Titanium alloy casting water immersion detection method and system based on industrial vision

Abstract

This invention discloses a water immersion inspection method and system for titanium alloy castings based on industrial vision, relating to the fields of non-destructive testing and industrial machine vision. The method includes acquiring industrial vision data of the titanium alloy casting in its current clamping state; registering the industrial vision data with the casting's CAD model to establish a multi-coordinate system transformation relationship; dividing the casting into inspection zones based on the registration results, geometric features, wall thickness information, and an acoustic process library; generating water immersion ultrasonic scanning paths and inspection parameters based on the zones; performing inspection along the scanning path and collecting ultrasonic echo data; identifying defects and mapping them to the model coordinate system; fusing visual features, zone information, and ultrasonic defect features to output defect judgment results and generate a report. This invention achieves fully automated, high-precision, and traceable quality inspection of complex titanium alloy castings through industrial vision-driven adaptive water immersion ultrasonic inspection, significantly improving inspection coverage, defect identification rate, and inspection efficiency.

Description

Technical Field

[0001] This invention relates to the fields of non-destructive testing and industrial machine vision technology, specifically a water immersion inspection method and system for titanium alloy castings based on industrial vision. Background Technology

[0002] Titanium alloy castings are widely used in key fields such as aerospace, shipbuilding and energy and chemical industry due to their high specific strength and excellent corrosion resistance. However, titanium alloys are prone to various internal and surface defects such as porosity, shrinkage porosity, inclusions and microcracks during casting and subsequent processing, which seriously threaten the service safety of components. Water immersion ultrasonic testing is often used to detect internal defects in castings due to its advantages of stable coupling and high signal-to-noise ratio. However, when facing titanium alloy castings with complex curved surfaces, variable wall thickness, stiffeners or cavity structures, traditional water immersion ultrasonic testing has the following problems: (1) It relies on manual clamping and path teaching, which is time-consuming, labor-intensive and inconsistent, making it difficult to meet the needs of batch testing.

[0003] (2) It is difficult to ensure that the probe maintains the best normal incidence and stable water distance on complex curved surfaces, resulting in signal deviation and missed detection.

[0004] (3) The detection parameters (such as frequency and gain) are fixed and cannot adapt to the differences in acoustic characteristics of different wall thicknesses and different grain noise levels.

[0005] (4) Visual appearance inspection and ultrasonic internal inspection data are separated, lacking a multi-source information fusion judgment mechanism, resulting in a high false alarm rate.

[0006] (5) The testing process and data are difficult to trace throughout the entire process, which is not conducive to quality management and process improvement.

[0007] Therefore, there is an urgent need for an intelligent and highly reliable water immersion inspection method that can adapt to the geometric features of complex castings. Summary of the Invention

[0008] The purpose of this invention is to provide a water immersion inspection method and system for titanium alloy castings based on industrial vision, so as to solve the problems mentioned in the background art.

[0009] To achieve the above objectives, the present invention provides the following technical solution.

[0010] A water immersion inspection method for titanium alloy castings based on industrial vision includes the following steps: S1: acquiring industrial vision data of the titanium alloy casting to be inspected in its current clamping state, wherein the industrial vision data includes image data and / or three-dimensional point cloud data.

[0011] S2: Register the industrial vision data with the casting CAD model and establish the transformation relationship between the vision coordinate system, motion coordinate system, ultrasonic coordinate system and casting model coordinate system.

[0012] S3: Based on the registration results, local geometric features of the casting, wall thickness information, and the pre-built acoustic process library for titanium alloy castings, the area to be inspected is divided into inspection zones.

[0013] S4: Generate a water immersion ultrasonic scanning path and corresponding detection parameters based on the detection zone. The detection parameters include at least one of probe posture, water distance, focal length, frequency, gain, gate, scanning step distance, and scanning speed.

[0014] S5: Control the water immersion ultrasonic testing module to inspect the titanium alloy casting along the scanning path, collect and record ultrasonic echo data and the corresponding sampling point coordinates.

[0015] S6: Perform defect identification on the ultrasonic echo data and map the defect identification results to the coordinate system of the casting model.

[0016] S7: Integrates visual features, detection zone information, and ultrasonic defect features to output defect type, 3D location, depth, equivalent size, and confidence level, and generates an inspection report.

[0017] Preferably, the detection zones include at least two of the following: a gentle, uniform thickness zone, a variable wall thickness zone, a stiffener intersection zone, a cavity edge zone, a gating and riser residual zone, a high-risk zone for hot spots, a no-scan zone, and a rescan zone.

[0018] Preferably, the acoustic process library for titanium alloy castings includes the correspondence between material grade, sound velocity, grain noise level, recommended probe frequency, recommended gain, recommended gate, recommended step size, and defect risk level.

[0019] Preferably, in step S4, a local encrypted scanning path, a multi-angle rescanning path, or multi-frequency detection parameters are generated for the high-risk detection zone.

[0020] Preferably, the defect identification of ultrasonic echo data in step S6 includes at least one analysis of the ultrasonic echo data, including amplitude characteristics, arrival time characteristics, frequency domain characteristics, connected domain characteristics, or spatial neighborhood characteristics.

[0021] As a preferred option, a rescan confirmation is triggered when the defect identification result is located in a high-risk detection zone or meets a preset spatial proximity relationship with an abnormal area on the visual surface.

[0022] The present invention also provides a water immersion inspection system for titanium alloy castings based on industrial vision, comprising: a water immersion inspection tank.

[0023] An industrial vision module is used to acquire images and / or 3D point clouds of titanium alloy castings to be inspected.

[0024] The casting support and positioning module is used to fix the titanium alloy casting to be inspected.

[0025] A multi-axis motion module is used to drive the movement of the water immersion ultrasonic testing module.

[0026] The water immersion ultrasonic testing module is used to transmit and receive ultrasonic waves to acquire ultrasonic echo data.

[0027] The calibration and coordinate registration module is used to establish the transformation relationships between the visual coordinate system, motion coordinate system, ultrasonic coordinate system, and casting model coordinate system.

[0028] The detection area partitioning module is used to partition the area to be inspected based on the registration results, local geometric features, wall thickness information, and the acoustic process library for titanium alloy castings.

[0029] The scanning path and parameter generation module is used to generate a water immersion ultrasonic scanning path and corresponding detection parameters based on the detection zone.

[0030] The defect fusion and judgment module is used to map the ultrasonic test results to a three-dimensional model and fuse visual features, test zone information and ultrasonic defect features to output the defect judgment result.

[0031] The report output module is used to generate test reports.

[0032] Preferably, the industrial vision module includes at least one of a structured light camera, a binocular camera, an RGB-D camera, a line laser profilometer, or a lidar.

[0033] Preferably, the water immersion ultrasonic testing module includes at least one of a water immersion focusing probe, a water immersion phased array probe, a frequency conversion probe, a multi-probe array, or a matrix array probe.

[0034] In addition, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described industrial vision-based water immersion detection method for titanium alloy castings.

[0035] Compared with existing technologies, the beneficial effects of this invention are: 1. Reduced reliance on manual labor, improved efficiency and consistency. By recognizing the actual posture of the casting through industrial vision and automatically correcting the path, no precision tooling or tedious manual teaching is required, significantly shortening the inspection preparation time.

[0036] 2. Achieve high-quality full-coverage detection of complex curved surfaces. The scanning path is adaptively generated based on the local curvature and normal of visual recognition, ensuring optimal probe incidence, effectively avoiding missed scans, and improving detection coverage and signal quality.

[0037] 3. Adaptive parameter detection improves defect recognition rate. Combining acoustic process library and geometric partitioning, the system automatically configures optimal detection parameters (such as frequency and step size) for different regions, and automatically executes a rescan strategy for high-risk areas, effectively suppressing coarse grain noise and distinguishing real defects.

[0038] 4. Multi-source information fusion for more reliable judgment. Spatial fusion judgment is performed on internal ultrasonic defects, visual surface anomalies, and process risk zoning, outputting quantitative results with confidence, reducing the misjudgment rate of single methods.

[0039] 5. End-to-end data traceability and quality record building. The system automatically saves original data from the entire process, including visual, ultrasonic, path, and parameter data, forming an auditable data package that provides complete evidence for quality traceability, process improvement, and AI model training. Attached Figure Description

[0040] Figure 1 This is a schematic diagram of a water immersion inspection system for titanium alloy castings based on industrial vision.

[0041] Figure 2 This is a schematic diagram of the structure of an ultrasonic probe.

[0042] Figure 3 This is a flowchart of a water immersion inspection method for titanium alloy castings based on industrial vision.

[0043] Figure 4 A flowchart for the retesting of complex titanium alloy castings.

[0044] In the diagram: 1. Control cabinet; 2. Multi-axis motion module; 3. Ultrasonic probe; 4. Casting bearing and positioning module; 5. Water immersion tank; 6. Industrial vision module. Detailed Implementation

[0045] The technical solution of this application will be further described in detail below with reference to specific embodiments.

[0046] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain this application, and should not be construed as limiting this application.

[0047] Please see Figure 1In one embodiment of the present invention, a water immersion inspection system for titanium alloy castings based on industrial vision includes: a water immersion tank 5, an industrial vision module 6, a casting support and positioning module 4, a multi-axis motion module 2, a water immersion ultrasonic inspection module (including an ultrasonic probe 3 and related acquisition cards), and a control cabinet 1 (containing a built-in control module, calibration and coordinate registration module, inspection area partitioning module, scanning path and parameter generation module, defect fusion judgment module, and report output module, etc.). The industrial vision module 6 is mounted on the ultrasonic probe 3, located above or to the side of the water immersion tank 5, and is used to acquire the three-dimensional point cloud or image of the casting through a transparent window before or during water immersion. The casting support and positioning module 4 is used to fix the titanium alloy casting and can cooperate with the multi-axis motion module 2 to achieve precise positioning. The structure of the ultrasonic probe 3 is as follows. Figure 2 As shown, it is used to transmit and receive ultrasonic waves.

[0048] The industrial vision module 6 includes at least one of a structured light camera, a binocular camera, an RGB-D camera, a line laser profilometer, or a lidar. The water immersion ultrasonic testing module includes at least one of a water immersion focusing probe, a water immersion phased array probe, a frequency conversion probe, a multi-probe array, or a matrix array probe. The report output module is used to output a traceable inspection report containing visual images, point cloud data, scanning path, inspection parameters, raw ultrasonic data, C-scan images, three-dimensional coordinates of defects, rescan records, and judgment conclusions.

[0049] Please see Figure 3 The overall process of the detection method of the present invention is as follows: Step S1, Acquire industrial vision data: Acquire industrial vision data of the titanium alloy casting to be inspected in its current clamping state. The industrial vision data includes image data and / or three-dimensional point cloud data. Specifically, after system initialization, the titanium alloy casting to be tested is placed on the casting support and positioning module 4 in the water immersion tank 5. The industrial vision module 6 is started, scans the workpiece, and acquires its complete three-dimensional point cloud data in its current clamping state. For complex curved surface workpieces such as titanium alloy castings, the ICP algorithm is used to complete fine registration, and normal constraints or scale factors are introduced to eliminate the sound velocity refraction distortion caused by the water immersion environment. The specific process is as follows: Extract the casting point cloud data (obtained by three-dimensional model or laser scanning) and the C-scan trajectory point cloud as the source point set; use ICP iteration to minimize the Euclidean distance between the two point sets, and combine the casting CAD model to constrain the registration degrees of freedom; finally, generate a joint matrix containing position information and parameter labels (such as the gain value of the corresponding scanning point), thereby realizing the unified expression of three-dimensional space and detection parameters.

[0050] Step S2, Coordinate Registration: The industrial vision data is registered with the casting CAD model to establish the transformation relationship between the vision coordinate system, motion coordinate system, ultrasonic coordinate system, and casting model coordinate system. Specifically, the calibration and coordinate registration module of the control module in control cabinet 1 performs ICP fine registration between the acquired point cloud data and the pre-stored casting CAD model to establish the accurate transformation relationship between the vision coordinate system, motion coordinate system (controlling the movement of ultrasonic probe 3), ultrasonic C-scan coordinate system, and CAD model coordinate system, thereby obtaining actual clamping deviation, local curvature, local normal, thickness estimation, and accessibility information.

[0051] Step S3, Inspection Zoning: Based on the registration results, local geometric features of the casting, wall thickness information, and the pre-built acoustic process library for titanium alloy castings, the area to be inspected is divided into inspection zones. Specifically, the inspection zone zoning module divides the surface and internal areas of the casting into several inspection zones based on the registration results, CAD nominal wall thickness, casting process information, and the pre-built acoustic process library for titanium alloys. These inspection zones include at least two of the following: smooth, uniform thickness zones, variable wall thickness zones, rib intersection zones, cavity edge zones, gating and riser residue zones, high-risk hotspot zones, no-scan zones, and rescan zones. Each zone is bound to sound velocity, recommended frequency, probe type, focal length, water distance, gate range, gain, step size, and scanning speed. The acoustic process library for titanium alloy castings includes the correspondence between material grade, sound velocity, grain noise level, recommended probe frequency, recommended gain, recommended gate, recommended step size, and defect risk level.

[0052] The construction method of the acoustic process library for titanium alloy castings is as follows: it is constructed based on two main lines: standard test block verification and physical simulation calibration. The specific contents include: 1. Sound velocity anisotropy model: by rotating the incident probe, the longitudinal and transverse wave sound velocities are measured under different grain orientations to form a direction-sound velocity mapping table with an accuracy of ±10m / s.

[0053] 2. Attenuation coefficient library: For typical casting defects, including porosity, shrinkage porosity, and microcracks, the scattering attenuation caused by the microstructure at different frequencies of 5-15MHz is measured and classified into a thickness-frequency comparison table.

[0054] 3. Interface compensation parameters: Record the curve of transmission loss at the water-titanium alloy interface as a function of incident angle for real-time gain compensation.

[0055] 4. Dynamic gate template: Based on the normal of the casting CAD surface and the probe scanning path, the distance amplitude correction curve DAC and the gate start / width set are generated in advance.

[0056] The process of constructing the process library includes: collecting representative samples, including good samples and artificially defective samples; performing full immersion automatic scanning to acquire A-mode waveforms; extracting acoustic feature extrema; forming a hierarchical library table through interpolation and clustering, wherein the clustering uses K-means to partition the attenuation characteristics; and finally embedding the hierarchical library table into the detection system.

[0057] Step S4: Generate scanning path and parameters: Generate a water immersion ultrasonic scanning path and corresponding detection parameters based on the detection zones. The detection parameters include at least one of probe posture, water distance, focal length, frequency, gain, gate, scanning step distance, and scanning speed. Specifically, the scanning path and parameter generation module automatically generates a water immersion ultrasonic scanning path based on the zone results and the accessibility of the multi-axis motion module 2. This path ensures that the center line of the ultrasonic probe 3's sound beam is incident along the normal direction of the local curved surface of the casting as much as possible, and controls the water distance within the optimal range of the probe's focal length. At the same time, optimal detection parameters are bound to each sampling point or each zone. For example, a small step distance (e.g., 0.3 mm) and multi-angle rescanning are set for high-risk areas of hot spots; a large step distance (e.g., 1.0 mm) and conventional C-scanning are set for flat areas. As a preferred implementation, locally encrypted scanning paths, multi-angle rescanning paths, or multi-frequency detection parameters are generated for high-risk detection zones.

[0058] Each sampling point in the path is associated with visual coordinates, CAD coordinates, probe orientation, expected wall thickness, and detection parameters. A strategy combining isoparametric sampling and dynamic programming is employed. During trajectory planning in the mathematical space after registration, the parameter domain of the casting surface is uniformly discretized to generate initial grid path points, and sampling is refined based on curvature changes. Then, at each path point, the optimal detection parameters are inversely derived using a physical simulation model. This typically involves an optimization process that minimizes the objective function (e.g., a weighted sum of signal-to-noise ratio and detection efficiency), which can be solved using gradient descent or genetic algorithms. The final output is a six-dimensional vector sequence containing spatial coordinates (x, y, z) and corresponding detection parameters (e.g., voltage, focal length), directly used to drive the immersion scanning equipment to perform adaptive detection.

[0059] Step S5: Perform water immersion ultrasonic testing: Control the water immersion ultrasonic testing module to inspect the titanium alloy casting along the scanning path, collect and record ultrasonic echo data and corresponding sampling point coordinates. Specifically, the system controls the multi-axis motion module 2 to drive the ultrasonic probe 3 to automatically scan the casting along the generated path. The water immersion ultrasonic testing module collects A-scan and C-scan data for each scanning point in real time and records the corresponding coordinates, probe posture, water distance, gain, and other parameters. For areas with high coarse-grained noise, a low-frequency probe can be automatically switched or multi-angle rescanning can be performed.

[0060] Step S6, Defect Identification and Coordinate Mapping: Defect identification is performed on the ultrasonic echo data, and the defect identification results are mapped to the coordinate system of the casting model. Specifically, the defect fusion judgment module identifies defects in the ultrasonic echo data, including analyzing at least one of the following: amplitude characteristics, arrival time characteristics, frequency domain characteristics, connected domain characteristics, or spatial neighborhood characteristics, to obtain potential internal defects. These potential defects are then fused with visual appearance anomalies, zoning risk levels, casting process risks, and historical samples to output the defect type, three-dimensional coordinates, depth, equivalent size, confidence level, and re-inspection recommendations. Defect types may include shrinkage cavities, porosity, gas porosity, inclusions, cracks, cold shut-related anomalies, and near-surface anomalies.

[0061] Specific integration methods.

[0062] (1) Joint eigenvectors ,in The characteristic dimension of the ultrasonic echo is d u Specifically, this includes: peak amplitude in the time domain, echo variance, center frequency offset, bottom echo loss rate, waveform rise time, and waveform fall time. The feature vector for appearance anomalies has dimension d. v Specifically, this includes: the area, aspect ratio, mean gray value, gray variance, and edge gradient intensity of the visually abnormal region; The risk level scalar is assigned a value based on the detection scalar type to which the defect belongs. Different types of scalars correspond to different prior risk levels.

[0063] (2) Abnormal appearance characteristics After the industrial vision module 6 acquires images of the titanium alloy casting surface, the images are extracted using an image segmentation algorithm. Specifically, this includes the geometric and textural features of visible abnormal areas such as surface scratches, pits, cold shuts, and gating remnants. If a defect point is spatially adjacent to a visually abnormal area (within a preset threshold, such as 2mm), its abnormal appearance feature vector is assigned the measured value; otherwise, all components of the abnormal appearance feature vector are assigned a value of 0.

[0064] (3) Feature normalization: Since the dimensions and value ranges of each feature are different, normalization is required. The Min-Max normalization method is used to scale each feature component to the [0,1] interval.

[0065] (4) Clustering: The K-Means clustering algorithm is used to group the normalized feature vectors. The specific steps are as follows: First, initialize the cluster centers. Based on prior knowledge, the number of clusters K=4, corresponding to: cluster 0 for normal regions, cluster 1 for crack defects, cluster 2 for porosity / looseness defects, and cluster 3 for dense porosity defects. The initial cluster centers are determined by randomly selecting 4 sample points from the training samples.

[0066] The second step is to assign samples. For each normalized feature vector, calculate its Euclidean distance to each cluster center and assign it to the cluster with the closest distance.

[0067] The third step is to update the cluster centers. The new cluster center for each cluster is the mean of all samples within that cluster.

[0068] Fourth, repeat steps two and three until the cluster centers no longer change or the preset maximum number of iterations is reached.

[0069] Step S7, Fusion Judgment and Report Output: The system fuses visual features, detection zone information, and ultrasonic defect features to output the defect type, 3D location, depth, equivalent size, and confidence level, and generates a detection report. Specifically, the system fuses ultrasonic defect features, the zone risk level from step S3, and surface scratches, dents, and other appearance anomalies identified by industrial vision module 6 to comprehensively judge each defect, outputting its type (e.g., porosity, crack), 3D coordinates, depth, equivalent size, and confidence level. Finally, the report output module automatically generates a complete detection report containing visual images, point clouds, scanning paths, C-scan images, and judgment criteria.

[0070] Specifically, when the defect identification result is located in a high-risk detection zone or meets a preset spatial proximity relationship with an abnormal area on the visual surface, a rescan confirmation is triggered.

[0071] Please see Figure 4 After the initial scan is completed (initial scan data input), the defect fusion judgment module will trigger a rescan based on preset judgment rules. The specific process is as follows: First, the system analyzes the initial scan data in real time to determine whether the rescan trigger conditions are met. These conditions include: the peak value of the ultrasonic echo exceeds the set evaluation line (condition 1: peak value threshold); the bottom echo loss rate exceeds 40% and the gradient change of adjacent points exceeds 20% (condition 2: bottom echo loss rate); abnormal jumps occur in the acoustic time or frequency characteristics (condition 3: characteristic jump).

[0072] Once any condition is met, the system triggers a rescan for confirmation and marks the area (triggering rescan confirmation, marking suspicious points). Subsequently, the scan path and parameter generation module performs local path planning for the suspicious area, reducing the step size and optimizing the scan range (local path planning: step size value, scan range). During the rescan, different detection strategies can be invoked for composite scanning, such as alternating between conventional frequency probes and high-frequency probes (composite scan data acquisition: conventional probe, high-frequency probe). Richer feature vectors (such as peak value, frequency band energy, waveform coefficients, etc.) are extracted from the rescan data, and clustering algorithms such as K-Means are used for region calculation (feature vector extraction, clustering region calculation) to distinguish defect types (defect judgment: type A: crack; type B: porosity / looseness; type C: dense porosity). Finally, the final judgment result after rescan confirmation is output and archived (result output and archiving).

[0073] Example 1: Water immersion C-scan inspection of complex thin-walled titanium alloy castings.

[0074] 1. Equipment configuration: The water immersion detection tank has a volume of 800mm×600mm×500mm, and the water temperature is controlled at 20±2℃. It is equipped with circulation filtration, degassing and liquid level monitoring; the industrial vision module 6 adopts a structured light camera with a point cloud sampling interval of no more than 0.5mm; the multi-axis motion module 2 adopts a three-axis gantry platform with A / B swing axis; the ultrasonic module adopts a 5MHz water immersion focusing probe with a sampling rate of no less than 100MS / s.

[0075] 2. Operation process: The titanium alloy casting is placed on a flexible support frame. The system collects point clouds and registers them with the CAD model to obtain the clamping deviation matrix. Based on the CAD wall thickness and the curvature of the point cloud, the system divides the area into a smooth, uniform thickness zone, a variable wall thickness zone, and a cavity edge zone. The system sets a step distance of 1.0 mm for the smooth zone, a step distance of 0.3 to 0.5 mm for the cavity edge zone, and a dynamic gate for the variable wall thickness zone.

[0076] 3. Inspection Process: The probe is controlled to perform a water immersion C-scan along the local normal, and the A-scan, probe orientation, water distance, and coordinates of each sampling point are recorded in real time. After the inspection is completed, a C-scan image and a 3D defect image are generated.

[0077] 4. Judgment: If the echo amplitude of a certain area exceeds the corresponding partition threshold and forms a connected region in adjacent sampling points, it is marked as a potential defect; if the potential defect is located in the high-risk area of ​​the hot spot or the area adjacent to the visual surface anomaly, the confidence level is increased and a local rescan is triggered.

[0078] Example 2: Phased array water immersion testing of titanium alloy castings with stiffeners and cavities.

[0079] 1. Equipment configuration: Water immersion phased array probe frequency 2.25~5MHz, array element number 32~64, equipped with phased array acquisition instrument; Industrial vision module 6 adopts a combination of line laser profilometer and area array camera; Multi-axis motion module 2 adopts a six-axis robotic arm plus water immersion probe end clamp.

[0080] 2. Zoning Strategy: The industrial vision module 6 identifies the rib intersection area, cavity edge area, abrupt change in inner and outer contour areas, and no-scan areas. The system generates the fan-scan angle, focal length, and beam incidence angle based on local thickness and curvature. Multi-angle rescanning is used in the rib intersection area, avoidance distances are set in the cavity edge area, and no path is generated in the no-scan area.

[0081] 3. Parameter examples: 5MHz frequency and 0.8mm step size in normal areas; 2.25MHz frequency and 0.5mm step size in thick-walled or coarse-grained noise areas; scanning speed of cavity edge areas is less than 40mm / s; high-risk areas are repeatedly sampled at least twice.

[0082] 4. Output: The system outputs the three-dimensional coordinates, depth, length, equivalent quantity, confidence level, and rescan conclusion of the defect, and overlays the results onto the CAD model.

[0083] Example 3: Variable frequency / multi-probe adaptive detection.

[0084] 1. Equipment Configuration and Strategy: For the thin-walled, medium-thick-walled, and thick-walled coarse-grained regions of the same titanium alloy casting, the system pre-establishes a material acoustic process library. A 7.5MHz high-frequency probe is used in the thin-walled region to improve near-surface resolution; a 5MHz probe is used in the medium-thick-walled region; and a 2.25MHz probe is used in the thick-walled coarse-grained region to improve penetration capability.

[0085] 2. Probe Selection and Switching: The industrial vision module 6 first estimates the wall thickness range based on the point cloud and CAD data, and then the parameter recommendation module selects the probe and detection parameters. After probe switching, the system automatically retrieves the sensitivity calibration results of the corresponding reference block.

[0086] 3. Multi-frequency comparison judgment: By comparing the multi-frequency echo responses of the same defect area, the grain noise, geometric boundary echo and real defect echo can be distinguished.

[0087] Comparative test cases

[0088] Comparative Example A

[0089] The traditional water immersion ultrasonic testing method is used to inspect titanium alloy castings. The specific steps are as follows: The operator manually clamps and fixes the titanium alloy casting in the water immersion tank 5; a fixed rectangular scanning area is set as the scanning path based on experience; the parameters of the water immersion ultrasonic testing module are set to a fixed 5MHz frequency, fixed gain, and fixed step distance; the ultrasonic probe 3 is controlled to perform C-scan inspection on the titanium alloy casting along the fixed rectangular scanning path, and ultrasonic echo data is collected; after the inspection is completed, the operator manually interprets the C-scan image to identify defects.

[0090] Example B

[0091] The method described in Embodiments 1, 2, or 3 of this invention is used to inspect titanium alloy castings of the same batch. Taking Embodiment 1 as an example, the specific steps are as follows: the industrial vision module 6 collects point cloud data of the casting in its current clamping state; the calibration and coordinate registration module in the control cabinet 1 registers the point cloud data with the casting CAD model to obtain the actual clamping deviation and local geometric features; the detection area partitioning module automatically divides the detection area according to the registration result, wall thickness information, and the pre-built acoustic process library of titanium alloy castings; the scanning path and parameter generation module automatically generates the water immersion ultrasonic scanning path and corresponding detection parameters according to the partitioning result; the multi-axis motion module 2 controls the ultrasonic probe 3 to perform adaptive detection along the generated scanning path and collect ultrasonic echo data; the defect fusion judgment module identifies defects in the ultrasonic echo data and maps the identification results to the casting model coordinate system; the visual features, detection partition information, and ultrasonic defect features are fused to output the defect judgment result; and the report output module generates an inspection report.

[0092] Evaluation indicators and expected results.

[0093] Recommended evaluation indicators include: clamping / teaching time, detection coverage, positioning error, detection time per unit area, defect detection rate, false alarm rate, and consistency of rescanning.

[0094] The expected result is that, compared with Comparative Example A, Embodiment B of the present invention can significantly reduce teaching time, improve the detection coverage of complex edge / cavity regions, and reduce misjudgments caused by water distance and surface normal deviation.

[0095] The above are merely preferred embodiments of the present invention. It should be noted that those skilled in the art can make several modifications and improvements without departing from the concept of the present invention, and these should also be considered within the scope of protection of the present invention. These will not affect the effectiveness of the implementation of the present invention or the practicality of the patent.

Claims

1. A water immersion inspection method for titanium alloy castings based on industrial vision, characterized in that, Includes the following steps: S1: Acquire industrial vision data of the titanium alloy casting to be inspected in the current clamping state, wherein the industrial vision data includes image data and / or three-dimensional point cloud data; S2: Register the industrial vision data with the casting CAD model and establish the transformation relationship between the vision coordinate system, motion coordinate system, ultrasonic coordinate system and casting model coordinate system; S3: Based on the registration results, local geometric features of the casting, wall thickness information, and the pre-built acoustic process library for titanium alloy castings, the area to be inspected is divided into inspection zones; S4: Generate a water immersion ultrasonic scanning path and corresponding detection parameters according to the detection zone. The detection parameters include at least one of probe posture, water distance, focal length, frequency, gain, gate, scanning step distance, and scanning speed. S5: Control the water immersion ultrasonic testing module to inspect the titanium alloy casting along the scanning path, collect and record ultrasonic echo data and the corresponding sampling point coordinates; S6: Perform defect identification on the ultrasonic echo data and map the defect identification results to the coordinate system of the casting model; S7: Integrates visual features, detection zone information, and ultrasonic defect features to output defect type, 3D location, depth, equivalent size, and confidence level, and generates an inspection report.

2. The water immersion inspection method for titanium alloy castings based on industrial vision according to claim 1, characterized in that, The detection zones include at least two of the following: a smooth, uniform thickness zone; a variable wall thickness zone; a ribbed intersection zone; a cavity edge zone; a gating and riser residue zone; a high-risk zone for hot spots; a no-scan zone; and a rescan zone.

3. The water immersion inspection method for titanium alloy castings based on industrial vision according to claim 1, characterized in that, The acoustic process library for titanium alloy castings includes the correspondence between material grade, sound velocity, grain noise level, recommended probe frequency, recommended gain, recommended gate, recommended step size, and defect risk level.

4. The water immersion inspection method for titanium alloy castings based on industrial vision according to claim 1, characterized in that, In step S4, local encrypted scanning paths, multi-angle rescanning paths, or multi-frequency detection parameters are generated for high-risk detection zones.

5. The water immersion inspection method for titanium alloy castings based on industrial vision according to claim 1, characterized in that, The defect identification of ultrasonic echo data in step S6 includes analyzing the ultrasonic echo data using at least one of amplitude characteristics, arrival time characteristics, frequency domain characteristics, connected domain characteristics, or spatial neighborhood characteristics.

6. The water immersion inspection method for titanium alloy castings based on industrial vision according to claim 1, characterized in that, When the defect identification result is located in a high-risk detection zone or meets the preset spatial proximity relationship with an abnormal area on the visual surface, a rescan confirmation is triggered.

7. A water immersion inspection system for titanium alloy castings based on industrial vision, used to implement the water immersion inspection method for titanium alloy castings based on industrial vision as described in any one of claims 1 to 6, characterized in that, include: Water immersion test tank; The industrial vision module is used to acquire images and / or 3D point clouds of the titanium alloy castings to be inspected. Casting support and positioning module, used to fix the titanium alloy casting to be inspected; A multi-axis motion module is used to drive the movement of the water immersion ultrasonic testing module; The water immersion ultrasonic testing module is used to transmit and receive ultrasonic waves to acquire ultrasonic echo data; The calibration and coordinate registration module is used to establish the transformation relationships between the visual coordinate system, the motion coordinate system, the ultrasonic coordinate system, and the casting model coordinate system. The detection area partitioning module is used to partition the area to be inspected based on the registration results, local geometric features, wall thickness information, and the acoustic process library for titanium alloy castings. The scanning path and parameter generation module is used to generate a water immersion ultrasonic scanning path and corresponding detection parameters based on the detection zone. The defect fusion and judgment module is used to map the ultrasonic inspection results to a three-dimensional model, and fuse visual features, inspection zoning information and ultrasonic defect features to output the defect judgment result. The report output module is used to generate test reports.

8. The water immersion inspection system for titanium alloy castings based on industrial vision according to claim 7, characterized in that, The industrial vision module includes at least one of a structured light camera, a binocular camera, an RGB-D camera, a line laser profilometer, or a lidar.

9. The water immersion inspection system for titanium alloy castings based on industrial vision according to claim 7, characterized in that, The water immersion ultrasonic testing module includes at least one of the following: water immersion focusing probe, water immersion phased array probe, frequency conversion probe, multi-probe array, or matrix array probe.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the water immersion inspection method for titanium alloy castings based on industrial vision as described in any one of claims 1 to 6.

Images