A dynamic three-dimensional foot scanning method

Abstract

The present invention discloses a dynamic three-dimensional foot scanning method, which collects foot photo image data in three directions through cameras set at three positions below and on both sides of the foot, and then obtains a three-dimensional foot model through computer algorithm calculation. The method is characterized in that when collecting data, the foot and the three cameras are controlled to realize relative movement along the front-back direction, and multiple groups of camera data are collected at multiple different positions, and the three-dimensional foot model is calculated based on the multiple groups of camera data. The present invention can realize a higher-precision scan of the foot based on a low-cost camera data collection method, and at the same time greatly reduces the accuracy requirements for the foot placement position, can better meet the scanning accuracy requirements of special foot types, and has the characteristics of convenient and fast operation.

Description

Technical Field

[0001] The invention relates to a foot scanning technology, and in particular to a dynamic three-dimensional foot scanning method. Background Art

[0002] With the continuous improvement of modern living standards, people's requirements for shoes, an important "tool" that cannot be missing in daily life, are also increasing. The comfort of shoes directly affects people's travel and daily life, so choosing a pair of suitable shoes is very important.

[0003] The non-contact technology of laser and infrared light sources is relatively mature in the market, and many companies, both at home and abroad, have launched laser-based 3D foot scanners. However, the cost of laser-based 3D scanners is relatively high, and the price of a 3D laser foot scanner is usually tens of thousands of RMB. Such a high-priced foot scanner undoubtedly greatly increases the cost of custom-made shoes for customers, restricting the development of the shoemaking industry.

[0004] Although photography-based 3D foot scanners are relatively cheap, they are currently relatively low in accuracy and are particularly difficult to meet the needs of 3D scanning for special foot shapes.

[0005] For example, among the existing three-dimensional foot scanning instruments, the invention application document with the announcement number CN102034264A discloses a three-dimensional foot scanner, which includes a foot scanning device and a foot image generation system. The scanning component can take images of the inner side of the foot, the outer side of the foot and the bottom of the foot, and correct and splice the images, and extract the complete inner, outer and bottom contours of the foot, and generate a three-dimensional foot model in combination with standard three-dimensional foot data. In this technical solution, although the data acquisition speed is fast and the cost is low, it belongs to fixed-point static data acquisition. During the acquisition process, the foot needs to be placed in a precise fixed position with a correct posture, and the accuracy of the foot placement position is very high. If the foot placement position has front-to-back or left-to-right deviations or the placement angle is incorrect, it will cause large errors. At the same time, fixed-point photography is used to generate a three-dimensional foot model. In the process of converting data from two-dimensional to three-dimensional, there is too little basic two-dimensional data and it relies too much on algorithms to convert it into three-dimensional data. In particular, after scanning, a new three-dimensional foot model needs to be generated in combination with previously existing standard three-dimensional foot data. Therefore, this method will also produce large errors, resulting in the device having significantly insufficient accuracy when scanning special foot shapes.

[0006] Therefore, how to design a three-dimensional foot scanning method that can generate a foot model with high accuracy based on low-cost video and photography data collection, while having low requirements on foot position placement accuracy, has become a problem to be considered and solved by technical personnel in this field. Summary of the invention

[0007] In view of the above-mentioned deficiencies in the prior art, the technical problem to be solved by the present invention is: how to provide a dynamic three-dimensional foot scanning method that can be based on a low-cost video and photographic data acquisition method, while having low requirements on the accuracy of foot position placement and generating a foot model with high accuracy, which can better meet the scanning accuracy requirements of special foot types and is also easy and quick to operate.

[0008] In order to solve the above technical problems, the present invention adopts the following technical solutions:

[0009] A dynamic three-dimensional foot scanning method, which simultaneously collects foot photo image data in three directions through cameras set at three positions below the foot and on the inner and outer sides (i.e., left and right sides), and then obtains a three-dimensional foot model through computer algorithm calculation. The method is characterized in that during data collection, the foot and the three cameras are controlled to achieve relative movement along the front-to-back direction, and multiple groups of camera data are collected at multiple different positions, and the three-dimensional foot model is calculated based on the multiple groups of camera data.

[0010] Thus, in this method, the foot and three cameras are controlled to move relative to each other in the front-to-back direction and multiple shots are taken, and the three-dimensional model of the foot is calculated based on the multiple sets of data obtained. Compared with the method of one-time positioning and shooting, the amount of camera data is increased, which greatly improves the accuracy of the calculated three-dimensional model of the foot in different front and back position areas, reduces or even eliminates the dependence on standard three-dimensional foot data in the calculation process, reduces the requirements for the placement of the foot, can better meet the scanning accuracy requirements of special foot types, and is also easy and quick to operate.

[0011] Furthermore, during the relative motion, multiple groups of camera data at different positions are collected based on the same spacing.

[0012] In this way, it is more conducive to splicing each set of photo data to obtain a three-dimensional foot model, thereby improving the accuracy of the final three-dimensional foot model.

[0013] Furthermore, during data collection, a distance measuring sensor fixed relative to the camera is used to detect the distance position of the foot, and the three-dimensional model of the foot is corrected.

[0014] In this way, the accuracy of the final foot three-dimensional model can be further improved, and the requirements for the foot placement position can be better reduced.

[0015] The method specifically includes the following steps: 1. using three groups of binocular cameras located at the bottom and inner and outer sides of the foot, in the process of relative movement with the foot in the front-back direction, taking multiple groups of photos of the inner side, outer side and bottom of the foot at different distances, in the process of using a laser ranging sensor fixed in parallel with each group of binocular cameras and facing the foot to obtain the distance measurement values ​​between multiple points at specific positions of the foot and the laser ranging sensor; 2. using image processing methods to process and splice the images of the foot taken by each group of binocular cameras in the three directions respectively, to obtain complete appearance images of the inner side, outer side and bottom of the foot in three directions and convert them into three-dimensional images; 3. In the image obtained in step 2, the image data is decomposed into position data of several points evenly distributed in space and a point cloud is formed; at the same time, the actual spatial coordinates of the multiple points at the specific positions of the foot obtained by the laser ranging sensor in step 1 and the laser ranging sensor are used to obtain the actual spatial coordinates; the point cloud data is then error corrected using the actual spatial coordinates, and the point clouds in three directions are spliced ​​to obtain a complete point cloud of the foot; 4 Using the corrected complete point cloud data of the foot, a Poisson surface reconstruction algorithm is used to reconstruct a three-dimensional model of the foot; 5 A model smoothing algorithm is then used to obtain a smooth three-dimensional model of the foot.

[0016] In this way, after multiple groups of binocular cameras are shot, the surfaces in three directions are first spliced ​​together, and then the surfaces in three directions are used to extract the foot point clouds in three directions. The foot point clouds contain the actual laser irradiation points, and then the coordinates of the laser irradiation points are extracted from the surfaces in three directions. Then, the laser irradiation points are used to correct the point clouds of the foot according to the actual values ​​obtained by laser ranging. At the same time, the point clouds in three directions are also spliced. Then, the corrected and spliced ​​point clouds are used to reconstruct the overall three-dimensional model of the foot. Finally, the three-dimensional model of the foot is smoothed to ignore tiny details such as skin wrinkles, and the three-dimensional model of the foot used for shoe last production is obtained. Therefore, the laser ranging sensor is cleverly used to realize the correction of the camera data, which better improves the accuracy of the obtained three-dimensional model of the foot, and has the characteristics of convenient implementation and simple operation. Among them, the Poisson surface reconstruction algorithm is an existing surface image processing algorithm, which can regenerate higher-precision overall surface image data based on the original surface image data and the higher-precision surface local position data. The specific process is an existing technology and will not be described in detail here. In addition, the model smoothing algorithm is also an existing mature algorithm for achieving surface smoothing processing, and the specific process will not be described in detail here.

[0017] Furthermore, the method is implemented by the following dynamic three-dimensional foot scanner, which includes a mounting bracket, a foot pedal made of a horizontally arranged transparent material on the upper end of the mounting bracket, and a scanning system, which includes three groups of binocular cameras respectively located on the left and right sides and below the foot pedal (as the name suggests, each group of binocular cameras includes two cameras arranged in parallel), and a computer connected to the binocular cameras; wherein, it also includes a scanning bracket that is overall U-shaped, the foot pedal is located inside the scanning bracket, and the three groups of binocular cameras are respectively installed on the lower upper surface and the inner surfaces on both sides of the scanning bracket; it also includes a scanning bracket sliding mechanism for driving the scanning bracket to slide horizontally along the length direction of the foot pedal.

[0018] In this way, when the scanner is used, the binocular camera is connected to the computer for communication, and then the foot to be scanned is placed on the foot pedal. The scanning bracket sliding mechanism drives the scanning bracket and the foot pedal to slide relative to each other horizontally. During the sliding process, the binocular camera is controlled to take multiple groups of foot photos, and then the multiple groups of foot photos are stitched together using an algorithm to obtain a more accurate three-dimensional foot model than that generated by a single photo shooting method.

[0019] Further, the scanning bracket sliding mechanism includes a guide rail fixed below the mounting bracket along the length direction of the foot pedal, the bottom surface of the lower part of the scanning bracket can be slidably matched on the guide rail, and the scanning bracket sliding mechanism also includes a pull rope fixed to one end of the lower part of the scanning bracket along the length direction of the foot pedal. Further, a pull ring is fixed at the outer end of the pull rope to facilitate pulling. Further, the lower end of the scanning bracket can be slidably matched on the guide rail through a fixedly arranged slider.

[0020] In this way, when scanning is convenient, the scanning bracket can be driven to slide horizontally along the guide rail and the length direction of the foot pedal by pulling the pull rope. The structure is simple, the operation is convenient and fast, and the sliding is stable and reliable.

[0021] Furthermore, the scanning support sliding mechanism also includes a position detection module connected to the computer, the position detection module is used to detect the sliding distance of the scanning support, and give a shooting signal after each sliding fixed distance to control the scanning system to collect camera signals.

[0022] In this way, when the scanner is in use, there is no need to control the speed of the pull rope when pulling the pull rope, and the foot image signals can be captured at uniform intervals, thereby ensuring the detection accuracy and reducing the difficulty of using the device.

[0023] Furthermore, the position detection module is a grating ruler arranged in parallel with the guide rail.

[0024] Grating ruler, also known as grating ruler displacement sensor, is a measurement feedback device that can accurately control the displacement. The measurement output signal is a digital pulse, which is convenient for direct signal control. It has the advantages of high precision and fast response speed. Of course, it can also be implemented with structures with similar functions such as magnetic grating and servo motor.

[0025] Furthermore, each set of binocular cameras includes two cameras of the same model and size that are arranged side by side and facing the foot with fixed relative position parameters. To better ensure the effect of binocular imaging after each shooting, the specific existing technology based on the calculation method and principle of binocular imaging will not be described in detail here.

[0026] Furthermore, the scanning system also includes three groups of laser distance measuring sensors respectively installed on the lower upper surface and the inner surfaces of both sides of the scanning bracket and arranged facing the feet, and the laser distance measuring sensors are connected to the computer.

[0027] In this way, it is convenient to detect the actual distance position of the foot through the laser ranging sensor, perform error correction on the foot image obtained by the binocular camera, and improve the final scanning accuracy. The working principle of the laser ranging sensor is to emit a laser directly to the foot, form a light spot on the foot, and receive the laser reflection signal. The distance between the laser emission position and the light spot can be calculated according to the time. The specific details are mature existing technologies and will not be described in detail here.

[0028] Furthermore, each group of laser ranging sensors includes three sensors arranged in parallel on the same vertical plane.

[0029] In this way, sufficient detection points can be obtained to improve the correction accuracy.

[0030] Furthermore, a foot image generation system is provided in the computer, and the foot image generation system includes:

[0031] An image processing module is used to stitch multiple groups of photos of the inside, outside and bottom of the foot taken at different distances by three groups of binocular cameras in three directions to obtain complete appearance images of the inside, outside and bottom of the foot in three directions, and record the position of the foot corresponding to the laser point emitted by the laser ranging sensor during each shooting in the image;

[0032] A binocular stereo vision module is connected to the image processing module and is used to decompose the shape images of the inner side, outer side and bottom of the foot into a number of evenly distributed points (using a binocular stereo vision algorithm). The position information of all the points forms point cloud data, which contains actual laser points.

[0033] A foot laser point cloud extraction module is connected to the binocular stereo vision module and is used to calibrate (using a grayscale algorithm) the point cloud data of the actual laser points to form a point cloud set to be calibrated;

[0034] The laser ranging sensor distance value conversion coordinate value module is used to convert the distance values ​​measured by the laser ranging sensor at the specific positions of multiple feet, obtain the actual two-dimensional coordinate values ​​of the laser point emitted by the laser ranging sensor during each shooting, and convert the actual two-dimensional coordinate values ​​into spatial coordinate points, and finally form a spatial coordinate point set;

[0035] The point cloud correction module is connected to the foot laser point cloud extraction module and the laser ranging sensor distance value conversion coordinate value module, and is used to match the point cloud set with the spatial coordinate point set, assign the spatial coordinates of each point in the spatial coordinate point set to the corresponding point in the point cloud set, and then correct the data of the inner point cloud of the foot, the outer point cloud of the foot, and the bottom point cloud of the foot, and at the same time, splice the point clouds in three directions to obtain a complete point cloud of the foot;

[0036] The model generation module is connected to the point cloud correction module and is used to reconstruct a three-dimensional model of the entire foot with higher accuracy based on the corrected three-dimensional foot point cloud (Poisson surface reconstruction algorithm may be used);

[0037] The model smoothing processing module is connected to the model generation module and uses a model smoothing algorithm (ignoring tiny details such as skin wrinkles) to obtain a smooth three-dimensional foot model.

[0038] Each module can be a computer entity module or a virtual module. The virtual module implements the corresponding function by a software program installed in the computer, but each module itself is a prior art. The contribution of the above scheme to the prior art is not in the modules themselves, but in combining these existing modules in a specific logical order to more conveniently, quickly and reliably realize the scanning and generation of the three-dimensional graphics of the foot and improve its accuracy, while also reducing the requirements for foot positioning during scanning.

[0039] Furthermore, the mounting bracket includes a mounting base plate horizontally arranged below, the mounting base plate and the foot pedal are arranged opposite to each other and are both rectangular, and a supporting structure is arranged on the mounting base plate to support the foot pedal and form a rectangular body structure as a whole; the scanning bracket includes a scanning bracket base plate horizontally arranged thereon along the width direction of the mounting base plate, and a scanning bracket side panel is vertically arranged at each end of the scanning bracket base plate, and three groups of binocular cameras and corresponding laser ranging sensors are respectively installed on the upper surface of the scanning bracket base plate and the inner side of the scanning bracket side panel; a folding and folding structure is arranged between the scanning bracket side panel and the scanning bracket base plate, and the folding and folding structure can realize the horizontal flipping and vertical rotation secondary folding of the scanning bracket side panel, so that the scanning bracket side panel can be folded into the space between the mounting base plate and the foot pedal through the folding and folding structure.

[0040] This is because the camera and laser distance sensor on the scanning bracket are all precision electronic instruments. However, the cameras and laser distance sensors on both sides of the scanning bracket in the equipment protrude from the outside of the mounting bracket due to work needs, which will make the equipment inconvenient to transport and pack, and the electronic instruments on both sides are easily damaged by collision during transportation. Therefore, after adding the above structure, the side panels of the scanning bracket can be folded into the space between the mounting base and the foot pedal, so that the whole equipment forms a rectangular frame structure, which greatly facilitates the maintenance and transportation of the equipment and avoids damage to the equipment during transportation.

[0041] Furthermore, the folding and folding structure comprises two integrally rectangular connecting plates connected to the two ends of the scanning bracket bottom plate, the connecting plates and the scanning bracket bottom plate are respectively overlapped by a step surface and rotatably connected by a vertical rotating shaft, the vertical surface on the inner side of the step surface at the end of the scanning bracket bottom plate and the inner end surface of the connecting plate are abutted against each other for limiting, the inner end of the connecting plate and the side adjacent to the pull rope are rounded so that the connecting plate can be rotated 90 degrees around the vertical rotating shaft to one end of the pull rope, and the other side of the inner end of the connecting plate is a right angle and is used for rotation limiting; the lower half of the outer end surface of the connecting plate is a vertical plane for limiting, one half of the upper half of the outer end surface of the connecting plate has a first arc-shaped protrusion protruding outward and upward, and the other half The side half has a concave first arc-shaped groove portion, and the outer half of the lower end surface of the scanning bracket side plate corresponding to the outer end of the connecting plate is a horizontal plane for limiting (when the upper end of the scanning bracket side plate is flipped downward to a horizontal state, the limitation can be achieved by fitting the horizontal plane of the outer half of the lower end surface of the scanning bracket and the vertical plane of the lower half of the outer end surface of the connecting plate). One side of the inner half of the lower end surface of the scanning bracket side plate is convex to form a second arc-shaped protrusion that fits in the first arc-shaped groove portion, and the other side is concave to form a second arc-shaped groove portion that fits in the first arc-shaped protrusion. The second arc-shaped protrusion and the first arc-shaped protrusion are located at the same height and are connected in rotation by a horizontal rotating shaft.

[0042] In this way, the folding and folding structure allows the device to be easily converted from the use state to the folding state by horizontal flipping and vertical rotation. When converting, first flip the upper ends of the scanning bracket side panels on both sides downward 90 degrees to a horizontal state, and limit the position by fitting the horizontal plane of the outer half of the end surface of the lower end of the scanning bracket and the vertical plane of the lower half of the outer end surface of the connecting plate; then rotate the connecting plate toward one end of the pull rope, driving the scanning bracket side panel and the connecting plate to rotate horizontally 90 degrees together, so that the scanning bracket side panel and the electrical components thereon are folded in the space between the installation base plate and the foot pedal, which is protected and convenient for packaging and transportation, avoiding damage to the electrical components during packaging and transportation, and greatly improving the safety of equipment transportation. At the same time, the folding and folding structure is simple and ingenious, and can limit the rotation angle while realizing flipping and folding by using its own structure. When the foot scanning is working, the structure itself can be used to stop the rotation, ensuring the stability and reliability of the scanning process.

[0043] Furthermore, when the side plate of the scanning support is flipped to a horizontal state, its upper and lower surfaces are flush with the upper and lower surfaces of the connecting plate and the bottom plate of the scanning support.

[0044] In this way, it is more convenient to realize the horizontal rotation and folding of the connecting plate.

[0045] In summary, the present invention can achieve a higher-precision scan of the foot based on a low-cost video and photo data collection method, while greatly reducing the accuracy requirements for the placement of the foot, and can better meet the scanning accuracy requirements of special foot types, and also has the characteristics of convenient and fast operation. The present invention reduces the cost of a three-dimensional foot scanner, can promote the realization of three-dimensional foot data collection in each shoe retail store, and also makes it possible to collect three-dimensional foot data through door-to-door service, making it possible for customers to customize their favorite and comfortable shoes; it promotes the development of customized shoes for customers. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is a schematic structural diagram of the dynamic three-dimensional foot scanner of the present invention in a folded transportation state.

[0047] Figure 2 for Figure 1 Schematic diagram of the structure of the dynamic three-dimensional foot scanner in scanning use state.

[0048] Figure 3 for Figure 1 Schematic diagram of the structure of the separate scanning bracket part.

[0049] Figure 4 for Figure 2 Schematic diagram of the structure of the separate scanning bracket part.

[0050] Figure 5 Schematic diagram of the module structure of the foot image generation system in the computer. DETAILED DESCRIPTION

[0051] The present invention is further described in detail below in conjunction with specific implementation modes.

[0052] Specific implementation method: A dynamic three-dimensional foot scanning method, which simultaneously collects foot photo image data in three directions through cameras set at three positions below the foot and on the inner and outer sides (i.e., left and right sides), and then obtains a three-dimensional foot model through computer algorithm calculation. Its characteristics are that during data collection, the foot and the three cameras are controlled to achieve relative movement along the front-to-back direction, and multiple groups of camera data are collected at multiple different positions, and the three-dimensional foot model is calculated based on the multiple groups of camera data.

[0053] Thus, in this method, the foot and three cameras are controlled to move relative to each other in the front-to-back direction and multiple shots are taken, and the three-dimensional model of the foot is calculated based on the multiple sets of data obtained. Compared with the method of one-time positioning and shooting, the amount of camera data is increased, which greatly improves the accuracy of the calculated three-dimensional model of the foot in different front and back position areas, reduces or even eliminates the dependence on standard three-dimensional foot data in the calculation process, reduces the requirements for the placement of the foot, can better meet the scanning accuracy requirements of special foot types, and is also easy and quick to operate.

[0054] During the relative motion process, multiple sets of camera data at different positions are collected based on the same spacing.

[0055] In this way, it is more conducive to splicing each set of photo data to obtain a three-dimensional foot model, thereby improving the accuracy of the final three-dimensional foot model.

[0056] During data collection, a distance measuring sensor fixed relative to the camera is used to detect the distance position of the foot and to correct the three-dimensional model of the foot.

[0057] In this way, the accuracy of the final foot three-dimensional model can be further improved, and the requirements for the foot placement position can be better reduced.

[0058] The method specifically includes the following steps: 1. using three groups of binocular cameras located at the bottom and inner and outer sides of the foot, in the process of relative movement with the foot in the front-back direction, taking multiple groups of photos of the inner side, outer side and bottom of the foot at different distances, in the process of using a laser ranging sensor fixed in parallel with each group of binocular cameras and facing the foot to obtain the distance measurement values ​​between multiple points at specific positions of the foot and the laser ranging sensor; 2. using image processing methods to process and splice the images of the foot taken by each group of binocular cameras in the three directions respectively, to obtain complete appearance images of the inner side, outer side and bottom of the foot in three directions and convert them into three-dimensional images; 3. In the image obtained in step 2, the image data is decomposed into position data of several points evenly distributed in space and a point cloud is formed; at the same time, the actual spatial coordinates of the multiple points at the specific positions of the foot obtained by the laser ranging sensor in step 1 and the laser ranging sensor are used to obtain the actual spatial coordinates; the point cloud data is then error corrected using the actual spatial coordinates, and the point clouds in three directions are spliced ​​to obtain a complete point cloud of the foot; 4 Using the corrected complete point cloud data of the foot, a Poisson surface reconstruction algorithm is used to reconstruct a three-dimensional model of the foot; 5 A model smoothing algorithm is then used to obtain a smooth three-dimensional model of the foot.

[0059] In this way, after multiple groups of binocular cameras are shot, the surfaces in three directions are first spliced ​​together, and then the surfaces in three directions are used to extract the foot point clouds in three directions. The foot point clouds contain the actual laser irradiation points, and then the coordinates of the laser irradiation points are extracted from the surfaces in three directions. Then, the laser irradiation points are used to correct the point clouds of the foot according to the actual values ​​obtained by laser ranging. At the same time, the point clouds in three directions are also spliced. Then, the corrected and spliced ​​point clouds are used to reconstruct the overall three-dimensional model of the foot. Finally, the three-dimensional model of the foot is smoothed to ignore tiny details such as skin wrinkles, and the three-dimensional model of the foot used for shoe last production is obtained. Therefore, the laser ranging sensor is cleverly used to realize the correction of the camera data, which better improves the accuracy of the obtained three-dimensional model of the foot, and has the characteristics of convenient implementation and simple operation. Among them, the Poisson surface reconstruction algorithm is an existing surface image processing algorithm, which can regenerate higher-precision overall surface image data based on the original surface image data and the higher-precision surface local position data. The specific process is an existing technology and will not be described in detail here. In addition, the model smoothing algorithm is also an existing mature algorithm for achieving surface smoothing processing, and the specific process will not be described in detail here.

[0060] In this embodiment, the method relies on Figure 1-5 The dynamic three-dimensional foot scanner shown is implemented, and the dynamic three-dimensional foot scanner includes a mounting bracket, the upper end of which has a horizontally arranged foot pedal 1 made of transparent material, and also includes a scanning system, the scanning system includes three groups of binocular cameras 2 respectively located on the left and right sides and below the foot pedal (as the name suggests, each group of binocular cameras includes two cameras arranged in parallel), and also includes a computer connected to the binocular cameras 2 (the structure of the computer is not shown in the figure); wherein, it also includes an overall U-shaped scanning bracket, the foot pedal 1 is located inside the scanning bracket, and the three groups of binocular cameras 2 are respectively installed on the lower upper surface and the inner surfaces on both sides of the scanning bracket; it also includes a scanning bracket sliding mechanism for driving the scanning bracket to slide horizontally along the length direction of the foot pedal.

[0061] In this way, when the scanner is used, the binocular camera is connected to the computer for communication, and then the foot to be scanned is placed on the foot pedal. The scanning bracket sliding mechanism drives the scanning bracket and the foot pedal to slide relative to each other horizontally. During the sliding process, the binocular camera is controlled to take multiple groups of foot photos, and then the multiple groups of foot photos are stitched together using an algorithm to obtain a more accurate three-dimensional foot model than that generated by a single photo shooting method.

[0062] The scanning support sliding mechanism includes a guide rail 3 fixed below the mounting support along the length direction of the foot pedal, and the bottom surface of the scanning support is slidably matched on the guide rail 3. The scanning support sliding mechanism also includes a pull rope 5 fixed at one end of the lower part of the scanning support along the length direction of the foot pedal. Further, a pull ring 6 is fixed at the outer end of the pull rope 5 to facilitate pulling. Further, the lower end of the scanning support is slidably matched on the guide rail 3 through a fixed slider 7.

[0063] In this way, when scanning is convenient, the scanning bracket can be driven to slide horizontally along the guide rail and the length direction of the foot pedal by pulling the pull rope. The structure is simple, the operation is convenient and fast, and the sliding is stable and reliable.

[0064] The scanning support sliding mechanism also includes a position detection module 8 connected to the computer. The position detection module 8 is used to detect the sliding distance of the scanning support and give a shooting signal after each sliding fixed distance to control the scanning system to collect the camera signal.

[0065] In this way, when the scanner is in use, there is no need to control the speed of the pull rope when pulling the pull rope, and the foot image signals can be captured at uniform intervals, thereby ensuring the detection accuracy and reducing the difficulty of using the device.

[0066] The position detection module 8 is a grating ruler arranged in parallel with the guide rail 3 .

[0067] Grating ruler, also known as grating ruler displacement sensor, is a measurement feedback device that can accurately control the displacement. The measurement output signal is a digital pulse, which is convenient for direct signal control. It has the advantages of high precision and fast response speed. Of course, it can also be implemented with structures with similar functions such as magnetic grating and servo motor.

[0068] Each set of binocular cameras 2 includes two cameras of the same size and model that are arranged side by side and facing the feet with fixed relative position parameters. To better ensure the effect of binocular imaging after each shooting, the existing technology based on the specific calculation method and principle of binocular imaging will not be described in detail here.

[0069] The scanning system further includes three groups of laser distance measuring sensors 9 respectively mounted on the lower upper surface and the inner surfaces of both sides of the scanning bracket and arranged facing the foot. The laser distance measuring sensors 9 are connected to the computer.

[0070] In this way, it is convenient to detect the actual distance position of the foot through the laser ranging sensor, perform error correction on the foot image obtained by the binocular camera, and improve the final scanning accuracy. The working principle of the laser ranging sensor is to emit a laser directly to the foot, form a light spot on the foot, and receive the laser reflection signal. The distance between the laser emission position and the light spot can be calculated according to the time. The specific details are mature existing technologies and will not be described in detail here.

[0071] Each group of laser distance measuring sensors 9 includes three sensors arranged in parallel on the same vertical plane.

[0072] In this way, sufficient detection points can be obtained to improve the correction accuracy.

[0073] The computer is provided with a foot image generation system. The structure of the foot image generation system is shown in FIG. Figure 5 ,include:

[0074] An image processing module is used to stitch multiple groups of photos of the inside, outside and bottom of the foot taken at different distances by three groups of binocular cameras in three directions to obtain complete appearance images of the inside, outside and bottom of the foot in three directions, and record the position of the foot corresponding to the laser point emitted by the laser ranging sensor during each shooting in the image;

[0075] A binocular stereo vision module is connected to the image processing module and is used to decompose the shape images of the inner side, outer side and bottom of the foot into a number of evenly distributed points (using a binocular stereo vision algorithm). The position information of all the points forms point cloud data, which contains actual laser points.

[0076] A foot laser point cloud extraction module is connected to the binocular stereo vision module and is used to calibrate (using a grayscale algorithm) the point cloud data of the actual laser points to form a point cloud set to be calibrated;

[0077] The laser ranging sensor distance value conversion coordinate value module is used to convert the distance values ​​measured by the laser ranging sensor at the specific positions of multiple feet, obtain the actual two-dimensional coordinate values ​​of the laser point emitted by the laser ranging sensor during each shooting, and convert the actual two-dimensional coordinate values ​​into spatial coordinate points, and finally form a spatial coordinate point set;

[0078] The point cloud correction module is connected to the foot laser point cloud extraction module and the laser ranging sensor distance value conversion coordinate value module, and is used to match the point cloud set with the spatial coordinate point set, assign the spatial coordinates of each point in the spatial coordinate point set to the corresponding point in the point cloud set, and then correct the data of the inner point cloud of the foot, the outer point cloud of the foot, and the bottom point cloud of the foot, and at the same time, splice the point clouds in three directions to obtain a complete point cloud of the foot;

[0079] The model generation module is connected to the point cloud correction module and is used to reconstruct a three-dimensional model of the entire foot with higher accuracy based on the corrected three-dimensional foot point cloud (Poisson surface reconstruction algorithm may be used);

[0080] The model smoothing processing module is connected to the model generation module and uses a model smoothing algorithm (ignoring tiny details such as skin wrinkles) to obtain a smooth three-dimensional foot model for the production of shoe lasts.

[0081] Each module can be a computer entity module or a virtual module. The virtual module implements the corresponding function by a software program installed in the computer, but each module itself is a prior art. The contribution of the above scheme to the prior art is not in the modules themselves, but in combining these existing modules in a specific logical order to more conveniently, quickly and reliably realize the scanning and generation of the three-dimensional graphics of the foot and improve its accuracy, while also reducing the requirements for foot positioning during scanning.

[0082] Among them, the mounting bracket includes a mounting base plate 10 horizontally arranged at the bottom, the mounting base plate 10 and the foot pedal 1 are arranged opposite to each other and are both rectangular, and a supporting structure is arranged on the mounting base plate 10 to support the foot pedal and form a rectangular body structure as a whole; the scanning bracket includes a scanning bracket base plate 11 horizontally arranged thereon along the width direction of the mounting base plate, and a scanning bracket side plate 12 is vertically arranged at each end of the scanning bracket base plate 11, and three groups of binocular cameras and corresponding laser ranging sensors are respectively installed on the upper surface of the scanning bracket base plate 11 and the inner side of the scanning bracket side plate 12; a folding and folding structure is arranged between the scanning bracket side plate 12 and the scanning bracket base plate 11, and the folding and folding structure can realize the horizontal flipping and vertical rotation secondary folding of the scanning bracket side plate, so that the scanning bracket side plate can be folded into the space between the mounting base plate and the foot pedal through the folding and folding structure.

[0083] This is because the camera and laser distance sensor on the scanning bracket are all precision electronic instruments. However, the cameras and laser distance sensors on both sides of the scanning bracket in the equipment protrude from the outside of the mounting bracket due to work needs, which will make the equipment inconvenient to transport and pack, and the electronic instruments on both sides are easily damaged by collision during transportation. Therefore, after adding the above structure, the side panels of the scanning bracket can be folded into the space between the mounting base and the foot pedal, so that the whole equipment forms a rectangular frame structure, which greatly facilitates the maintenance and transportation of the equipment and avoids damage to the equipment during transportation.

[0084] The folding and folding structure includes two integrally rectangular connecting plates 13 connected to the two ends of the scanning support base plate. The connecting plates 13 and the scanning support base plate 11 are respectively overlapped by a step surface 14 and are rotatably connected by a vertical rotating shaft 15. The vertical surface on the inner side of the step surface at the end of the scanning support base plate 11 and the inner end surface of the connecting plate 13 are abutted against each other for limiting position. The inner end of the connecting plate 13 and the side adjacent to the pull rope 5 are rounded so that the connecting plate 13 can be rotated 90 degrees around the vertical rotating shaft toward one end of the pull rope. The other side of the inner end of the connecting plate 13 is a right angle and is used for rotation limiting position. The lower half of the outer end surface of the connecting plate 13 is a vertical plane for limiting position. One side half of the upper half of the outer end surface of the connecting plate 13 has a first arc-shaped protrusion 16 protruding outward and upward. , the other half has a concave first arc-shaped groove portion 17, the outer half of the lower end surface of the scanning support side plate 12 corresponding to the outer end of the connecting plate 13 is a horizontal plane for limiting (when the upper end of the scanning support side plate is flipped downward to a horizontal state, the limiting can be achieved by the horizontal plane of the outer half of the lower end surface of the scanning support and the vertical plane of the lower half of the outer end surface of the connecting plate), one side of the inner half of the lower end surface of the scanning support side plate 12 is convexly formed with a second arc-shaped protrusion 18 that fits in the first arc-shaped groove portion 17, and the other side is concavely formed with a second arc-shaped groove portion 19 that matches the first arc-shaped protrusion 16, the second arc-shaped protrusion 18 and the first arc-shaped protrusion 16 are located at the same height and are rotatably connected by a horizontal rotating shaft.

[0085] In this way, the folding and folding structure allows the device to be easily converted from the use state to the folding state by horizontal flipping and vertical rotation. When converting, first flip the upper ends of the scanning bracket side panels on both sides downward 90 degrees to a horizontal state, and limit the position by fitting the horizontal plane of the outer half of the end surface of the lower end of the scanning bracket and the vertical plane of the lower half of the outer end surface of the connecting plate; then rotate the connecting plate toward one end of the pull rope, driving the scanning bracket side panel and the connecting plate to rotate horizontally 90 degrees together, so that the scanning bracket side panel and the electrical components thereon are folded in the space between the installation base plate and the foot pedal, which is protected and convenient for packaging and transportation, avoiding damage to the electrical components during packaging and transportation, and greatly improving the safety of equipment transportation. At the same time, the folding and folding structure is simple and ingenious, and can limit the rotation angle while realizing flipping and folding by using its own structure. When the foot scanning is working, the structure itself can be used to stop the rotation, ensuring the stability and reliability of the scanning process.

[0086] When the scanning support side plate 12 is flipped to a horizontal state, its upper and lower surfaces are flush with the upper and lower surfaces of the connecting plate 13 and the scanning support bottom plate 11 .

[0087] In this way, it is more convenient to realize the horizontal rotation and folding of the connecting plate.

[0088] Among them, the supporting structure includes vertical beams 20 vertically arranged around the mounting base 10, and also includes cross beams 21 transversely arranged above the vertical beams at both ends of the mounting base in the length direction, and also includes a supporting frame 22 of a rectangular frame structure erected above the cross beams, and the foot pedal 1 is installed above the supporting frame.

[0089] In this way, the overall structure of the device and the support for the foot pedaling are made more stable and reliable.

[0090] In addition, when the above-mentioned device is used, after the installation is converted to the use state, each camera of the binocular camera can be calibrated separately to obtain the intrinsic parameter matrix and distortion coefficient matrix of each camera, and stored in the computer software. This step only needs to be performed once when the device is installed and fixed for the first time. Then, each binocular camera is calibrated to obtain the intrinsic matrix, basic matrix, rotation matrix and translation matrix of each binocular camera, and stored in the computer software. Finally, each binocular camera is stereo corrected separately and stored in the computer software.

[0091] In specific implementation, after all binocular cameras are calibrated, each laser ranging sensor can be further calibrated to ensure that the laser emitted by each laser ranging sensor is perpendicular to its own installation surface, and the lasers emitted by all laser ranging sensors are within one plane, and the relative positions of each laser point in the inner laser ranging sensor group, the outer laser ranging sensor group and the bottom laser ranging sensor group are measured respectively, and stored in the computer software. The computer converts the information into coordinates, and can set the coincidence point of the laser point emitted by a laser ranging sensor in the vertical direction and the laser point emitted by a laser ranging sensor in the horizontal direction as the zero point. The emission direction of the laser of the outer laser ranging sensor group can be used as the positive direction of the X axis, the emission direction of the laser of the bottom laser ranging sensor group can be used as the positive direction of the Z axis, and the direction from the starting point of the position detection module to the end point can be used as the positive direction of the Y axis to establish a spatial coordinate system.

[0092] At the beginning of scanning, the operator needs to pull the pull ring to make the scanning bracket move horizontally on the guide rail. After the operator pulls the scanning bracket, according to the feedback of the position detection module, the three groups of binocular cameras located on both sides and the bottom of the scanning bracket simultaneously take pictures and transmit the picture data to the foot image generation system of the computer. At the same time, the three groups of laser ranging sensors on both sides and the bottom of the scanning bracket of the present invention will also start ranging, and feed back their own ranging results to the foot image generation system in real time.

[0093] After the scanning starts, the three groups of binocular cameras will take pictures and the three groups of laser ranging sensor groups will measure the distance according to the feedback of the position detection module. The position detection module used in the foot scanning device of the present invention is a grating ruler. Obviously, without any effort, practitioners in this industry can easily replace the grating ruler with a magnetic grid, servo motor or other equipment with similar functions. The three groups of binocular cameras of the present invention can be executed as follows: whenever the scanning module moves 0.5mm, the three groups of binocular cameras will perform the photo action once, and transmit the picture data to the computer's foot image generation system; the three groups of laser ranging sensor groups of the present invention will also perform the ranging action as follows: whenever the scanning module moves 0.5mm, the three groups of laser ranging sensor groups will perform the ranging action once, and their own ranging results will also be fed back to the foot image generation system in real time. And the high-precision three-dimensional model of the foot is finally generated by calculation in the foot image generation system.

Claims

1. A dynamic three-dimensional foot scanning method, which uses cameras set at three locations below the foot and at the inner and outer sides to simultaneously collect foot photo image data in three directions, and then obtains a three-dimensional foot model through computer algorithm calculation, characterized in that: During data collection, the foot and three cameras are controlled to move relative to each other in the front-to-back direction, and multiple sets of camera data are collected at multiple different positions, and a three-dimensional model of the foot is calculated based on the multiple sets of camera data; During the relative motion, multiple sets of camera data at different positions are collected based on the same spacing; The method is implemented by a dynamic three-dimensional foot scanner, which includes a mounting bracket, a foot pedal made of a horizontally arranged transparent material on the upper end of the mounting bracket, and a scanning system, which includes three groups of binocular cameras respectively located on the left and right sides and below the foot pedal, and a computer connected to the binocular cameras; wherein, it also includes a scanning bracket that is overall U-shaped, the foot pedal is located inside the scanning bracket, and the three groups of binocular cameras are respectively installed on the lower upper surface and the inner surfaces on both sides of the scanning bracket; and also includes a scanning bracket sliding mechanism for driving the scanning bracket to slide horizontally along the length direction of the foot pedal; The scanning bracket sliding mechanism includes a guide rail fixed below the mounting bracket along the length direction of the foot pedal, the bottom surface of the lower part of the scanning bracket is slidably matched on the guide rail, and the scanning bracket sliding mechanism also includes a pull rope fixed at one end of the lower part of the scanning bracket along the length direction of the foot pedal; The scanning bracket sliding mechanism also includes a position detection module connected to the computer, the position detection module is used to detect the sliding distance of the scanning bracket, and give a shooting signal after each sliding fixed distance to control the scanning system to collect camera signals; the position detection module is a grating ruler arranged in parallel with the guide rail; The mounting bracket comprises a mounting base plate horizontally arranged at the bottom, the mounting base plate and the foot pedal are arranged opposite to each other and are both rectangular, a supporting structure is arranged on the mounting base plate to support the foot pedal and form a rectangular body structure as a whole; the scanning bracket comprises a scanning bracket base plate horizontally arranged thereon along the width direction of the mounting base plate, a scanning bracket side plate is vertically arranged at each end of the scanning bracket base plate, three groups of binocular cameras and corresponding laser ranging sensors are respectively installed on the upper surface of the scanning bracket base plate and the inner side of the scanning bracket side plate; a folding and folding structure is arranged between the scanning bracket side plate and the scanning bracket base plate, and the folding and folding structure can realize the horizontal flipping and vertical rotation secondary folding of the scanning bracket side plate, so that the scanning bracket side plate can be folded into the space between the mounting base plate and the foot pedal through the folding and folding structure; The folding and folding structure comprises two integrally rectangular connecting plates connected to the two ends of the scanning support base plate. The connecting plates and the scanning support base plate are respectively overlapped by a step surface and are connected by a vertical rotating shaft. The vertical surface on the inner side of the step surface at the end of the scanning support base plate and the inner end surface of the connecting plate are abutted against each other for limiting position. The inner end of the connecting plate and the side adjacent to the pull rope are rounded so that the connecting plate can be rotated 90 degrees around the vertical rotating shaft toward one end of the pull rope. The other side of the inner end of the connecting plate is a right angle and is used for rotation limiting position. The lower half of the end surface of the outer end of the connecting plate is a vertical plane for limiting position. One side half of the upper half of the end surface has a first arc-shaped protrusion protruding outward and upward, and the other side half has a first arc-shaped groove portion concave inwardly. The outer half of the lower end surface of the scanning bracket side plate corresponding to the outer end of the connecting plate is a horizontal plane for limiting. One side of the inner half of the lower end surface of the scanning bracket side plate is convexly formed with a second arc-shaped protrusion that fits in the first arc-shaped groove portion, and the other side is concavely formed with a second arc-shaped groove portion that fits in the first arc-shaped protrusion. The second arc-shaped protrusion and the first arc-shaped protrusion are located at the same height and are connected in rotation by a horizontal rotating shaft.

2. The dynamic three-dimensional foot scanning method according to claim 1, characterized in that: During data collection, a distance measuring sensor fixed relative to the camera is used to detect the distance position of the foot and to calibrate the three-dimensional model of the foot.

3. The dynamic three-dimensional foot scanning method according to claim 2, characterized in that: The specific steps include:

1. Using three sets of binocular cameras located at the bottom and inside and outside of the foot, in the process of relative movement with the foot in the front-back direction, multiple sets of photos of the inside, outside and bottom of the foot are taken at different distances. In this process, a laser ranging sensor fixed in parallel with each set of binocular cameras and facing the foot is used to obtain the distance measurement values ​​between multiple points at specific positions of the foot and the laser ranging sensor; 2. Using image processing methods, the images of the foot taken by each set of binocular cameras in the three directions are processed and spliced ​​separately to obtain complete appearance images of the inside, outside and bottom of the foot and convert them into three-dimensional images; 3. The images obtained in step 2 are In the image, the image data is decomposed into position data of several points evenly distributed in space and a point cloud is formed; at the same time, the actual space coordinates of the points at the specific positions of the multiple feet obtained by the laser ranging sensor in step 1 and the laser ranging sensor are used to obtain the actual space coordinates; the point cloud data is then error corrected using the actual space coordinates, and the point clouds in three directions are spliced ​​to obtain a complete point cloud of the foot; 4 using the corrected complete point cloud data of the foot, a Poisson surface reconstruction algorithm is used to reconstruct and generate a three-dimensional model of the foot; 5 a model smoothing algorithm is used to obtain a smooth three-dimensional model of the foot.

4. The dynamic three-dimensional foot scanning method according to claim 1, characterized in that: The scanning system also includes three groups of laser distance measuring sensors respectively mounted on the lower upper surface and the inner surfaces of both sides of the scanning bracket and arranged facing the foot, and the laser distance measuring sensors are connected to the computer; Each group of laser distance measuring sensors includes three laser distance measuring sensors arranged in parallel on the same vertical plane.

5. The dynamic three-dimensional foot scanning method according to claim 4, characterized in that: The computer is provided with a foot image generation system, which includes: An image processing module is used to stitch multiple groups of photos of the inside, outside and bottom of the foot taken at different distances by three groups of binocular cameras in three directions to obtain complete appearance images of the inside, outside and bottom of the foot in three directions, and record the position of the foot corresponding to the laser point emitted by the laser ranging sensor during each shooting in the image; A binocular stereo vision module is connected to the image processing module and is used to decompose the appearance images of the inner side, outer side and bottom of the foot into a number of evenly distributed points. The position information of all the points forms point cloud data, which contains actual laser points. A foot laser point cloud extraction module is connected to the binocular stereo vision module and is used to calibrate and extract the point cloud data of the actual laser points to form a point cloud set to be calibrated; The laser ranging sensor distance value conversion coordinate value module is used to convert the distance values ​​measured by the laser ranging sensor at the specific positions of multiple feet, obtain the actual two-dimensional coordinate values ​​of the laser point emitted by the laser ranging sensor during each shooting, and convert the actual two-dimensional coordinate values ​​into spatial coordinate points, and finally form a spatial coordinate point set; The point cloud correction module is connected to the foot laser point cloud extraction module and the laser ranging sensor distance value conversion coordinate value module, and is used to match the point cloud set with the spatial coordinate point set, assign the spatial coordinates of each point in the spatial coordinate point set to the corresponding point in the point cloud set, and then correct the data of the inner point cloud of the foot, the outer point cloud of the foot, and the bottom point cloud of the foot, and at the same time, splice the point clouds in three directions to obtain a complete point cloud of the foot; The model generation module is connected to the point cloud correction module and is used to reconstruct a three-dimensional model of the entire foot with higher accuracy based on the corrected three-dimensional foot point cloud; The model smoothing processing module is connected to the model generation module and uses a model smoothing algorithm to obtain a smooth three-dimensional foot model.

Images