Method, device and equipment for guiding crochet hook looping through three-dimensional depth measurement

Abstract

The invention provides a method, device and equipment for guiding crochet hook looping through three-dimensional depth measurement, and the method comprises the steps: obtaining a knitting instruction, the knitting instruction comprises a looping path of a knitted fabric in a current knitting layer, a path point in the looping path represents a crochet hook position of the current knitting layer, and a coil connection type under the crochet hook position; if the current knitted layer is not the first knitted layer, generating a three-dimensional point cloud of the surface of the knitted fabric based on three-dimensional depth measurement; based on the position of the previous coil, the position of the next crochet hook and the three-dimensional point cloud of the surface of the knitted fabric, the form and the position of a to-be-operated coil are estimated, and the to-be-operated coil is the coil, needing to be strung and sleeved in the current crochet hook looping process, of the previous knitted layer; according to the form and the position of the to-be-operated coil and the coil connection type of the next crochet needle position, a motion path of next crochet needle looping is generated, and the knitted fabric or the target crochet needle is controlled to move according to the motion path so as to complete the next crochet needle looping operation. The knitted fabric in any shape can be woven.

Description

Technical Field

[0001] The present invention relates to the field of software technology, and more specifically, to a method, device and equipment for guiding loop formation of a crochet hook by three-dimensional depth measurement. Background Art

[0002] Three-dimensional solid knitting refers to using knitting technology to knit a three-dimensional solid object corresponding to a given three-dimensional model of any shape. In three-dimensional solid knitting, a basic action is to use a crochet hook to achieve the stringing and connection of loops in three-dimensional (X, Y, and Z) coordinate directions, so that the knitted fabric has loop connections in three dimensions and can knit a three-dimensional solid object of any shape.

[0003] Therefore, how to achieve three-dimensional solid knitting of any shape with loop stringing connections in three dimensions is an urgent problem to be solved in the current three-dimensional knitting field. Summary of the Invention

[0004] In view of this, to solve the above problems, the present invention provides a method, device and equipment for guiding loop formation of a crochet hook by three-dimensional depth measurement, and the technical solutions are as follows:

[0005] A method for guiding loop formation of a crochet hook by three-dimensional depth measurement, the method for guiding loop formation of a crochet hook by three-dimensional depth measurement includes:

[0006] Obtain a knitting instruction, where the knitting instruction includes the loop formation path of the knitted fabric in the current knitting layer, and the path points in the loop formation path represent the crochet hook positions in the current knitting layer and the loop connection types under the crochet hook positions;

[0007] If the current knitting layer is not the first knitting layer, generate a three-dimensional point cloud on the surface of the knitted fabric based on three-dimensional depth measurement;

[0008] Estimate the shape and position of the coil to be operated based on the position of the previous loop, the position of the next crochet hook, and the three-dimensional point cloud on the surface of the knitted fabric, where the coil to be operated is the loop of the previous knitting layer that needs to be strung during the loop formation process of the current crochet hook;

[0009] Generate a movement path for the next crochet hook to form a loop according to the shape and position of the coil to be operated and the loop connection type of the next crochet hook position, and control the knitted fabric or the target crochet hook to move according to the movement path to complete the operation of forming a loop with the next crochet hook.

[0010] Preferably, the generating a three-dimensional point cloud on the surface of the knitted fabric based on three-dimensional depth measurement includes:

[0011] Generate an original three-dimensional point cloud on the surface of the knitted fabric based on a three-dimensional depth sensor;

[0012] Perform processing operations on the original three-dimensional point cloud, where the processing operations include one or more of point cloud fusion, point cloud hole filling, and filtering.

[0013] Preferably, generating the three-dimensional point cloud of the knitted fabric surface based on three-dimensional depth measurement includes:

[0014] Obtain the layer height coefficient where the current knitting layer is located;

[0015] If the layer height coefficient meets the preset measurement conditions, generate the three-dimensional point cloud of the knitted fabric surface for this time based on three-dimensional depth measurement;

[0016] If the layer height coefficient does not meet the preset measurement conditions, retrieve the three-dimensional point cloud of the knitted fabric surface generated based on three-dimensional depth measurement last time.

[0017] Preferably, estimating the shape and position of the coil to be operated based on the position of the previous coil, the position of the next hook needle, and the three-dimensional point cloud of the knitted fabric surface includes:

[0018] Determine the coil to be operated according to the position of the previous coil and the position of the next hook needle;

[0019] Extract the three-dimensional point clouds of each coil from the three-dimensional point cloud of the knitted fabric surface;

[0020] Obtain the shape of the wire material of the knitted fabric, and fit the three-dimensional point cloud of the coil to be operated according to the shape of the wire material to obtain the shape and position of the coil to be operated.

[0021] Preferably, extracting the three-dimensional point clouds of each coil from the three-dimensional point cloud of the knitted fabric surface includes:

[0022] Determine the pixel points on the three-dimensional point cloud of the knitted fabric surface where the gradient depths in all directions are not in a decreasing state, and search for the three-dimensional point clouds of the pixels belonging to the same coil on the knitted fabric surface with the three-dimensional point clouds of the determined pixel points as the initial values; or

[0023] Obtain the two-dimensional image of the knitted fabric surface, and input the three-dimensional point cloud of the knitted fabric surface and the two-dimensional image into a pre-trained segmentation model to determine the three-dimensional point clouds of each coil on the knitted fabric surface.

[0024] Preferably, the method for guiding the hook needle to form a loop by three-dimensional depth measurement further includes:

[0025] Obtain the moving progress of the knitted fabric or the target hook needle, and verify the process of the next hook needle forming a loop when the moving progress meets the preset verification conditions.

[0026] Preferably, verifying the process of the next hook needle forming a loop includes:

[0027] Obtain the force feedback value of the next crochet loop formation, where the force feedback value represents the feedback force from the knitted fabric borne by the target crochet hook or the yarn feeding mechanism;

[0028] Verify the process of the next crochet loop formation according to the force feedback value.

[0029] Preferably, verifying the process of the next crochet loop formation further includes:

[0030] If the verification result is that the loop formation fails, obtain the number of failures of the next crochet loop formation, and when the number of failures is less than a preset number threshold, return to execute the step of generating the three-dimensional point cloud of the surface of the knitted fabric based on the three-dimensional depth measurement.

[0031] A device for guiding crochet loop formation by three-dimensional depth measurement, the device for guiding crochet loop formation by three-dimensional depth measurement includes:

[0032] An instruction acquisition module, configured to acquire a knitting instruction, where the knitting instruction includes the loop formation path of the knitted fabric in the current knitting layer, and the path points in the loop formation path represent the crochet hook positions in the current knitting layer and the coil connection types under the crochet hook positions;

[0033] A point cloud generation module, configured to generate the three-dimensional point cloud of the surface of the knitted fabric based on the three-dimensional depth measurement if the current knitting layer is not the first knitting layer;

[0034] A crochet loop formation module, configured to estimate the shape and position of the coil to be operated based on the position of the previous coil, the position of the next crochet hook, and the three-dimensional point cloud of the surface of the knitted fabric, where the coil to be operated is the coil of the previous knitting layer that needs to be strung during the current crochet loop formation process; generate the movement path of the next crochet loop formation according to the shape and position of the coil to be operated and the coil connection type of the next crochet hook position, and control the knitted fabric or the target crochet hook to move according to the movement path to complete the operation of the next crochet loop formation.

[0035] A device, the device includes: at least one memory and at least one processor; the memory stores an application program, and the processor calls the application program stored in the memory, and the application program is used to implement the method for guiding crochet loop formation by three-dimensional depth measurement.

[0036] Compared with the prior art, the beneficial effects achieved by the present invention are:

[0037] The present invention provides a method, device and equipment for guiding loop formation of a crochet hook by three-dimensional depth measurement, which obtains a knitting instruction. The knitting instruction includes the loop formation path of the knitted fabric in the current knitting layer, and the path points in the loop formation path represent the position of the crochet hook in the current knitting layer and the type of loop connection under the position of the crochet hook. If the current knitting layer is not the first knitting layer, a three-dimensional point cloud of the surface of the knitted fabric is generated based on three-dimensional depth measurement. Based on the position of the previous loop, the position of the next crochet hook and the three-dimensional point cloud of the surface of the knitted fabric, the shape and position of the loop to be operated are estimated. The loop to be operated is the loop of the previous knitting layer that needs to be strung during the loop formation process of the current crochet hook. According to the shape and position of the loop to be operated and the type of loop connection of the next crochet hook position, a movement path for the next crochet hook to form a loop is generated, and the knitted fabric or the target crochet hook is controlled to move according to the movement path to complete the operation of the next crochet hook to form a loop. The present invention guides loop formation of a crochet hook based on three-dimensional depth measurement, is not limited by the number of crochet hooks, can realize three-dimensional solid knitting of any shape with loop connections in three-dimensional directions, can knit knitted fabrics of any shape, greatly simplifies the product structure and reduces the product size. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only the embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained according to the provided drawings.

[0039] Figure 1 It is a flowchart of the method for guiding loop formation of a crochet hook by three-dimensional depth measurement provided by an embodiment of the present invention;

[0040] Figure 2 It is a system structure diagram of the three-dimensional depth measurement guiding crochet hook loop formation provided by an embodiment of the present invention;

[0041] Figure 3 It is a schematic diagram of the surface of an uneven knitted fabric provided by an embodiment of the present invention;

[0042] Figure 4 It is a measurement point distribution diagram of the three-dimensional point cloud of the surface of a coil provided by an embodiment of the present invention;

[0043] Figure 5 It is a schematic diagram of annotation based on deep learning segmentation provided by an embodiment of the present invention;

[0044] Figure 6 It is a schematic diagram of the segmentation effect of the segmentation model provided by an embodiment of the present invention;

[0045] Figure 7Schematic diagram of the device for guiding loop formation by a crochet hook with three-dimensional depth measurement provided by an embodiment of the present invention. Detailed implementation manners

[0046] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0047] For the convenience of understanding the present invention, the following first explains the related concepts involved in the present invention:

[0048] Loop formation: The action process of forming a loop with a wire.

[0049] Three-dimensional depth measurement: Based on a camera sensor to obtain depth information to estimate the X, Y, and Z three-dimensional coordinate values of all points or key points on the surface of an object.

[0050] Loop connection type: The common two-dimensional knitted fabric loop connection types are weft knitting and warp knitting. Among them, weft knitted fabrics include plain weft-knitted fabric, rib fabric, purl fabric, double rib fabric, etc.; warp knitted fabrics include chain stitch fabric, plain warp-knitted fabric, satin stitch fabric, etc.; in addition, there are various loop forming and connection methods corresponding to hand crochet knitting methods, such as long stitch, half double crochet, single crochet, slip stitch, etc.

[0051] To make the above objects, features, and advantages of the present invention more obvious and understandable, the present invention will be further described in detail below in conjunction with the accompanying drawings and specific implementation manners.

[0052] See Figure 1 , Figure 1 which is the method flowchart of the method for guiding loop formation by a crochet hook with three-dimensional depth measurement provided by an embodiment of the present invention. As Figure 1 shown, the method for guiding loop formation by a crochet hook with three-dimensional depth measurement includes the following steps:

[0053] S10. Obtain a knitting instruction, where the knitting instruction includes the loop formation path of the knitted fabric in the current knitting layer, and the path points in the loop formation path represent the crochet hook positions in the current knitting layer and the loop connection type under the crochet hook positions.

[0054] See Figure 2 , Figure 2 which is the system structure diagram of the system for guiding loop formation by a crochet hook with three-dimensional depth measurement provided by an embodiment of the present invention. As Figure 2As shown in the figure, the system for guiding loop formation of a three-dimensional depth measurement crochet hook includes a device 10 that executes the method for guiding loop formation of a three-dimensional depth measurement crochet hook provided by the present invention and a host computer 20. The device 10 executes operations including system initialization, knitting initialization, loop formation operation of the first knitting layer, loop formation operation starting from the second knitting layer, moving the transmission mechanism to a safe position, etc. The host computer 20 executes operations such as three-dimensional model management and knitting instruction generation. Among them, the first knitting layer is the first knitting layer of the knitted fabric, and the second knitting layer is other layers of the knitted fabric except the first knitting layer. It should be noted that the types of the host computer 20 include but are not limited to cloud servers, local computing terminals, other CPU cores of the device 10, etc., and the embodiments of the present invention do not limit this.

[0055] The host computer 20 is responsible for three-dimensional model management of the knitted fabric, and knitting instruction generation is a process of cutting the three-dimensional model of the knitted fabric into a series of two-dimensional planes (also called knitting layers) according to a specified thickness, and planning loop formation paths and knitting instructions for each plane. Each plane is each knitting layer. The loop formation path is composed of a series of discrete path points. The discrete path points include the crochet hook positions of the knitting layer where they are located and the type of loop connection at the crochet hook positions. When the host computer 20 plans the loop formation path, it determines the crochet hook positions and loop connection types of the path points according to the set optimization goals. The set optimization goals include but are not limited to the isotropy of loop connection forces, the flatness of the knitted fabric surface, etc. All path points are stored in a data structure according to the topological structure of each loop and the order is marked.

[0056] When the device 10 initializes the system, it will start a task thread that communicates with the host computer 20. This thread will maintain a connection with the host computer 20 and receive knitting instructions and new knitting tasks from the host computer 20 at any time. When the device 10 receives a new knitting task, it will stop the current task, start executing the new task through knitting initialization, and then receive the planned loop formation path from the host computer 20 and store it in the cache for use in the loop formation operation of each knitting layer. The system initialization of the device 10 includes the reset or configuration of the control circuit, motor drive module, transmission mechanism, system parameters, system status, etc. The knitting initialization of the device 10 includes but is not limited to the reset of the transmission mechanism, parameter configuration, status detection and recovery, etc. The loop formation operation of the first knitting layer of the device 10 is a traditional two-dimensional loop formation process with only X and Y coordinate connections. The loop formation operation of the device 10 starting from the second knitting layer is a process based on three-dimensional depth measurement to guide the crochet hook to form loops. After the loop formation operations of all knitting layers are completed, the device 10 moves the transmission mechanism to a safe position to avoid safety accidents.

[0057] Specifically, since the loop-forming process of the first knitting layer is a two-dimensional loop-forming process, it can be completed by using a mechanical transmission mechanism to string and form loops, without the need to rely on three-dimensional depth measurement for loop formation. Therefore, the present invention can design a mechanical transmission mechanism to achieve the loop-forming process of the first layer. Starting from the second knitting layer, the loop-forming process of each knitting layer needs to use three-dimensional depth measurement to identify and guide the stringing connection of the loops of the previous knitting layer and the current knitting layer, so as to achieve three-dimensional connection in the X, Y, and Z coordinate directions. The present invention adopts a method of using three-dimensional depth measurement to guide the crochet hook to form loops, and focuses on the loop-forming operation starting from the second knitting layer. The loop-forming process of the first knitting layer is not described in detail herein.

[0058] In this regard, after receiving the knitting instruction output by the host computer, the device 10 can obtain the loop-forming path of the knitted fabric in the current knitting layer from it. This loop-forming path is composed of discrete path points, and each path point represents the crochet hook position of the knitted fabric in the current knitting layer and the loop connection type at this crochet hook position.

[0059] In addition, in practical applications, the device 10 performs initialization of the current knitting layer, that is, reads the knitting instruction from the instruction buffer received from the host computer 20 for initialization, including setting whether to perform subsequent loop-forming control verification, loop-forming verification strategy, loop-forming failure handling strategy, post-processing parameters, setting the initial value of the current crochet hook position, etc.

[0060] S20, if the current knitting layer is not the first knitting layer, generate a three-dimensional point cloud of the surface of the knitted fabric based on three-dimensional depth measurement.

[0061] In the embodiment of the present invention, if the current knitting layer is not the first knitting layer, but any knitting layer starting from the second knitting layer, the crochet hook can be guided to form loops based on three-dimensional depth measurement. Here, a three-dimensional point cloud of the surface of the knitted fabric can be generated based on three-dimensional depth measurement.

[0062] Specifically, the sensors used for three-dimensional depth measurement include but are not limited to structured light cameras, stereo vision cameras, TOF sensors (such as laser range sensors, ultrasonic range sensors, infrared range sensors), etc., which are sensors that can obtain depth or distance data. Using these sensors, the depth of the surface of the knitted fabric can be measured. Coupled with the estimated or known sensor pose information, the three-dimensional point cloud of the surface of the knitted fabric can be calculated. The specific principles for calculating the three-dimensional point cloud of the surface of the knitted fabric include but are not limited to the following two:

[0063] 1) The structured light camera and the stereo vision camera calculate the three-dimensional point cloud of the knitted fabric surface. Both the structured light camera and the stereo vision camera use multi-camera systems. Every two cameras form a binocular camera. The following takes the binocular camera as an example to illustrate its principle. The camera is installed on the robotic arm, or on other transmission mechanisms, or on a fixed mechanical structure. The two cameras of the binocular camera are usually placed at a certain distance, which is called the baseline distance. The two cameras are relatively fixed and capture the same scene simultaneously, forming two two-dimensional images. When the same target (specifically, the surface of the knitted fabric in the embodiments of the present invention) appears in both the left and right images at the same time, due to the baseline distance between the cameras, its position in the two images will shift. This shift is called parallax. The parallax can be obtained by calculating the distance difference between the corresponding pixel points in the two images. Through the parallax, the distance difference between the corresponding pixel points of the same target in the two images can be obtained. At the same time, with known parameters such as the baseline distance of the two cameras, the three-dimensional coordinates of the target can be calculated through the principle of triangulation, and then the three-dimensional point cloud can be obtained through data conversion.

[0064] 2) The TOF sensor calculates the three-dimensional point cloud of the knitted fabric surface. The TOF sensor is installed on the robotic arm, or on other transmission mechanisms, or on a fixed mechanical structure. TOF (Time of flight) is literally translated as "time of flight". Its ranging principle is to continuously send light pulses to the target, and then use the sensor to receive the light returned from the target. The distance to the target object is obtained by detecting the flight (round-trip) time of the light pulse. This technology is basically similar to the principle of the three-dimensional laser sensor, except that the three-dimensional laser sensor scans point by point, while the TOF camera obtains the depth (distance) information of the entire image at the same time. Thus, the three-dimensional coordinates of the target can be obtained, and then the three-dimensional point cloud can be obtained through data conversion.

[0065] Since the wire diameter can be as small as less than 1 mm, for the reliability of the crochet loop formation, the depth measurement accuracy of the knitted fabric surface is required to be within 0.1 mm. In actual use, when measuring the depth values of at least 20 points on a 1-mm length, when the depth estimation error is within one measurement point spacing, the depth measurement error is within 0.05 mm. See Figure 3 , Figure 3 is a schematic diagram of the uneven knitted fabric surface provided by the embodiments of the present invention. As Figure 3 shown, due to the unevenness and irregularity of the surface of the flexible wire, the depth measurement method of the knitted fabric surface is required to have strong robustness. Generally, the depth measurement accuracy of a flat surface can reach within 0.01 mm, but the depth measurement accuracy of an uneven surface fluctuates greatly. For this, in the embodiments of the present invention, step S20 "generating a three-dimensional point cloud of the knitted fabric surface based on three-dimensional depth measurement" can adopt the following steps to improve the reliability of depth measurement:

[0066] Generate the original three-dimensional point cloud of the knitted fabric surface based on a three-dimensional depth sensor; perform processing operations on the original three-dimensional point cloud, and the processing operations include one or more of point cloud fusion, point cloud hole filling, and filtering.

[0067] In the embodiments of the present invention, for the three-dimensional point cloud of the knitted fabric surface output by three-dimensional depth sensors such as structured light cameras, stereo vision cameras, TOF sensors, etc., as the original three-dimensional point cloud, one or more processing operations of point cloud fusion, hole filling, and filtering can be performed on it to improve the accuracy and stability of the three-dimensional point cloud. See Figure 4 , Figure 4 is a measurement point distribution diagram of the three-dimensional point cloud of the surface of a single loop provided by the embodiments of the present invention, Figure 4 the wire diameter of the middle loop is 2.5 mm, Figure 4 is at Figure 3 Based on the three-dimensional point cloud of the surface of a loop that meets the accuracy requirements measured. The following will separately explain point cloud fusion, hole filling, and filtering:

[0068] 1) Point cloud fusion. When the camera takes pictures, it is inevitable to generate some noise points. Fusing the three-dimensional point clouds obtained by taking pictures multiple times can make up for the deficiencies in point cloud accuracy and point cloud density and improve the calculation accuracy.

[0069] 2) Hole filling. Due to the uneven surface of the flexible wire, when structured light is projected onto it, it may cause uneven brightness or partial slight occlusion, resulting in the appearance of some holes. Using the point cloud hole filling method can reduce the appearance of holes. Combining the point cloud around the hole point cloud and the two-dimensional image near the hole, use traditional methods such as fitting the surface around the hole and then interpolating the point cloud of the missing part of the surface, and use deep learning methods (such as methods like ProxyFormer) for hole filling.

[0070] 3) Filtering. Use a method based on the number of neighborhood points or statistics to remove noise points and improve the stability of the three-dimensional point cloud.

[0071] S30, Based on the position of the previous loop, the position of the next crochet hook, and the three-dimensional point cloud of the knitted fabric surface, estimate the shape and position of the coil to be operated, and the coil to be operated is the coil of the previous knitting layer that needs to be strung during the current crochet hook loop formation process.

[0072] In the embodiments of the present invention, the surface of the knitted fabric is composed of coils formed by the mutual looping and connection of wires. Therefore, the three-dimensional point cloud on the surface of the knitted fabric is the three-dimensional point cloud on the surface of the coils. Since the wires may be flexible and each coil is subjected to different forces, the three-dimensional shapes and positions of the coils are different. Based on the given wire shape and the three-dimensional point cloud on the surface of the knitted fabric, the three-dimensional shape and position of the coil to be operated can be estimated. The given wire shape is generally a cylindrical shape with a given diameter, or it can also be a square or other shape. In the following, the present invention will take the cylindrical shape as an example for illustration. It should be noted that the coil to be operated is the coil of the previous knitting layer that has been looped and needs to be looped by the current crochet hook during the looping process.

[0073] In the specific implementation process, step S30, "estimating the shape and position of the coil to be operated based on the position of the previous coil, the position of the next crochet hook, and the three-dimensional point cloud on the surface of the knitted fabric", can be carried out by the following steps:

[0074] Determine the coil to be operated according to the position of the previous coil and the position of the next crochet hook; extract the three-dimensional point clouds of each coil from the three-dimensional point cloud on the surface of the knitted fabric; obtain the wire shape of the knitted fabric, and fit the three-dimensional point cloud of the coil to be operated according to the wire shape to obtain the shape and position of the coil to be operated.

[0075] In the embodiments of the present invention, a traditional three-dimensional point cloud or a method based on deep learning segmentation can be used to determine the coil to be operated according to the position of the coil of the previous loop formation and the position of the next crochet hook (i.e., the crochet hook position of the coil to be looped), and then extract the three-dimensional point clouds of each coil from the three-dimensional point cloud on the surface of the knitted fabric. Among them, the three-dimensional point cloud of the coil to be operated is included. Further, the three-dimensional point cloud of the coil to be operated is fitted according to the wire shape of the knitted fabric to obtain the shape and position of the coil to be operated. In the embodiments of the present invention, the shape refers to the shapes of different coils under the stress state.

[0076] Specifically, when extracting the three-dimensional point clouds of each coil from the three-dimensional point cloud on the surface of the knitted fabric, the following two methods can be used:

[0077] 1) Determine the pixel points on the three-dimensional point cloud on the surface of the knitted fabric where the gradient depths in all directions are not in a decreasing state, and use the three-dimensional point cloud of the determined pixel points as the initial value to search for the three-dimensional point clouds of the pixel points belonging to the same coil on the surface of the knitted fabric.

[0078] In the embodiments of the present invention, a traditional three-dimensional point cloud method can be used to determine the pixel points on the three-dimensional point cloud on the surface of the knitted fabric where the depth gradients in all directions do not decrease. These pixel points are local minimum points. The window size for calculating the direction gradient can be set according to the wire diameter in proportion, and the size of the proportion is determined according to the actual test results; start searching for the three-dimensional point clouds of the pixel points belonging to the same coil from the pixel points of the local minimum points.

[0079] 2) Obtain a two-dimensional image of the knitted fabric surface, and input the three-dimensional point cloud and the two-dimensional image of the knitted fabric surface into a pre-trained segmentation model to determine the three-dimensional point cloud of each loop on the knitted fabric surface.

[0080] In the embodiments of the present invention, the three-dimensional point clouds of loops with different shapes and colors can be collected in advance, and the edge positions of the loops in the three-dimensional point cloud and the two-dimensional image are segmented and labeled. Refer to Figure 5 , Figure 5 which is a schematic diagram of the annotation based on deep learning segmentation provided by the embodiments of the present invention. Input the three-dimensional point cloud, the two-dimensional image, and the annotation data of the two-dimensional image into a deep neural network for training to obtain a segmentation model for segmenting loops. The deep neural network used for training the segmentation model includes but is not limited to networks such as Mask R-CNN, UNet, FCN, DeepLabV3+, PointNet, PointNet++, PCT (Point Cloud Transformer), JSNet, etc., and the embodiments of the present invention do not make any limitations in this regard. Refer to Figure 6 , Figure 6 which is a schematic diagram of the segmentation effect of the segmentation model provided by the embodiments of the present invention. Through the segmentation model, all the loops in the three-dimensional point cloud of the knitted fabric can be segmented to obtain the three-dimensional point cloud of each loop.

[0081] On the basis of obtaining the three-dimensional point cloud of each loop, the pixel points belonging to the same loop can be fitted according to the wire shape of the knitted fabric to determine the shape and position of the loop center line, that is, the shape and position of the loop to be operated. Since the loop formation by the crochet needle is continuously executed in sequence for the planned path points, the position of the next loop to be operated can be predicted according to the current position of the crochet needle, and only the loops within the predicted range are estimated for their three-dimensional shapes and positions. The loop to be operated refers to the previously knitted loop on the upper layer that needs to be strung during the current loop formation process of the crochet needle in the planned path.

[0082] In addition, based on the implementation of the above method based on deep learning segmentation, in some scenarios, the shape and position of the loop to be operated can also be detected in an end-to-end manner based on deep learning. Specifically, the method based on deep learning segmentation divides the position recognition process of the loop to be operated into two steps: segmentation and post-processing, and these two steps can be combined into one network by means of deep learning to improve the overall robustness and speed. For one-time loop position recognition, only the loops near the previous loop formation position need to be concerned, and it is not necessary to process all the loops in the point cloud after segmentation. The specific process of the end-to-end manner based on deep learning is as follows:

[0083] For each acquired three-dimensional point cloud, annotate the shape and position of the coil to be operated, and record the position of the next crochet hook in the knitting instruction. Take the three-dimensional point cloud, two-dimensional image, the position of the previous coil, and the position of the next crochet hook as four-way inputs and input them into the network simultaneously. Then, perform operations such as feature extraction, feature fusion, coil shape and coordinate output in sequence, and regress the shape and position of the coil to be operated. The accuracy of coil shape and position estimation is premised on meeting the accuracy requirements of crochet loop formation.

[0084] It should be noted that the position of the coil to be operated obtained here is the position coordinate in the camera coordinate system. In order to enable the motion mechanism to actually move for loop formation, the position needs to be transformed to the motion mechanism coordinate system. Use the hand-eye calibration method to calibrate the matrix for transforming the camera coordinate system to the motion mechanism coordinate system, and multiply the position of the coil to be operated by the matrix to transform it to the motion mechanism coordinate system.

[0085] It should be noted that there are generally multiple coils in the field of view after obtaining the three-dimensional point cloud by taking a single photo. The above description is about the shape and position estimation of one coil. In actual operation, the present invention also adopts another method, calculating the shapes and positions of multiple coils each time and returning the position points in sequence. For this, the step "generating the three-dimensional point cloud of the knitted fabric surface based on three-dimensional depth measurement" in step S20 can be carried out as follows:

[0086] Obtain the layer height coefficient of the current knitting layer; if the layer height coefficient meets the preset measurement conditions, generate the three-dimensional point cloud of the current knitted fabric surface based on three-dimensional depth measurement; if the layer height coefficient does not meet the preset measurement conditions, retrieve the three-dimensional point cloud of the knitted fabric surface generated based on three-dimensional depth measurement last time.

[0087] In the knitting operations of different knitting layers, the knitted fabric of the lower knitting layer has obvious deformation due to the lack of force constraints between layers and has greater flexibility. Each time a loop formation action is performed, it will affect the adjacent coils. Therefore, when operating at a lower layer, taking only one loop formation action per photo has a higher accuracy. While when operating at a higher layer, since the knitted fabric has stronger stability and each loop formation has little impact on the surrounding coils, multiple loop formation actions can be performed per photo to improve efficiency. For this, the present invention sets a corresponding coefficient representing the layer height for different knitting layers, that is, the layer height coefficient. When the layer height coefficient of the current knitting layer is less than the preset coefficient threshold, it indicates that the current knitting layer belongs to the lower layer, and fewer loop formation times are performed per photo, such as forming one loop per photo. While when the layer height coefficient of the current knitting layer is greater than or equal to the preset coefficient threshold, it indicates that the current knitting layer belongs to the upper layer, and more loop formation times can be performed per photo, such as forming multiple loops per photo.

[0088] In practical applications, different layer height coefficients can be measured as parameters according to different materials, and the corresponding layer height coefficients can be initialized based on the set different materials to improve the knitting efficiency as much as possible.

[0089] S40. According to the shape and position of the coil to be operated, as well as the coil connection type at the position of the next hook needle, generate the movement path of the next hook needle to form a loop, and control the knitted fabric or the target hook needle to move according to the movement path to complete the operation of the next hook needle to form a loop.

[0090] In the embodiment of the present invention, before planning the movement path of the hook needle to form a loop, the coil connection type corresponding to the position of the next hook needle is also obtained. The initial value of the hook needle position is set during the initialization of the current knitting layer. The current hook needle position is updated after each post-processing. The position of the next hook needle can be deduced from the current hook needle position, and the coil connection type of the next hook needle position is obtained from the cache.

[0091] When planning the movement path of the next hook needle to form a loop, first move the hook needle to the initial value of the hook needle position during the process of the next hook needle forming a loop. Then, according to the shape and position of the coil to be operated, the coil connection type, etc., and adding the position margin for eliminating the cumulative error of the three-dimensional depth measurement and the hook needle movement control at the same time, the position of the constraint point for the hook needle movement can be determined. Based on the position of the constraint point, a smooth movement path of the hook needle can be fitted. The present invention takes into account the cumulative error of the three-dimensional depth measurement and the hook needle movement control, and when planning the movement path of the hook needle, the influence of this cumulative error is eliminated to the greatest extent.

[0092] Further, according to the movement path of the next hook needle to form a loop, the hook needle or the knitted fabric can be controlled to rotate and translate through a robotic arm or other transmission mechanisms, so that the hook needle moves according to the planned path. At this time, the wire feeding device also follows to perform the corresponding wire feeding action to cooperate with the hook needle to complete the loop forming operation. It should be noted that the robotic arm or the transmission mechanism needs to meet the condition of controlling the end to move from one point to another according to any three-dimensional space trajectory to ensure that the hook needle moves according to the planned path.

[0093] In addition, after the operation of the next hook needle to form a loop is completed, post-processing can be further performed, and the current position of the hook needle can be updated, and then enter the next loop forming process. The post-processing here includes fixing the loop after forming, transferring the loop, moving the three-dimensional depth measurement sensor or feeding the wire and other necessary processes.

[0094] It should be noted that the initialization of the current knitting layer allows setting the parameters of the loop forming process and the post-processing according to the knitting layer. The current knitting layer initialization only needs to be performed once for one layer, and the loop forming processes need to be continuously executed one by one.

[0095] In some scenarios, it is also possible to verify the loop formation control process. There are various ways to verify the loop formation control process. During the process of controlling the rotation and translation of the crochet hook or the knitted fabric, verification can be carried out while controlling, or verification can be carried out after the crochet hook moves to the target position (this target position can be the end point of the loop formation control process), or other verification methods. The specific verification strategy is set during the initialization of the current layer.

[0096] In this regard, the method for guiding loop formation of a three-dimensional depth measurement crochet hook provided by the embodiments of the present invention further includes the following steps:

[0097] Obtain the movement progress of the knitted fabric or the target crochet hook, and verify the process of the next crochet hook forming a loop when the movement progress meets the preset verification conditions.

[0098] In the embodiments of the present invention, the movement progress of the knitted fabric or the target crochet hook can be used to represent the progress of the loop formation control process. In this regard, when the preset verification conditions are met, such as moving to half or moving to the target position, the process of the next crochet hook forming a loop can be verified.

[0099] Specifically, during the process of controlling the rotation and translation of the crochet hook or the knitted fabric, verification is carried out while controlling. When controlling the movement of the crochet hook or the fabric, sensor data is obtained at regular time intervals, and then a correctness judgment is made to see if there are any errors during the movement process or if it meets the expectations. If an error is found, stop immediately and perform loop formation failure processing to prevent irreparable errors in a timely manner. However, because this method requires continuous acquisition and calculation of sensor data, it consumes a relatively large amount of computing performance and affects efficiency. The time interval for obtaining data needs to be adjusted appropriately to balance efficiency and correctness.

[0100] In addition, during the process of controlling the rotation and translation of the crochet hook or the knitted fabric, verification is carried out after the crochet hook moves to the target position. After the control loop formation is completed, sensor data is obtained and then a judgment is made on whether the loop formation is successful or failed. This method is relatively simple, only judging the final result once, and the speed is relatively fast, but it may result in an irrecoverable error in the final loop formation due to an error occurring midway.

[0101] Specifically, the process of verifying the next crochet hook forming a loop can be verified by measuring with a force sensor. In this regard, the step "verifying the process of the next crochet hook forming a loop" can adopt the following steps:

[0102] Obtain the force feedback value of the next crochet hook forming a loop. The force feedback value represents the feedback force from the knitted fabric received by the target crochet hook or the wire feeding mechanism; verify the process of the next crochet hook forming a loop according to the force feedback value.

[0103] In the embodiments of the present invention, during the threading process, the crochet hook and the thread feeding mechanism will bear the feedback force from the knitting structure of the knitted fabric. If the threading fails or the loop formation fails, the feedback force will be different from the feedback force of normal loop formation. By comparing the curve characteristics of the force feedback value of the next crochet hook loop formation obtained with the calibrated force feedback curve characteristics, it can be determined whether the loop formation is successful or failed.

[0104] In addition, in some other scenarios, other direct or indirect verification methods such as three-dimensional point cloud recognition and two-dimensional image recognition can also be used. Specifically:

[0105] 1) Three-dimensional point cloud recognition. There are obvious differences in the point cloud characteristics between successful loop formation and failed loop formation. For example, after successful loop formation, an additional loop will appear. Traditional point cloud recognition methods can be used to detect whether this loop exists and is normal, such as three-dimensional arc detection and other methods. It is also possible to extract the nearby point cloud and train a point cloud classification network for whether the loop formation is successful, such as networks like PointNet, PointNet++, PointCert, etc.

[0106] 2) Two-dimensional image recognition. If it is only to judge whether the loop formation is successful rather than guiding the operation by position, accurate judgment can also be made using two-dimensional images. An image classification network can be trained, and pictures of successful and failed loop formations are collected for binary classification training to obtain a classification network model. The selectable network models include but are not limited to networks such as ViT, MobileViT, Swin-Transformer, MobileNetV3, ConvNeXt, EfficientNetV2, etc. In addition, traditional image processing methods can also be used to detect the edges in the picture and extract the corresponding loop edges to judge the integrity of the loop.

[0107] On this basis, after successful loop formation, post-processing can be performed and the current position of the crochet hook can be updated, and then enter the next loop formation process. After the loop formation fails, various methods for dealing with loop formation failure can also be carried out. The crochet hook can be retracted to the initial position after the loop formation fails and then perform a loop formation operation again, or an alarm can be issued to prompt manual intervention operation, or other processing methods, or no processing can be done, which is set during the initialization of the current layer.

[0108] Specifically, the step of "verifying the process of the next crochet hook loop formation" further includes the following steps:

[0109] If the verification result is that the loop formation fails, obtain the number of times of failed loop formation of the next crochet hook, and when the number of failures is less than the preset number threshold, return to execute step S20, "generating a three-dimensional point cloud on the surface of the knitted fabric based on three-dimensional depth measurement".

[0110] That is to say, in the embodiments of the present invention, if the verification result is that the loop formation fails, the number of loop formation failures is further obtained. If the number of failures is less than the preset number threshold, the return hook is retracted to the initial position and a loop formation operation is performed again until the number of failures reaches the number threshold, at which time an alarm is issued to prompt manual intervention. Until the loop formation is successful, post-processing is performed, the current position of the hook is updated, and then the next loop formation process is entered.

[0111] In the method for guiding the hook to form a loop by three-dimensional depth measurement provided by the embodiments of the present invention, the selection of using three-dimensional depth measurement to guide the hook to form a loop instead of completely using a mechanical transmission method is because the three-dimensional depth measurement to guide the hook to form a loop is not limited by the number of hooks and can knit objects of any shape, which can greatly simplify the product structure and reduce the product size. The present invention has the following advantages:

[0112] 1) Greatly simplifies the mechanical and transmission structures of the device, making the device miniaturized and more suitable for home application scenarios.

[0113] 2) The loop formation of the hook is not restricted by the mechanical structure and is flexible in operation. Therefore, three-dimensional solid objects of any shape can be knitted, which is suitable for home application scenarios.

[0114] 3) Considers the unevenness and irregularity of the surface loops of the knitted fabric and has better robustness.

[0115] 4) In the hook loop formation motion path planning, the elimination of the cumulative errors of three-dimensional depth measurement and hook motion control is considered, and the reliability is better.

[0116] 5) Allows the three-dimensional depth measurement to guide the hook to form a loop to fail and uses remedial measures to improve the operating reliability of the device.

[0117] Based on the method for guiding the hook to form a loop by three-dimensional depth measurement provided in the above embodiments, the embodiments of the present invention also correspondingly provide a device for executing the method for guiding the hook to form a loop by three-dimensional depth measurement. The structural schematic diagram of the device is as Figure 7 shown.

[0118] An instruction acquisition module 101, configured to acquire a knitting instruction, where the knitting instruction includes the loop formation path of the knitted fabric in the current knitting layer, and the path points in the loop formation path represent the hook positions in the current knitting layer and the loop connection types under the hook positions.

[0119] A point cloud generation module 102, configured to generate a three-dimensional point cloud on the surface of the knitted fabric based on three-dimensional depth measurement if the current knitting layer is not the first knitting layer.

[0120] The crochet loop forming module 103 is used to estimate the shape and position of the loop to be operated based on the position of the previous loop, the position of the next crochet hook, and the three-dimensional point cloud of the knitted fabric surface. The loop to be operated is the loop of the previous knitting layer that needs to be looped through during the current crochet loop forming process. According to the shape and position of the loop to be operated and the loop connection type of the next crochet hook position, a motion path for the next crochet loop forming is generated, and the knitted fabric or the target crochet hook is controlled to move along the motion path to complete the next crochet loop forming operation.

[0121] Optionally, the point cloud generation module 102 for generating the three-dimensional point cloud of the knitted fabric surface based on three-dimensional depth measurement is specifically used for:

[0122] Generating the original three-dimensional point cloud of the knitted fabric surface based on a three-dimensional depth sensor; performing processing operations on the original three-dimensional point cloud, and the processing operations include one or more of point cloud fusion, point cloud hole filling, and filtering.

[0123] Optionally, the point cloud generation module 102 for generating the three-dimensional point cloud of the knitted fabric surface based on three-dimensional depth measurement is specifically used for:

[0124] Obtaining the layer height coefficient where the current knitting layer is located; if the layer height coefficient meets the preset measurement conditions, generating the three-dimensional point cloud of the current knitted fabric surface based on three-dimensional depth measurement; if the layer height coefficient does not meet the preset measurement conditions, retrieving the three-dimensional point cloud of the knitted fabric surface generated based on the previous three-dimensional depth measurement.

[0125] Optionally, the crochet loop forming module 103 for estimating the shape and position of the loop to be operated based on the position of the previous loop, the position of the next crochet hook, and the three-dimensional point cloud of the knitted fabric surface is specifically used for:

[0126] Determining the loop to be operated according to the position of the previous loop and the position of the next crochet hook; extracting the three-dimensional point cloud of each loop from the three-dimensional point cloud of the knitted fabric surface; obtaining the shape of the wire material of the knitted fabric, and fitting the three-dimensional point cloud of the loop to be operated according to the wire material shape to obtain the shape and position of the loop to be operated.

[0127] Optionally, the crochet loop forming module 103 for extracting the three-dimensional point cloud of each loop from the three-dimensional point cloud of the knitted fabric surface is specifically used for:

[0128] Determining the pixel points on the knitted fabric surface where the gradient depths in all directions are not in a decreasing state from the three-dimensional point cloud of the knitted fabric surface, and searching for the three-dimensional point cloud of each pixel point belonging to the same loop on the knitted fabric surface with the three-dimensional point cloud of the determined pixel points as the initial value; or obtaining the two-dimensional image of the knitted fabric surface, and inputting the three-dimensional point cloud and the two-dimensional image of the knitted fabric surface into a pre-trained segmentation model to determine the three-dimensional point cloud of each loop on the knitted fabric surface.

[0129] Optionally, the crochet looping module 103 is further configured to:

[0130] Obtain the movement progress of the knitted fabric or the target crochet hook, and verify the process of the next crochet looping when the movement progress meets the preset verification condition.

[0131] Optionally, the crochet looping module 103 for verifying the process of the next crochet looping is specifically configured to:

[0132] Obtain the force feedback value of the next crochet looping, where the force feedback value represents the feedback force from the knitted fabric received by the target crochet hook or the wire feeding mechanism; verify the process of the next crochet looping according to the force feedback value.

[0133] Optionally, the crochet looping module 103 for verifying the process of the next crochet looping is further configured to:

[0134] If the verification result is that the looping fails, obtain the number of failure times of the next crochet looping, and when the number of failure times is less than the preset number threshold, return to trigger the point cloud generation module 102 to execute the step of generating the three-dimensional point cloud on the surface of the knitted fabric based on three-dimensional depth measurement.

[0135] It should be noted that for the refined functions of each module in the embodiments of the present invention, reference can be made to the corresponding disclosed parts in the method embodiments of the above three-dimensional depth measurement-guided crochet looping, which will not be elaborated here.

[0136] Based on the method for three-dimensional depth measurement-guided crochet looping provided in the above embodiments, an embodiment of the present invention further provides a device, which includes: at least one memory and at least one processor; the memory stores an application program, and the processor calls the application program stored in the memory, and the application program is used to implement the above method for three-dimensional depth measurement-guided crochet looping.

[0137] The above has introduced in detail a method, device and equipment for three-dimensional depth measurement-guided crochet looping provided by the present invention. Specific examples are used in this article to elaborate on the principle and implementation manner of the present invention. The description of the above embodiments is only used to help understand the method and its core idea of the present invention; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation to the present invention.

[0138] It should be noted that the embodiments in this specification are all described in a progressive manner. Each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the description of the method part.

[0139] It should also be noted that in this text, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that the elements inherent in a process, method, article or device comprising a series of elements, or those further comprising elements inherent in these process, methods, articles or devices. Without further limitation, an element defined by the statement "comprising an..." does not exclude the presence of additional identical elements in the process, method, article or device comprising the said element.

[0140] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for guiding the loop formation of a crochet hook by three-dimensional depth measurement, characterized in that, The method for guiding loop formation of a crochet hook by three-dimensional depth measurement includes: Obtaining a knitting instruction, which includes the loop formation path of the knitted fabric in the current knitting layer. The path points in the loop formation path represent the positions of the crochet hooks in the current knitting layer and the type of coil connection under the positions of the crochet hooks; If the current knitting layer is not the first knitting layer, generating a three-dimensional point cloud of the surface of the knitted fabric based on three-dimensional depth measurement; Estimating the shape and position of the coil to be operated based on the position of the previous coil, the position of the next crochet hook, and the three-dimensional point cloud of the surface of the knitted fabric. The coil to be operated is the coil of the previous knitting layer that needs to be looped during the loop formation process of the current crochet hook; Generating a movement path for the next crochet hook to form a loop according to the shape and position of the coil to be operated and the type of coil connection at the position of the next crochet hook, and controlling the knitted fabric or the target crochet hook to move according to the movement path to complete the loop formation operation of the next crochet hook.

2. The method for guiding crochet loop formation by three-dimensional depth measurement according to claim 1, wherein The generating of the three-dimensional point cloud of the surface of the knitted fabric based on three-dimensional depth measurement includes: Generating an original three-dimensional point cloud of the surface of the knitted fabric based on a three-dimensional depth sensor; Performing a processing operation on the original three-dimensional point cloud, and the processing operation includes one or more of point cloud fusion, point cloud hole filling, and filtering.

3. The method for guiding loop formation of a three-dimensional depth measurement crochet hook according to claim 1, wherein The generating of the three-dimensional point cloud of the surface of the knitted fabric based on three-dimensional depth measurement includes: Obtaining the floor height coefficient where the current knitting layer is located; If the floor height coefficient meets the preset measurement conditions, generating the three-dimensional point cloud of the surface of the knitted fabric for this time based on three-dimensional depth measurement; If the floor height coefficient does not meet the preset measurement conditions, retrieving the three-dimensional point cloud of the surface of the knitted fabric generated based on three-dimensional depth measurement last time.

4. The method for guiding the crochet loop formation by three-dimensional depth measurement according to claim 1, characterized in that, The estimating of the shape and position of the coil to be operated based on the position of the previous coil, the position of the next crochet hook, and the three-dimensional point cloud of the surface of the knitted fabric includes: Determining the coil to be operated according to the position of the previous coil and the position of the next crochet hook; Extracting the three-dimensional point clouds of each coil from the three-dimensional point cloud of the surface of the knitted fabric; Obtaining the shape of the wire material of the knitted fabric, and fitting the three-dimensional point cloud of the coil to be operated according to the shape of the wire material to obtain the shape and position of the coil to be operated.

5. The method for guiding the crochet loop formation in three-dimensional depth measurement according to claim 4, characterized in that, The extracting of the three-dimensional point clouds of each coil from the three-dimensional point cloud of the surface of the knitted fabric includes: Determining the pixel points on the surface of the knitted fabric where the gradient depths in all directions are not in a decreasing state from the three-dimensional point cloud of the surface of the knitted fabric, and searching for the three-dimensional point clouds of the pixel points belonging to the same coil on the surface of the knitted fabric with the three-dimensional point clouds of the determined pixel points as the initial values; or Obtaining a two-dimensional image of the surface of the knitted fabric, and inputting the three-dimensional point cloud of the surface of the knitted fabric and the two-dimensional image into a pre-trained segmentation model to determine the three-dimensional point clouds of each coil on the surface of the knitted fabric.

6. The method for guiding loop formation of a three-dimensional depth measurement crochet hook according to claim 1, wherein The method for guiding loop formation of a crochet hook by three-dimensional depth measurement further includes: Obtaining the moving progress of the knitted fabric or the target crochet hook, and verifying the process of the next crochet hook to form a loop when the moving progress meets the preset verification conditions.

7. The method for guiding the crochet loop formation in three-dimensional depth measurement according to claim 6, characterized in that, The verifying of the process of the next crochet hook to form a loop includes: Obtain the force feedback value for the next crochet loop formation, where the force feedback value represents the feedback force from the knitted fabric received by the target crochet hook or the yarn feeding mechanism; Verify the process of the next crochet loop formation according to the force feedback value.

8. The method for guiding crochet loop formation in three-dimensional depth measurement according to claim 7, characterized in that, The verification of the process of the next crochet loop formation further includes: If the verification result is that the loop formation fails, obtain the number of failures of the next crochet loop formation, and when the number of failures is less than the preset number threshold, return to execute the step of generating the three-dimensional point cloud of the knitted fabric surface based on the three-dimensional depth measurement.

9. A device for guiding the loop formation of a three-dimensional depth measurement crochet hook, characterized in that, The device for guiding crochet loop formation by three-dimensional depth measurement includes: An instruction acquisition module, configured to acquire a knitting instruction, where the knitting instruction includes the loop formation path of the knitted fabric in the current knitting layer, and the path points in the loop formation path represent the crochet hook positions in the current knitting layer and the type of coil connection at the crochet hook positions; A point cloud generation module, configured to generate the three-dimensional point cloud of the knitted fabric surface based on the three-dimensional depth measurement if the current knitting layer is not the first knitting layer; A crochet loop formation module, configured to estimate the shape and position of the coil to be operated based on the position of the previous coil, the position of the next crochet hook, and the three-dimensional point cloud of the knitted fabric surface, where the coil to be operated is the coil of the previous knitting layer that needs to be strung during the current crochet loop formation process; generate the movement path of the next crochet loop formation according to the shape and position of the coil to be operated and the type of coil connection at the position of the next crochet hook, and control the knitted fabric or the target crochet hook to move according to the movement path to complete the operation of the next crochet loop formation.

10. A device, characterized in that, The device includes: at least one memory and at least one processor; the memory stores an application program, and the processor calls the application program stored in the memory, and the application program is used to implement the method for guiding crochet loop formation by three-dimensional depth measurement according to any one of claims 1-8.

Images