Automatic wheel mounting device

Abstract

The present application provides a kind of wheel automatic installation device of installing wheel according to the inclination of the wheel hub of vehicle, can improve the efficiency of wheel automatic installation device.Wheel automatic installation device includes detection device.Detection device is imaged to the wheel hub or wheel of vehicle, and the position of wheel hub bolt or the position of wheel bolt hole is detected.Detection device includes acquisition unit and circular extraction unit.Acquisition unit obtains the basic position information of the position of wheel hub bolt or the position of wheel bolt hole.Circular extraction unit extracts multiple circles existing in the imaged image.In the combination of multiple circles, circular extraction unit extracts the combination consistent with basic position information as the combination of wheel hub bolt or the combination of bolt hole.

Description

Technical Field

[0001] This invention relates to an automatic wheel installation device, and more particularly to an automatic wheel installation device that detects the tilt of a vehicle's wheel hub to install the wheel. Background Technology

[0002] Since then, efforts have continued to be made to mitigate climate change or reduce its impacts, and research and development related to reducing carbon dioxide emissions are underway to achieve these goals.

[0003] In existing automatic wheel mounting devices, a stereo camera consisting of a front-facing camera and an oblique-facing camera is used to detect the circular position of the front end of the wheel hub bolts, and the center position (X, Y, Z) of the wheel hub and the posture of the wheel hub surface (toe, camber and phase) are determined based on the detection results.

[0004] When performing circular detection on the front end of wheel hub bolts, a preliminary step of roughly detecting the bolt positions (i.e., detection without requiring precise positioning) is used to prevent false detections of circular components such as countersunk screws and service hole covers on the wheel hub surface. As a current rough detection method, the edge shape of the bolts can be pre-registered as a model, and one bolt can be detected through pattern matching. Then, using the known phase relationships of the bolt positions (72° each for 5 bolts, 90° each for 4 bolts), the area where the bolts can be identified, including the positions of the remaining bolts, is determined.

[0005] [Existing Technical Documents]

[0006] [Patent Literature]

[0007] [Patent Document 1]: Japanese Patent Application Publication No. 2009-178829 Summary of the Invention

[0008] However, current coarse inspection methods, such as pattern matching, may mistakenly detect countersunk screws and maintenance hole covers when inspecting bolts. This is because, due to the nature of pattern matching, if the stored model differs from the appearance of the bolts captured during wheel installation, accurate detection becomes difficult. Furthermore, since the success of coarse bolt inspection depends on the pattern matching result of a single bolt, this can easily lead to errors, prolonging inspection time, reducing production efficiency, and increasing carbon dioxide emissions. Improving the efficiency of inspection and operation is a key challenge for designers in this field in reducing carbon dioxide emissions.

[0009] In view of the above, this invention aims to improve the efficiency of the automatic wheel mounting device in order to solve the aforementioned problems. Furthermore, it helps to mitigate or reduce the impact of climate change.

[0010] To achieve the aforementioned objective, the present invention provides an automatic wheel mounting device for mounting wheels according to the tilt angle of a vehicle's wheel hub. The automatic wheel mounting device includes a detection unit. The detection unit captures an image of the vehicle's wheel hub or wheel and detects the position of the wheel hub bolts or the position of the wheel bolt holes. The detection unit includes an acquisition unit and a circular extraction unit. The acquisition unit acquires basic positional information about the position of the wheel hub bolts or the position of the wheel bolt holes. The circular extraction unit extracts multiple circles present in the captured image. Among the combinations of multiple circles, the circular extraction unit extracts the combination that matches the basic positional information as the combination of wheel hub bolts or the combination of bolt holes.

[0011] In one embodiment of the present invention, when there is no combination consistent with the basic position information among the multiple circles, at least one imaginary circle is created. Among the multiple circles and the imaginary circle, the circle extraction unit extracts the combination consistent with the basic position information as the combination of hub bolts or bolt holes.

[0012] Based on the above, the automatic wheel installation device of the present invention does not perform a difficult search by pattern matching from objects with similar shapes. Instead, it initially detects all multiple circles on the wheel hub surface, including bolts, countersunk screws, maintenance hole covers, etc. Furthermore, by reducing the number of candidate bolt circles (representative circles) and adding conditions during the search, a search range that prevents the erroneous detection of other circular parts can be determined, reducing false detections during coarse detection and maintaining real-time processing within a line takt (approximately 1.0 second for the entire sensing process). This allows for the extraction of only candidate bolt circles, determining a search range that prevents the erroneous detection of other circular parts, and improving the efficiency of the automatic wheel installation device.

[0013] To make the above features and advantages of the present invention more apparent and understandable, embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0014] Figure 1 This is a schematic block diagram illustrating an embodiment of the automatic wheel mounting device of the present invention.

[0015] Figure 2 A flowchart illustrating a rough detection method according to an embodiment of the present invention.

[0016] Figure 3 This is a flowchart illustrating a method for detecting multiple circles according to an embodiment of the present invention.

[0017] Figure 4 This illustrates an example of an edge extraction method according to an embodiment of the present invention.

[0018] Figure 5 This illustrates an example of a region generation method according to an embodiment of the present invention.

[0019] Figure 6 A flowchart illustrating a method for determining the combination of multiple circles according to an embodiment of the present invention.

[0020] Figure 7 This illustrates an example of a method for generating a bounding box according to an embodiment of the present invention.

[0021] [Explanation of Symbols]

[0022] 100: Automatic wheel mounting device

[0023] 110: Detection device

[0024] 112: Acquisition Department

[0025] 114: Camera Department

[0026] 116: Circular extraction section Detailed Implementation

[0027] Reference will now be made in detail to exemplary embodiments of the invention, examples of which are illustrated in the accompanying drawings. Wherever possible, the same element references are used in the drawings and description to denote the same or similar parts.

[0028] Figure 1 This is a schematic block diagram illustrating an embodiment of the automatic wheel mounting device of the present invention. The automatic wheel mounting device 100 can mount wheels according to the inclination of the vehicle's wheel hub at a wheel mounting station in a vehicle manufacturing plant. Please refer to... Figure 1The automatic wheel mounting device 100 includes a detection device 110. The detection device 110 can capture images of the vehicle's wheel hub or wheel and detect the position of the wheel hub bolts or the position of the wheel bolt holes. The detection device 110 includes an acquisition unit 112, a camera unit 114, and a circular extraction unit 116. The acquisition unit 112 can acquire basic position information IB of the wheel hub bolts or the position of the wheel bolt holes. For example, the acquisition unit 112 can acquire pre-stored basic position information IB from a memory or cloud server, or the operator can input the vehicle's basic position information IB to the acquisition unit 112 via an input device such as an input button, remote control, touchpad, or touch panel with resistive, capacitive, or other types of touch sensing components. The basic position information IB includes information such as the position and number (4 or 5) of the wheel hub bolts on the production line, the position and number (4 or 5) of the wheel bolt holes, and the spacing between each bolt or bolt hole. The acquisition unit 112 can be implemented, for example, by a circuit capable of processing information data, such as a large-scale integrated circuit (LSI) or an application-specific integrated circuit (ASIC).

[0029] The camera unit 114 is, for example, a camera that can capture images of the vehicle's wheel hub or wheels to generate a video image CI, and transmit the video image CI to the circular extraction unit 116.

[0030] The circular extraction unit 116 is, for example, a central processing unit (CPU), or other programmable general-purpose or special-purpose microprocessor, digital signal processor (DSP), programmable controller, application-specific integrated circuit (ASIC), or other similar components or combinations thereof. The circular extraction unit 116 is coupled to the acquisition unit 112 and the imaging unit 114. The circular extraction unit 116 can extract multiple circles present in the captured image CI. Among the extracted combinations of multiple circles, the circular extraction unit 116 can extract combinations consistent with the basic position information IB as combinations of hub bolts or bolt holes.

[0031] Figure 2 This is a flowchart illustrating a coarse detection method according to an embodiment of the present invention. Please also refer to... Figure 1 and Figure 2 The detection method in this embodiment is applicable to Figure 1The automatic wheel mounting device 100 can analyze camera images (CI) of the wheel hub or wheel of a vehicle to filter out the circles formed by the wheel hub bolts or bolt holes, so as to facilitate accurate wheel mounting by the automatic mounting device 100. The following describes the detection method of this embodiment of the invention in conjunction with the various components of the automatic wheel mounting device 100.

[0032] First, in step S201, the circle extraction unit 116 detects multiple circles in the camera image CI. Next, in step S202, the circle extraction unit 116 determines combinations of the detected circles. Finally, in step S203, the circle extraction unit 116 generates a bounding box around each circle.

[0033] Figure 2 For a detailed explanation of the steps in step S201 regarding the detection method for multiple circles, please refer to [link / reference]. Figure 3 Each step of the process. Figure 3 This is a flowchart illustrating a method for detecting multiple circles according to an embodiment of the present invention. First, in step S301, the circle extraction unit 116 extracts edges from the camera image CI. Specifically, the circle extraction unit 116 scans the camera image CI and searches for edges that can form part of a circle for extraction. In step S302, the circle extraction unit 116 divides the extracted edges into line segments and arcs or elliptical arcs. Next, in step S303, the circle extraction unit 116 removes edges that are too short. Next, in step S304, the circle extraction unit 116 extracts edges with circular elements. Next, in step S305, the circle extraction unit 116 detects circles formed by edges with circular elements. Next, in step S306, the circle extraction unit 116 extracts circles within a set radius range from the detected circles. The set radius depends on the radius of the wheel hub bolts or the bolt hole radius of the vehicle, thereby allowing the circle extraction unit 116 to remove circles that cannot conform to the wheel hub bolts or bolt holes.

[0034] Figure 4 This illustrates an example of an edge extraction method according to an embodiment of the present invention. Figure 4 (a) represents the camera image CI of the hub bolt. Figure 4 (b) represents the case of edge extraction from the camera image CI. Since it extracts all edges that can form a portion of a circle from the camera image CI, in... Figure 4 (b) will have many very short edges. Figure 4 (c) indicates the case after removing edges that are too short. Figure 4 Compared to (b), Figure 4 (c) will only have longer edges that are more likely to form a complete circle. Figure 4(d) represents the situation after extracting circles within a set radius from the detected circles, and further filtering to the remaining circles that may match the hub bolt.

[0035] exist Figure 3 In step S307, the circular extraction unit 116 generates a region that fills the area within each circle. Next, in step S308, the circular extraction unit 116 generates a sum region (set) of all regions. Specifically, overlapping circular regions are merged into a single region. Next, in step S309, the circular extraction unit 116 separates the previously separated regions.

[0036] Figure 5 This illustrates an example of a region generation method according to an embodiment of the present invention. Figure 5 (a) represents the case where the region inside each circle is filled. Figure 5 (b) represents the case where overlapping circular regions are integrated into a single region, and the separate regions are divided. For example, in Figure 5 (b) includes 8 regions A.

[0037] exist Figure 3 In step S310, the circle extraction unit 116 presets i to 0. Next, in step S311, the circle extraction unit 116 determines whether the selected number i of representative circles for a region is less than the number of regions. Specifically, the circle extraction unit 116 can compare the number of segmented regions with the currently selected number i of representative circles for each region. When the selected number i is less than the number of regions, in step S312, the circle extraction unit 116 determines whether there are more than two circles in one of the regions. If there are not more than two circles in this region, it means there is only one circle in this region, and the process proceeds to step S315, where the circle extraction unit 116 uses this circle as the representative circle for this region.

[0038] When the circular extraction unit 116 determines in step S312 that there are two or more circles in this area, in step S313, the circular extraction unit 116 classifies circles with close center positions into the same group. Specifically, the circular extraction unit 116 can classify circles whose center positions are less than a set value into the same group. The smaller the set value, the longer the processing time, and the higher the probability that the circle at the front end of the bolt is used as the representative circle; the larger the set value, the shorter the processing time. The set value can be set by the operator according to actual needs.

[0039] In step S314, the circle extraction unit 116 selects the circle with the highest roundness from each group. Next, in step S315, the circle extraction unit 116 uses the circle with the highest roundness as the representative circle of the region. Next, in step S316, the circle extraction unit 116 increments the selected number i. Then, it returns to step S311 to determine again whether the selected number i of the representative circles of the region is less than the number of regions. If not, the detection method of this embodiment ends. If so, processing continues for the next region in step S312.

[0040] Figure 2 For a detailed explanation of the method for determining the combination of multiple circles in step S202, please refer to [link / reference]. Figure 6 Each step of the process. Figure 6 This is a flowchart illustrating a method for determining the combination of multiple circles according to an embodiment of the present invention. First, in step S401, the circle extraction unit 116 determines whether the set number of bolts is 4 or 5 based on the basic position information IB. When the number of bolts is 4, in step S402, the circle extraction unit 116 determines whether the number of region-representing circles is 4 or more. When the number of region-representing circles is 4, in step S403, the circle extraction unit 116 determines the square relationship of the circle positions. Specifically, the distance ratio between each corner of the square is fixed, and the circle extraction unit 116 can calculate whether the relationship between the four region-representing circles conforms to this ratio. For example, the circle extraction unit 116 can calculate the distance from any region-representing circle to other region-representing circles (the other three). In this embodiment, the distances are sorted from shortest to longest as d1, d2, d3, and the error value e of the distance ratio is calculated. The error value e satisfies, for example, the following relationship: Starting with each of the four circles, calculate each 'e'. The total error value E of the square satisfies, for example, the following relationship: The circular extraction unit 116 can generate a relationship evaluation value for the combination of multiple circles based on the calculated total error value E.

[0041] On the other hand, when the circle extraction unit 116 determines in step S402 that the number of circles representing the region is 5 or more, in step S404, the circle extraction unit 116 sets n to the number of circles, and then in step S405, it selects 4 circles from the n circles as a combination. Next, in step S406, the circle extraction unit 116 determines whether the distance between each circle in the combination is greater than a threshold. If yes, in step S407, the circle extraction unit 116 determines the square relationship of the circle positions. If no, it proceeds directly to step S408. In step S408, the circle extraction unit 116 determines whether there are any undetermined combinations among the n circles. If yes, it returns to step S405 to continue processing the next combination. If no, in step S409, the circle extraction unit 116 extracts the combination that maximizes the relationship evaluation value, which is, for example, the combination with the lowest calculated total error value E.

[0042] When the circular extraction unit 116 determines the number of bolts to be 5 in step S401, in step S410, the circular extraction unit 116 determines whether the number of region-representing circles is 5 or more. When the number of region-representing circles is 5, in step S411, the circular extraction unit 116 determines the regular pentagonal relationship of the circle positions. Specifically, the distance ratio between each corner of the regular pentagon is fixed, and the circular extraction unit 116 can calculate whether the relationship between the 5 region-representing circles conforms to this ratio. For example, the circular extraction unit 116 can calculate the distance from any region-representing circle to other region-representing circles (the other 4). In this embodiment, the distances are sorted from shortest to longest as d1, d2, d3, d4, and the error value e of the distance ratio is calculated. The error value e satisfies, for example, the following relationship: Starting with each of the five circles, calculate each 'e'. The total error value E of the regular pentagon satisfies, for example, the following relationship: The circular extraction unit 116 can generate a relationship evaluation value for the combination of multiple circles based on the calculated total error value E.

[0043] On the other hand, when the circle extraction unit 116 determines in step S410 that the number of circles representing the region is 6 or more, in step S412, the circle extraction unit 116 sets n to the number of circles, and then in step S413, selects 5 circles from the n circles as a combination. Next, in step S414, the circle extraction unit 116 determines whether the distance between each circle in the combination is greater than a threshold. If yes, then in step S415, the circle extraction unit 116 determines the regular pentagonal relationship of the circle positions. If no, it proceeds directly to step S416. In step S416, the circle extraction unit 116 determines whether there are any undetermined combinations among the n circles. If yes, it returns to step S413 to continue processing the next combination. If no, then in step S417, the circle extraction unit 116 extracts the combination that maximizes the relationship evaluation value, which is, for example, the combination with the lowest calculated total error value E.

[0044] Therefore, in the steps Figure 2 In S203, the circular extraction unit 116 generates a bounding box around each circle in the combination that maximizes the relation evaluation value, so that the automatic installation device 100 can accurately install the wheel.

[0045] Figure 7 This illustrates an example of a method for generating a bounding box according to an embodiment of the present invention. Figure 7 (a) represents the combination that maximizes the evaluation value of the extracted relationship. Figure 7 As shown in (a), the combination that maximizes the evaluation value includes circles C1 to C5, and its position is closest to the regular pentagon combination as can be seen from the dashed line in the figure. Figure 7 (b) represents the case where the circle with the highest roundness is selected. Figure 7 In (b), a quadrilateral region surrounding the extracted circle is generated as a bounding box BD, which serves as a reference for alignment when installing the wheel, thereby completing the coarse detection method of this embodiment.

[0046] In one embodiment, when there is no combination consistent with the basic position information IB among the multiple combinations of circles, the circle extraction unit 116 can create at least one imaginary circle based on the basic position information IB, and extract the combination consistent with the basic position information IB from the multiple combinations of circles and the created imaginary circle as the combination of hub bolts or bolt holes.

[0047] It should be noted that, although in Figure 4 , Figure 5 and Figure 7 The example used wheel hub bolts for illustration, but similar testing methods can be applied to bolt holes on wheels to find the combination with the lowest error value.

[0048] In one embodiment, wheel bolts identified by the above method can be further identified through machine learning. Using images from a front-facing camera during mass production, the trained artificial intelligence (AI) detects four or five wheel bolts. For wheel bolt matching rates below a certain threshold, automatic annotation and additional learning are performed. Furthermore, if one of the four or five wheel bolts is not detected, the positions of the remaining three or four bolts are estimated and detected based on image values, and automatically annotated to improve detection accuracy through additional learning. Annotation involves adding relevant information (metadata) to data. Metadata is typically marked using descriptive formats such as XML.

[0049] In summary, the automatic wheel mounting device of the present invention does not perform a difficult search by pattern matching among similarly shaped objects. Instead, it initially detects all multiple circles on the wheel hub surface, including bolts, countersunk screws, maintenance hole covers, etc. Furthermore, by reducing the number of candidate bolt circles (representative circles) and adding conditions during the search, a search range that prevents the erroneous detection of other circular portions can be determined, reducing false detections during coarse inspection. This improves the efficiency of the automatic wheel mounting device and helps mitigate or reduce the impact of climate change.

[0050] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A wheel automatic mounting device for mounting a wheel in accordance with the inclination of a wheel hub of a vehicle, characterized by, Comprising: a detection device that images a wheel hub or a wheel of the vehicle and detects positions of wheel hub bolts or positions of bolt holes of the wheel, the detection device comprising: an acquisition section that acquires basic position information of the positions of the wheel hub bolts or the positions of the bolt holes of the wheel; and a circular extraction section that extracts edges with circular elements from the imaged image, detects a plurality of first circles constituted by the edges with the circular elements, extracts a plurality of second circles from the detected plurality of first circles according to a set radius, classifies the plurality of second circles into a plurality of groups according to center positions, and selects a plurality of representative circles from the plurality of groups according to circularities, in a combination of the plurality of representative circles, the circular extraction section extracts a combination that coincides with the basic position information as a combination of the wheel hub bolts or a combination of the bolt holes to perform wheel mounting. when there is no combination that coincides with the basic position information in the combination of the plurality of representative circles, at least one imaginary circle is created, and in a combination of the plurality of representative circles and the imaginary circle, the circular extraction section extracts a combination that coincides with the basic position information as a combination of the wheel hub bolts or a combination of the bolt holes.

2. The automatic wheel mounting apparatus according to claim 1, characterized by ​

Images