile, and calculates the Vickers hardness value of the measurement point based on the geometric features of the indentation profile; A light intensity analysis unit performs light intensity distribution analysis on the preprocessed transmission light field distribution image in the second imaging set to acquire light intensity distribution features and calculate the optical transmittance of the measurement point according to the light intensity distribution features.

5. The image processing based optical crystalline Vickers hardness and transmittance measurement system of claim 4, wherein: The feature extraction and calculation module calculates the Vickers hardness value of the measuring point in the indentation analysis unit, and specifically includes: extracting the sub-pixel coordinate position of the indentation edge in the image; fitting all the extracted sub-pixel edge point sets into four straight line equations respectively using the least square method; calculating the sub-pixel coordinates of the four indentation vertexes by simultaneously solving the equations of adjacent straight lines; calculating the first diagonal physical length and the second diagonal physical length according to the vertex coordinates using the Euclidean distance formula; converting the pixel length into physical length by calling the physical scale transmitted by the image preprocessing module; automatically calculating the hardness value according to the Vickers hardness calculation formula.

6. The image processing based optical crystalline Vickers hardness and transmittance measurement system of claim 4, wherein: The feature extraction and calculation module calculates the optical transmittance of the measuring point in the light intensity analysis unit, and specifically includes: For all pixels in the segmented effective light spot region, the arithmetic mean of the gray value is calculated as the transmission light intensity measurement value of the measurement point ; The reference white field value transmitted by the image preprocessing module is called The standard light intensity value when the transmittance is 100% is called the standard light intensity value, and the optical transmittance of the measurement point is calculated , which is specifically represented as 7. The image processing based optical crystalline Vickers hardness and transmittance measurement system of claim 4, wherein: The feature extraction and calculation module further includes a data verification unit, and after the calculation of the indentation analysis unit and the light intensity analysis unit is completed, it performs: checking the residual sum of squares of the four straight lines fitted by the least square method, if any residual is greater than a preset threshold, marking the Vickers hardness value of the measuring point with a low confidence flag; checking whether the calculated diagonal length is within a reasonable physical range based on the test force, if it is out of range, marking it as invalid data; checking whether the transmitted light intensity measurement value is within the linear response interval of the sensor, if not, marking the transmittance data with a suspicious flag; all data flag bits will be included in the output of the optical crystal feature set.

Images