TWI633278B - Three-dimensional space touch panel detection system and method - Google Patents

Abstract

The invention discloses a three-dimensional space touch panel detection system and method, which includes a three-axis mobile platform, a multi-axis rotation mechanism, a test touch lever, a light source, an image capture device, and a control module. The multi-axis rotating mechanism is provided on the third driving mechanism of the three-axis moving platform. The test touch lever can perform touch detection on the panel under test with the rotation of the multi-axis rotation mechanism. The light source is arranged above the multi-axis rotation mechanism and is used to project the light source toward the panel to be tested. The image capturing device is used to capture the light spot image of the reflective light source of the panel under test. The control module controls the three-axis mobile platform to make the corresponding displacement according to the preset path, and then performs image processing and true circle arithmetic operation on the light spot image, and controls the multi-axis rotating mechanism to make the corresponding uniaxial axis according to the calculation result Or the two axial rotations are used to make the normal of the contact point of the test touch lever and the touch surface the same axis, so as to reduce the production cost and simplify the test standard path, so as to achieve the best surface measurement and inspection equipment level.

Description

Three-dimensional space touch panel detection system and method

The invention relates to a three-dimensional space touch panel detection system and method, in particular to a touch panel detection technology that can reduce production costs and simplify test standard paths to achieve the best curved surface measurement and detection equipment level.

According to a Nikkei news report, Sumitomo Chemical will increase the production of curved touch panel products for smartphones equipped with OLED panels. It aims to expand the capacity of curved touch panels to 100 million mobile phones per year in early 2018. The main reason is that in recent years, mobile phones, tablets, computers, and multimedia TVs of Samsung, Apple, and other international manufacturers have gradually moved toward curved panel design. At present, most of the testing equipment used in factory production and production are flat touch detection. The panel device does not have a dedicated three-dimensional curved touch panel detection device.

Although a few manufacturers have developed curved touch panel detection devices, most of these curved touch panel detection devices use mechanical arms to detect and process 3D curved touch panels, although robotic arms can be used to detect 3D curved touch panels; However, the cost of applying a robotic arm to process 3D curved touch panel inspection equipment is too high, which affects the manufacturer's willingness to install, so that the robotic arm cannot be popularized as a necessary inspection device for detecting curved touch panel; not only that, the required displacement Path control will lengthen the learning curve of engineers, which will greatly increase the cost of manpower and time, thus causing inconvenience and great troubles in the detection of curved touch panels.

In view of this, the planar touch panel inspection equipment has been in operation for many years, the maturity of the product technology has not been perfected, and there is still a need for further improvement. Therefore, how to develop a set that can independently pursue the curved surface to achieve touch inspection Purpose and simplification of touch panel inspection technology for generating test standard paths has become a technical issue that is urgently needed to be challenged and overcome by relevant industry and academics.

The technical problem to be solved by the present invention is to solve the problems of high detection cost of the 3D curved touch panel currently detected by the mechanical arm detection processing and the long learning curve of the engineer.

The first object of the present invention is to provide a three-dimensional space touch panel detection system and method, which is mainly based on a planar touch panel detection device, and is equipped with an independent tracking surface vertical distance on the Z axis of the planar touch panel detection device In addition to the control mechanism that can realize the independent tracking of the surface to achieve the purpose of surface touch detection, it can also reduce production costs and simplify the production of standard test paths, thereby achieving the best level of surface measurement and testing equipment. The technical means to achieve the first goal and solve the problem include three-axis mobile platform, multi-axis rotation mechanism, test touch lever, light source, image capture device and control module. The multi-axis rotating mechanism is provided on the third driving mechanism of the three-axis moving platform. The test touch lever can perform touch detection on the panel under test with the rotation of the multi-axis rotation mechanism. The light source is arranged above the multi-axis rotation mechanism and is used to project the light source toward the panel to be tested. The image capturing device is used to capture the light spot image of the reflective light source of the panel under test. The control module controls the three-axis mobile platform to make the corresponding displacement according to the preset path, and then performs image processing and true circle arithmetic operation on the light spot image, and controls the multi-axis rotating mechanism to make the corresponding uniaxial axis according to the calculation result Or two axial rotations, so that the normal of the contact point of the test touch lever and the touch surface is the same axis or at an angle close to the axis.

The second object of the present invention is to provide a method which can effectively shorten the detection time. A three-dimensional space touch panel detection system and method thereof, without the need to create curved 3D preset path data, only the operation of 2D preset path data can detect the 3D curved surface to be tested, thereby effectively shortening the detection time Effect. The technical means adopted to achieve the second most obvious objective include three-axis mobile platform, multi-axis rotation mechanism, test touch lever, light source, image capture device and control module. The multi-axis rotating mechanism is provided on the third driving mechanism of the three-axis moving platform. The test touch lever can perform touch detection on the panel under test with the rotation of the multi-axis rotation mechanism. The light source is arranged above the multi-axis rotation mechanism and is used to project the light source toward the panel to be tested. The image capturing device is used to capture the light spot image of the reflective light source of the panel under test. The control module controls the three-axis mobile platform to make the corresponding displacement according to the preset path, and then performs image processing and true circle arithmetic operation on the light spot image, and controls the multi-axis rotating mechanism to make the corresponding uniaxial axis according to the calculation result Or two axial rotations, so that the normal of the contact point of the test touch lever and the touch surface is the same axis or at an angle close to the axis. Wherein, this embodiment mainly lies in that the above-mentioned predetermined path is the path data of the 2D plane.

10‧‧‧Three axis mobile platform

11‧‧‧ First driving mechanism

12‧‧‧ Second drive mechanism

13‧‧‧ Third drive mechanism

20‧‧‧Multi-axis rotating mechanism

21‧‧‧horizontal rotating mechanism

22‧‧‧Longitudinal rotation mechanism

30‧‧‧Test Touch Bar

40‧‧‧Light source

41‧‧‧Image capture device

Lens

50‧‧‧Control module

60‧‧‧ Panel to be tested

61‧‧‧Touch surface

FIG. 1 is a schematic diagram of the implementation of the functional blocks of the present invention.

FIG. 2 is a schematic view of the implementation of the specific architecture of the present invention.

FIG. 3 is a schematic diagram of a specific implementation of the multi-axis rotation mechanism of the present invention.

FIG. 4 is a schematic diagram of an embodiment of the present invention for testing touch bars to test touch surfaces with different shapes.

5 is a schematic diagram of the normal distribution of the touch surface of the present invention.

FIG. 6 is a schematic diagram of the operation of the real circle calculation method of the present invention.

7 is a schematic diagram of the flow control implementation of the image processing and true circle arithmetic method of the present invention.

FIG. 8 is a schematic diagram of an image comparison implementation of the first experimental example of the present invention.

9 is a schematic diagram of an image comparison implementation of a second experimental example of the present invention.

FIG. 10 is a schematic diagram of an image comparison implementation of a third experimental example of the present invention.

FIG. 11 is a schematic diagram of an image comparison implementation of a fourth experimental example of the present invention.

FIG. 12 is a schematic diagram of an image comparison implementation of a fifth experimental example of the present invention.

FIG. 13 is a schematic diagram of an image comparison implementation of a sixth experimental example of the present invention.

In order for your reviewer to further understand the overall technical features of the present invention and the technical means for achieving the purpose of the invention, the specific embodiments and drawings are used to explain in detail as follows: Please refer to Figures 1 to 5 for the invention The specific embodiment of the first object includes technical features such as a three-axis mobile platform 10, a multi-axis rotation mechanism 20, a test touch lever 30, a light source 40, an image capture device 41, and a control module 50. The multi-axis rotating mechanism 20 is provided on the third driving mechanism 13 of the three-axis moving platform 10. The three-axis mobile platform 10 can accommodate at least one panel under test 60. The three-axis mobile platform 10 includes a first driving mechanism 11 (that is, an X-axis displacement) that can be displaced to three different axial directions, and a second The driving mechanism 12 (that is, the displacement toward the Y axis) and a third driving mechanism 13 (that is, the displacement toward the Z axis). The multi-axis rotating mechanism 20 is provided on the third driving mechanism 13 to make at least two kinds of axial (namely horizontal and vertical) rotations. The test touch lever 30 (ie testing the copper head) is provided on the multi-axis rotating mechanism 20. The multi-axis rotation mechanism 20 can be displaced and rotated to perform touch detection on the test panel 60. A light source 40 (such as a light-emitting diode LED; or a laser) is disposed on the multi-axis rotation mechanism 20 near the test touch The position of the lever 30 is used to project the light source toward the touch surface 61 of the panel under test 60. The image capturing device 41 is provided at a position of the multi-axis rotation mechanism 20 close to the light source 40 for capturing from the panel under test 60 The light spot image of the reflected light source. The control module 50 has at least one preset path built in, and controls the three-axis mobile platform 10 to make a corresponding displacement according to the preset path, and then performs image processing and true circle on the light spot image The calculation of the arithmetic method, and according to the calculation result, the multi-axis rotating mechanism 20 is controlled to make corresponding single-axis or two-axis rotation to test the touch lever 30 The axis of the touch point and the normal of the touch point 61 are the same axis; or close to the same axis; in other words, the axis coincides with the normal, or is close to the state of coincidence. As for "near the same axis" and "nearly coincident state", it means that the angle between the axis of the touch lever 30 and the normal of the contact point is less than 3 degrees.

Please refer to the embodiment shown in FIG. 3, the image capturing device 41 and the light source 40 are located near the top of the test touch lever 30. Specifically, the multi-axis rotation mechanism 20 includes a horizontal rotation mechanism 21 that can rotate horizontally, and a vertical rotation mechanism 22 that can rotate vertically. The horizontal rotation mechanism 21 is provided on the end of the actuation lever 130 of the third drive mechanism 13. The vertical rotation mechanism 22 is provided at the bottom of the horizontal rotation mechanism 21, and the image capturing device 41 and the light source 40 are provided on the vertical rotation mechanism 22.

Please refer to FIGS. 1 to 5 for a specific embodiment of achieving the second object of the invention, which includes a three-axis mobile platform 10, a multi-axis rotation mechanism 20, a test touch lever 30, a light source 40, an image capture device 41 and Control module 50 and other technical features. The multi-axis rotating mechanism 20 is provided on the third driving mechanism 13 of the three-axis moving platform 10. The three-axis mobile platform 10 can accommodate at least one panel under test 60. The three-axis mobile platform 10 includes a first driving mechanism 11 (that is, an X-axis displacement) that can be displaced to three different axial directions, and a second The driving mechanism 12 (that is, the displacement toward the Y axis) and a third driving mechanism 13 (that is, the displacement toward the Z axis). The multi-axis rotating mechanism 20 is provided on the third driving mechanism 13 to make at least two kinds of axial (namely horizontal and vertical) rotations. The test touch lever 30 (ie testing the copper head) is provided on the multi-axis rotating mechanism 20. The multi-axis rotation mechanism 20 can be displaced and rotated to perform touch detection on the test panel 60. A light source 40 (such as a light-emitting diode LED; or a laser) is disposed on the multi-axis rotation mechanism 20 near the test touch The position of the lever 30 is used to project the light source toward the touch surface 61 of the panel under test 60. The image capturing device 41 is provided at a position of the multi-axis rotation mechanism 20 close to the light source 40 for capturing from the panel under test 60 The light spot image of the reflected light source. The control module 50 has at least one preset path built in, and controls the three-axis mobile platform 10 to make a corresponding displacement according to the preset path, and then The light spot image is processed by image processing and true circle arithmetic, and the multi-axis rotation mechanism 20 is controlled to make corresponding uniaxial or two-axis rotation according to the calculation result, which is used to test the touch lever 30 and touch The normal of the contact point of the control surface 61 is the same axis; or close to the same axis, that is, the angle between the test touch lever 30 and the axis is within an angle range of 0.1 to 3 degrees. Among them, this embodiment mainly lies in that the above-mentioned preset path is 2D plane path data. Here, it is possible to operate only with 2D default path data without making curved 3D default path data The 3D curved surface to be tested panel 60 is detected.

Specifically, please refer to FIG. 7. The image processing performed by the control module 50 includes step 1: converting the captured color spot image into a grayscale image. Step 2: Use Canny algorithm to obtain the image edge of the light spot image in gray scale. Step 3: Perform image binarization on the image edge.

Specifically, please refer to FIG. 7. The real circle calculation method performed by the control module 50 includes step 1: inputting the number of edge points of the light spot image. Step 2: Randomly select four edge points. Step 3: Use the theory of Klama formula to determine the candidate circle, solve the center and radius of the circle, and then compare the straight axis and the intersection axis of the circle to determine whether it is a true circle. Step 4: Determine whether the true circle failure number is less than the tolerance number; if the judgment result is yes, return to step two; if the judgment result is no, go to step six. Step 6: Move the lens of the image capturing device 41 to the center of the circle.

The system structure of the present invention is to establish a control module 50 (such as a microcontroller; or a combination of computer and software) for autonomously tracking the vertical distance of the curved surface in three-dimensional space control, in which the Z axis is passed on X, Y, Z mobile platforms The control module 50 for self-tracking the vertical distance of the curved surface is under control. Two sets of movable A-axis and B-axis are installed on the Z-axis gantry. An image capturing device 41 with a photographic lens is mounted on the movable A-axis. A laser or LED is placed on the same axis to form a phase in front of the lens. The angle captured by the lens adjusts the angle of the probe so that the vertical normal of the touch surface (ie the test surface) is the same axis as the test touch lever 30, and Control the touch circle area of the test touch lever 30 (that is, the test copper head) and The contact surface to be tested is completely tight, and the finished panel to be tested 60 can be reported via the built-in wifi or Bluetooth interface, and the control module 50 can trace the path.

The completed three-dimensional space control can provide complete support for the detection problems of the three-dimensional curved surface touch panel, so as to reduce the cost of purchasing a mechanical arm to handle the 3D curved surface touch panel, and increase the test flexibility by reporting the panel 60 to be reported via wifi or Bluetooth interface , Reduce the learning curve of engineers and reduce the man-hours of transfer lines.

The system structure to be developed by the present invention is shown in FIG. 3. A group of light sources 40 with LEDs and an image capturing device 41 with lenses are arranged on the A and B axes of the device on the Z axis. The touch surface 61 of the panel under test 60 is phased in front of the lens to form a circular light spot image. The distance between the straight axis and the intersecting axis of the circular light spot in the image is used to determine the change of the reflective surface of the light spot. , And rotate the Z axis and the two axes of A and B to adjust the lens axis until it overlaps the normal of the touch surface 61; that is, the normal of the contact point of the test touch lever 30 and the touch surface 61 For the same axis, as shown in Figures 4 and 5.

In general, the original image of the light spot image reflected from the panel under test 60 must undergo image processing to color the original image of the light source image of the LED light source 40 reflected by the panel under test 60 through the image processing software of the control module 50 After grayscale conversion, binarization and edge extraction, the LED spot is tracked using circle detection. In circle detection, the set of edge points is obtained from the captured spot image, usually using the following equation To express: V = {( x , y )} (1)

( x - a ) ² +( y - b ) ² = r ² (2)

Assuming that the given edge points are ( a , b ) and the set is V, we randomly pick four points from V. The four points can determine four circles, as shown in Figure 6a. Assuming that the selected four points are all from the same circle, it can be said that the circle determined by these four points is a candidate circle.

The control system for the calculation of the control module 50 of the present invention adopts a four-core high-speed microprocessor or microcontroller. As for the image acquisition, true circle calculation, and adjustment of the touch detection angle, the built-in image processing and The real circle arithmetic is done automatically.

The following is a description of each part. The theory adopted in the present invention is the theory of Klama formula to determine the candidate circle. Rewriting the circle equation in equation (2), we can obtain the following formula: 2 xa + 2 yb + d = x ² + y ² (3)

Here d = r ² - a ² - b ² . Let v _i =( x _i , y _i ), i =1,2,3 be three randomly selected edge points in the image. If v ₁ , v ₂ , and v _{3 are} not collinear, they can determine a circle C ₁₂₃ and get the center of the circle ( a ₁₂₃ , b ₁₂₃ ) and the radius r ₁₂₃ .

Center and radius: Substitute the three edge points v ₁ =( x ₁ , y ₁ ), v ₂ =( x ₂ , y ₂ ), v ₃ =( x ₃ , y ₃ ) into (3)

Here .

Using the solution of the center of the circle ( a ₁₂₃ , b ₁₂₃ ), we can get the radius of the circle as follows:

The theory adopted by the present invention is the theory of Klama formula to determine the candidate circle, and the circle equation in formula (2) is rewritten to obtain the following formula: when v ₁ , v ₂ , and v _{3 are} collinear: The three edge point values selected by the formula derivation conclusion can unfortunately satisfy the equation ( x ₂ - x ₁ )( y ₃ - y ₁ )-( x ₃ - x ₁ )( y ₂ - y ₁ )=0 means That is, the three randomly selected edge points v ₁ , v ₂ and v _{3 are} collinear. In other words, they cannot form a circle. Let v ₄ = ( x ₄ , y ₄ ) be the fourth selected edge point, and let the distance from this point to the circle C ₁₂₃ be d _4→123 , the following formula can be obtained:

If v _{4 is} on circle C ₁₂₃ , then equation (7) is zero. If the value in equation (7) is small enough, we all regard v _{4 to be} on the boundary of circle C ₁₂₃ , as shown in Figure 6b, and give four random edge points v _i =( x _i , y _i ), i =1 ,2,3,4, these four side points make up to four circles. The circle table caused by v _i , v _j , v _k is C _ijk and its center and radius are expressed as ( a _ijk , b _ijk ) and r _ijk . Let the distance from v _l to the circle C _ijk be d _{l → ijk} (see equations (5)-(7)). Then formula (7) can be changed to:

The main goal here is to decide from three randomly selected edge points which three points can form a digital circle. At the same time, the fourth edge point also falls on the edge of the circle. Give four points, a total of This possible circle needs further examination to determine who is the most probable circle. If there are enough points, they can be combined into a digital circle, as shown in Figure 6c.

In the non-ideal situation, this non-ideal situation occurs when two of the three agent points are very close. Here, this possible circle has a high probability of not being a true circle. In order to avoid this undesirable situation, we hope that the distance between any two agent points must exceed a threshold. Adding this condition will make the candidate circle determined by the three proxy points have stronger evidence as a true circle, as shown in Figure 6d.

When determining the true circle, suppose that the above methods v _i , v _j , v _k determine a possible circle with a circle center ( a _ijk , b _ijk ) and radius r _ijk , then we add a threshold value T _g To check if this candidate circle is a real circle. Let the initial value of counter C be 0. We select any edge point v _l from the edge point set V , and then check whether the distance d _{l → ijk} is less than the threshold value T _d . If it is, the value of C is increased by one. Then, we pick one more point from the remaining set of edge points, and continue the above distance calculation and comparison until all the edge points are processed. If the value of C is greater than the threshold value T _{g at} this time, the candidate circle formed by v _i , v _j , v _{k is} the true circle. Otherwise, the candidate circle is a false circle. Next, we return these C edge points to the edge point set V. The method of formulating the threshold value of C is assuming that the center and radius of the candidate circle are ( a _ijk , b _ijk ) and r _ijk , then It needs to be greater than a threshold value T _r , for example, T _r = 0.8 means that the edge points on the digital circle need to occupy 80% of the circumference. After all, the number of points on the circumference is proportional to the radius. Such a set threshold T _r less susceptible to the influence of the size of the circle.

The control module 50 used in the present invention is a microcontroller with four cores, 64Bit, 1.2GHz, and memory capacity of 1G. The system of the present invention operates in the LINUX operating system, with high automation and parallel calculation functions. It can process images in parallel, track specific targets, peripheral interfaces and sensor processing, and can perform curved touch panel detection without the support of specific 3D scanning devices. The touch panel detection object is finished products; or semi-finished products, semi-finished products can be touched The control IC of the panel reports to the control module 50 via I2C, SPI, CAN BUS and other signal transmission interfaces. In addition, the finished product is controlled by the program APK installed by the host under test through the built-in Bluetooth or Wifi communication module. Group 50 reports. The Z axis of the system automatically adjusts the curvature of the surface of the object to be measured and is fixed to the touch copper head next to the lens, so that the angle of the copper head is parallel to the normal of the surface to be measured on the surface of the object to be measured. The detection system does not need to make a curved 3D path. The operation of the 2D path drawing can detect the 3D curved surface and shorten the detection time.

Please refer to FIG. 7 for the following, and the method for manufacturing the tracking image processing for touch detection under test of the present invention is as follows:

Step 1: First use a steering lens module fixed on the Z axis, and use the LED light source 40 on the lens to illuminate the spot image formed on the surface of the panel 60 to be tested.

Step 2: Binaryize the obtained color light spot image, and filter out unnecessary background through the software filter, and strengthen the LED light spot feature.

Step 3: Use the Canny algorithm to obtain the edge of the image to make the round spot of the LED light spot more prominent.

Step 4: Use the theory of Klama formula to determine the candidate circle, solve for the center and radius of the circle, and then determine the true circle by comparing the straight axis and the intersection axis of the circle.

Step 5: Through the real circle detection, confirm that the lens axis is parallel to the surface normal of the object to be measured.

The experimental example of the lens acquisition calculation of the present invention has the following preliminary experimental results:

Experimental example one: Take a picture of the front of the touch curved surface. The left side of FIG. 8 is the sampling image, and the right side of FIG. 8 is the image after image processing.

Experimental example 2: Long-distance image acquisition on the front of the touch curved surface. The side of FIG. 9 is the sampling image, and the right side of FIG. 9 is the image after image processing.

Experiment 3: Right-distance image acquisition of the touch surface. The left side of Fig. 10 is the sampling image, and the right side of Fig. 10 is the image after image processing.

Experiment 4: Right-biased capture of the touch surface. The left side of Fig. 11 is the sampling image, and the right side of Fig. 11 is the image after image processing.

Experiment 5: The left-handed image of the touch surface is taken. The left side of FIG. 12 is a sampling image, and the right side of FIG. 12 is an image after image processing.

Experiment 6: Touch curved convex image acquisition, the left side of Fig. 13 is the sampling image, and the right side of Fig. 13 is the image after image processing.

The above is only a feasible embodiment of the present invention and is not intended to limit the patent scope of the present invention. Any equivalent implementation of other changes based on the content, features and spirit described in the following claims should be Included in the patent scope of the present invention. The structural features of the invention specifically defined in the claim are not found in similar items, and are practical and progressive. They have met the requirements of the invention patent. You have filed an application in accordance with the law, and I would like to ask the Jun Bureau to approve the patent in accordance with the law to maintain this. The applicant's legal rights and interests.

Claims (9)

A three-dimensional space touch panel detection system includes: a three-axis mobile platform for placing at least one panel to be tested, the three-axis mobile platform includes a first driving mechanism that can be displaced to three different axial directions, A second drive mechanism and a third drive mechanism; a multi-axis rotation mechanism, which is arranged on the third drive mechanism, to make at least two kinds of axial rotation; at least one test touch lever, which is arranged on The multi-axis rotation mechanism can follow the displacement and rotation of the multi-axis rotation mechanism to perform touch detection on a touch surface of the panel to be tested; a light source is provided on the multi-axis rotation mechanism close to the test touch The position of the rod is used to project the light source toward the touch surface of the panel under test; an image capture device is provided at a position of the multi-axis rotation mechanism close to the light source to capture the reflection of the panel under test A light spot image of the light source; and a control module, which has at least one preset path built in, and controls the three-axis mobile platform to make a corresponding displacement according to the preset path, and then performs image processing on the light spot image And true circle arithmetic, and control the multi-axis rotation mechanism to make corresponding single-axis or two-axis rotation according to the calculation result, so as to make the axis of the test touch lever and the touch surface The normal of the contact point is the same axis; or close to the same axis. The three-dimensional space touch panel detection system according to claim 1, wherein the image capturing device and the light source are located at a position near the top of the test touch rod. The three-dimensional space touch panel detection system according to claim 1, wherein the multi-axis rotation mechanism includes a horizontal rotation mechanism capable of rotating horizontally, and a vertical rotation mechanism capable of rotating longitudinally; the horizontal rotation mechanism is provided at The end of the actuating rod of the third driving mechanism; the longitudinal rotation mechanism is arranged at the bottom of the horizontal rotation mechanism; the image capturing device and the light source are arranged on the longitudinal rotation mechanism. The three-dimensional space touch panel detection system according to claim 1, wherein the predetermined path is path data of a 2D plane. The three-dimensional space touch panel detection system according to claim 1, wherein the image processing includes the following steps: Step 1: Convert the captured color light spot image into grayscale image; Step 2: Use Canny algorithm to obtain The image edge of the light spot image in gray scale; and Step 3: Perform image binarization processing on the image edge. The three-dimensional space touch panel detection system according to claim 1, wherein the true circle algorithm includes the following steps: Step 1: input the number of edge points of the light point image; Step 2: randomly select four edges in the edge point set Point; step three: use the theory of Klama formula to determine the candidate circle, and solve for the center and radius of the circle, and then determine whether it is a true circle through the straight axis of the circle and the intersection axis; step four: determine whether the number of true circle failures is less than the tolerance number ; If the judgment result is yes, go back to step two; if the judgment result is no, go to step six; and Step six: move one lens of the image capture device to align with the center of the circle A three-dimensional space touch panel detection method, which includes: providing a three-axis mobile platform, a multi-axis rotation mechanism, at least one test touch lever, a light source, an image capture device, and a control module; wherein, the three The axis moving platform includes a first drive mechanism, a second drive mechanism, and a third drive mechanism that are respectively displaceable in three different axial directions; the multi-axis rotation mechanism is provided on the third drive mechanism to make at least Two axial rotations; the test touch lever is provided on the multi-axis rotation mechanism, which can touch at least one panel to be tested placed on the three-axis mobile platform as the multi-axis rotation mechanism is displaced and rotated Detection; the light source is provided at a position of the multi-axis rotation mechanism close to the test touch lever for projecting the light source toward a touch surface of the panel to be tested; the image capture device is provided at the multi-axis rotation mechanism close to the The position of the light source is used to capture the light point image of the light source reflected by the panel under test; and at least one preset path is built in the control module, and the three-axis mobile platform is controlled according to the preset path to make Corresponding displacement, then perform image processing and true circle arithmetic operation on the spot image, and control the multi-axis rotation mechanism to make corresponding single-axis or two-axis rotation according to the calculation result, so as to make The normal of the contact point of the test touch rod and the touch surface is the same axis; or close to the same axis. The method for detecting a three-dimensional space touch panel according to claim 7, wherein the image processing includes the following steps: Step 1: convert the captured color spot image into a grayscale image; Step 2: use the Canny algorithm to obtain The image edge of the light spot image in gray scale; and Step 3: Perform image binarization processing on the image edge. The method for detecting a three-dimensional space touch panel according to claim 7, wherein the true circle operation method includes the following steps: Step 1: input the number of edge points of the light point image; Step 2: randomly select four edges in the edge point set Point; step three: use the theory of Klama formula to determine the candidate circle, and solve for the center and radius of the circle, and then determine the true circle by comparing the straight axis and the intersection axis of the circle; step four: determine whether the number of true circle failures is less than the tolerance number; If the judgment result is yes, go back to step two; if the judgment result is otherwise, go to step six; and step six: move one lens of the image capturing device to align the center of the circle.