Industrial CT scanning measurement

Since its inception, CT technology has been first used in medical diagnosis and material testing. With the advancement of CT technology and the improvement of measurement accuracy, its application scope has expanded to the field of industrial product measurement and gradually emerged.

For traditional contact or optical non-contact coordinate measuring equipment, non-destructive measurement of the internal structural dimensions of objects is one of the difficulties in production practice. Industrial CT technology provides an effective way to solve such problems.

In the field of industrial measurement, industrial CT technology can non-destructive measure the overall dimensions of the internal and external structures of products: a single industrial CT scan can simultaneously complete the process of product size measurement and material defect assessment; The industrial CT measurement process is not affected by the surface condition of the workpiece (roughness, color, curvature); The high-density point cloud obtained from industrial CT measurement can be used for overall evaluation of the internal and external dimensions of the scanned workpiece body model; Industrial CT technology can measure objects during assembly, and can be used for component failure analysis, tracking quality control and tolerance assessment in the manufacturing process of industrial products.

1、 Principles of Industrial CT Imaging

The imaging process of industrial CT includes: the X-ray source generates X-rays and penetrates the tested sample, and the sample undergoes attenuation due to absorption or scattering of the rays. The amount of attenuation is determined by the thickness and composition of the irradiated sample; After attenuation, the rays are incident on the detector to form a two-dimensional grayscale projection image; The detector captures two-dimensional projection images from different angles; Obtain a three-dimensional voxel model of the sample by reconstructing multiple continuous CT images of the sample from the projected image; The analysis and visualization process of industrial CT data is completed through threshold segmentation and edge detection of 3D volume data.

In summary, a complete industrial CT scanning and data processing process includes: projection acquisition, data reconstruction, edge detection, and data analysis.

Starting from the imaging process and basic components of industrial CT, we will classify and analyze the factors that may affect the performance of industrial CT, and summarize the possible measures to improve the measurement accuracy of industrial CT.

2、 Analysis of factors affecting the measurement accuracy of industrial CT

There are many factors that affect the measurement accuracy of industrial CT, which can be summarized as: system hardware related (ray source, motion system, detector), software and data processing related (data reconstruction, threshold carving, wheel magic extraction, data calibration), measured object (geometric structure, material), experimental environment temperature conditions, and operator.

2.1 Ray source

The influencing factors related to the radiation source come from the equipment itself, such as the target material, radiation spectrum, stability, etc., and from the equipment operator on the other hand. The key technical indicators of the radiation source include voltage, current, and focal size.

Usually, operators choose the voltage and current conditions used in experiments within a certain range based on their own experience, which leads to subjectivity in measurement results and measurement not being carried out under optimal conditions. Generally speaking, the higher the voltage, the stronger the radiation penetration ability; The higher the current, the greater the intensity of the radiation. Doubling the current only affects the intensity of the radiation, while voltage doubling affects both the energy and intensity of the radiation.

Another important influencing factor related to radiation sources is the focal size of the radiation source. Based on the focal size of the radiation source, it can be divided into nanofocal (S1pm), microfocal (1-200pm), and conventional focal (2200pm) radiation sources. The smaller the focal size, the sharper the edges of the imaging image, and the larger the focal size of the ray source, the more blurry the imaging image may be due to the penumbra effect. At the same magnification, the smaller the focal size, the sharper the imaging effect. However, the local heat concentration caused by the small focal point can cause the target material to overheat or even be broken down, which limits the energy of the micro focal point X-ray source to a lower range. Currently, the ray energy of the micro focal point X-ray machine is generally below 225kV.

The positional relationship between the radiation source, sample, and detector has a significant impact on the measurement accuracy of industrial CT. The closer the sample position is to the radiation source, the greater the magnification, and correspondingly, more pixels are used on the detector, which theoretically improves spatial resolution. However, at the same time, the penumbra effect cancels out the blurry edges of the imaging image. The reconstruction of cone beam industrial CT is extremely sensitive to horizontal offset between the radiation source, sample, and detector, which can be corrected through the "fine string method".

2.2 Factors related to software and data processing

The image processing of industrial CT data includes two main processes:

(1) Collect projection images and reconstruct 3D voxel models from 2D projection data:

(2) The 3D model undergoes edge detection and thresholding segmentation to complete the subsequent measurement process.

At present, the reconstruction of industrial CT data is mostly carried out using the classical FDK method, and for image edge detection and structural classification, threshold based methods are often used. The main influencing factors come from ray hardening and ray scattering. If these artifacts are not properly corrected, they can lead to a decrease in the reliability of the measurement data.

3、 Ways to improve the measurement accuracy of industrial CT

Previously, we briefly summarized the main factors that affect the performance of industrial CT. In this section, we will provide corresponding improvement measures from both the system hardware and software processing aspects to improve the measurement accuracy of industrial CT in response to these factors.

3.1 Improving the Hardware Performance of Industrial CT Systems

The improvement of hardware performance mainly depends on the continuous improvement of the system by device manufacturers. For equipment operators, it is necessary to pay attention to the indicator parameters of the radiation source and detector equipped in the industrial CT system during the experimental process, such as the probe size, signal-to-noise ratio, dynamic range of the detector: the focus size, maximum voltage, maximum current and power of the radiation source, etc.; the motion accuracy and motion mode (step or continuous) of the turntable system. Other hardware improvement measures, such as using granite bases, air bearings, and servo motors as system components, are also taken to ensure high motion accuracy and stability of industrial CT systems.

3.2 Improve industrial CT software and data post-processing capabilities

Post processing of industrial CT data refers to the software processing of reconstructed 3D volume data, which is transformed from grayscale image data to point cloud data. In this process, the main source of error is the selection of boundary thresholds; 2. Determination of calibration scale.

After industrial CT reconstruction, a three-dimensional volume model of the object is obtained. Before subsequent measurement of this data, it is necessary to first select appropriate threshold values to segment the boundaries between the material and air or different materials, which is known as threshold segmentation. The traditional threshold segmentation algorithm uses the IS050% method to determine the material edge, which selects the middle position between the air and the material peak on the image grayscale histogram as the material edge. However, this method is easily affected by image quality and has a large error. Therefore, by improving the edge detection algorithm and using an edge detection algorithm based on actual surfaces, it has been proven that the accuracy of edge detection can indeed be improved by searching for pixel changes in the normal direction of the image.

Previously, we introduced two main factors that have an adverse impact on the reconstruction process: ray hardening and scattered radiation: reducing the effect of ray hardening can be adjusted by placing a pre filter plate during data acquisition, or improved by subsequent hardening correction algorithms. Using polynomial fitting method to correct hardening artifacts, the corrected hardening artifacts are effectively controlled.

3.3 Optimizing the selection of industrial CT scanning parameters

The selection and setting of parameters directly affect the imaging quality and detection results of X-ray during industrial CT scanning. In the actual X-ray testing process, it usually takes a long time from experimental preparation to obtaining experimental results. When studying the impact of a certain parameter on imaging quality, it is often necessary to repeatedly adjust the parameters, and the entire adjustment process is time-consuming and labor-intensive. Especially when it comes to testing products with complex and bulky structures. Meanwhile, operators often choose experimental parameter combinations based on their own experience, resulting in subjective and non optimal experimental results. X-ray simulation tools can be used to simulate the impact of parameter changes on industrial CT imaging, simulate the real detection process, adjust parameter settings to achieve the best detection effect, and thus obtain optimized detection plans, greatly reducing the detection cycle.

Simulate the real detection process through X-ray simulation tools, obtain the three-dimensional structural information of the workpiece by reading the CAD file of the tested sample, and simulate the X-ray detection process of the workpiece without the need for a real workpiece, obtaining the X-ray simulation image of the workpiece.

The simulation program can quickly obtain the minimum detectable defect size under different parameter conditions. By changing the combination of scanning voltage and current, dark defects inside the workpiece can be detected from scratch, allowing for a clear and intuitive assessment of the defect detection ability of the projected image, thereby optimizing the acquisition parameter settings.

3.4 Reducing measurement system errors by using standard molds

By using standard molds with specific structures and materials, on the one hand, it can be used to study the measurement characteristics of industrial CT. On the other hand, the measurement results of industrial CT images (pixels) can be converted into international standard units of measurement (m) by measuring known size molds. The process of establishing industrial CT value sources can also be used as recognized molds to compare measurement accuracy between different devices, or to compare the measurement accuracy of CT systems with traditional CMM equipment, Establish uncertainty in industrial testing and promote the standardization process of industrial CT measurement.

3.5 Improving the measurement accuracy of industrial CT through the combination of traditional measurement methods

Traditional industrial measurement methods, such as contact based coordinate measuring instruments, can only measure limited points on complex surfaces and cannot fully reflect the shape of the surface. Although their single point measurement accuracy is high, using limited points to describe complex surfaces can actually lead to a decrease in overall accuracy. On the contrary, optical measurement methods such as industrial CT are very suitable for digitizing free-form surfaces and complex surfaces that require large-scale measurement points.

By using higher precision optical scanners to measure the external dimensions of the sample, the advantage of using traditional CMM equipment to measure external dimensions with higher accuracy is achieved. To correct the measurement results of industrial CT. This method combines the advantages of non-destructive measurement of the internal structure of objects using industrial CT with higher external accuracy using traditional methods, and indeed improves the measurement accuracy of industrial CT in practical applications.

By integrating other measuring devices (such as optics) into the industrial CT system, both industrial CT measurement and optical measurement can be completed simultaneously on a single industrial CT device, truly transforming industrial CT into a specialized non-contact industrial measurement device. This system advantage. By integrating optical measurement and industrial CT measurement data, the measurement accuracy of the system can be enhanced. It is convenient to directly compare industrial CT scanning with traditional optical scanning data for measurement error analysis.