Cone yarn color identification and sorting equipment based on machine vision

Abstract

The invention relates to the technical field of sorting equipment, and discloses cone yarn color recognition and sorting equipment based on machine vision, which comprises a mechanical arm, clamping jaws and an air cylinder are mounted at the movable end of the mechanical arm, and the clamping jaws are circumferentially distributed on the outer side of the air cylinder; the middle of one face of the connecting plate is connected with the output end of the air cylinder, the edge of the connecting plate is connected with the clamping jaw, and the connecting plate is used for driving the clamping jaw to move along with operation of the air cylinder; the camera is installed in the middle of the other face of the connecting plate and used for collecting the cone yarn color; and the support is installed on the outer wall of the connecting plate, and a light supplementing lamp used for providing illumination for the camera is installed in the support. According to the method, the target is separated by adopting the Mask-R-CNN model, and visual interference on the surface of the cone yarn is weakened by utilizing a channel attention mechanism. And in combination with a CIE-Lab color space insensitive to illumination, various colors can still be accurately and stably identified under complex illumination and surface conditions, and particularly, similar color numbers with extremely small chromatic aberration can be effectively distinguished.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of sorting equipment, in particular to a bobbin color identification and sorting equipment based on machine vision. BACKGROUND

[0002] Bobbin refers to a yarn packaging form that is convenient for storage, transportation and subsequent processing (such as weaving, knitting) in the textile production process. In modern textile industry, in order to produce color-uniform and design-required cloth, color sorting of bobbin is a crucial quality control link. Especially in the scene of producing multi-color fabric or needing to splice yarns of different colors, it is necessary to ensure that each batch of bobbin used has accurate and consistent color number, so as to avoid quality defects such as color difference and color grade in the final product.

[0003] The traditional manual sorting method relies on visual judgment, and has problems of low efficiency, high labor intensity, strong subjectivity of color difference judgment, and accurate identification of similar color numbers. SUMMARY

[0004] In view of the deficiencies of the prior art, the present application provides a bobbin color identification and sorting equipment based on machine vision, which solves the problems of the prior art that the traditional manual sorting of bobbin relies on visual judgment, has low efficiency, high labor intensity, strong subjectivity of color difference judgment, and cannot accurately identify similar color numbers.

[0005] To achieve the above purpose, the present application is realized by the following technical scheme: a bobbin color identification and sorting equipment based on machine vision, comprising:

[0006] A machine arm has an active end integrated with an end effector for bobbin grabbing and identification. The end effector includes a cylinder and a plurality of clamps distributed circumferentially outside the cylinder;

[0007] The middle part of one side of the connecting plate is connected with the output end of the cylinder, and the edge part is connected with each clamp;

[0008] A camera is installed in the middle part of the other side of the connecting plate, and its lens faces outward for collecting images of bobbins within the grabbing range of the clamps; a plurality of supports are installed on the outer wall of the connecting plate, staggered with the clamps, and a light supplement lamp is installed inside each support for providing illumination to the camera's collection area;

[0009] As a preferred layout, the support is provided in multiple and is symmetrically distributed in a ring outside the camera, the multiple built-in light compensation lamps collectively form a ring light source, and the ring light source is coaxially arranged around the lens of the camera. This structure can provide uniform and shadow-free lighting conditions for the cylindrical surface of the cheese, and can minimize the interference caused by uneven lighting and high light reflection on color recognition.

[0010] The controller is also included, which is in communication or electrical connection with the machine arm, cylinder, camera and light compensation lamp; the controller has an operation logic inside to achieve the purpose of the application, which can be specifically divided into the following functional modules:

[0011] An image data acquisition module is used to receive the digital image containing one or more cheeses collected by the camera in real time;

[0012] A signal processing and color recognition module is used to algorithmically process the received cheese image to determine the final color number of each cheese;

[0013] A sorting control module is used to generate and send motion control instructions to the machine arm and cylinder according to the determined final color number to drive it to perform corresponding sorting actions.

[0014] The signal processing and color recognition module is the key to achieving high-precision color recognition of the application, and the algorithm flow inside includes the following units:

[0015] An instance segmentation unit adopts a Mask-R-CNN (Mask Region-based Convolutional Neural Network) model to output a pixel-level mask region (Mask) of each cheese target in the image after receiving the original image, thereby effectively separating the cheeses.

[0016] A feature enhancement unit is integrated in the Mask-R-CNN model, specifically after the backbone network of the model. This unit adopts a spatial-channel attention mechanism (Spatial-and-Channel-Attention-Mechanism). After the backbone network extracts the initial feature map of the image, the feature map is sent to this unit. This unit uses spatial attention mechanism to make the model focus more computing resources on the spatial position features corresponding to the mask region; at the same time, through the channel attention mechanism, it enhances the weight of the feature channel most sensitive to color information, and suppresses noise irrelevant to the color recognition task, such as feather and winding texture feature channels on the surface of the cheese.

[0017] A color feature extraction unit receives the image data containing only the mask region of the cone yarn after being processed by the feature enhancement unit. In order to eliminate the serious influence of the workshop ambient light changes on the color recognition, the unit first converts the image data from the RGB color space sensitive to light to the CIE-Lab color space which is not sensitive to light changes and more in line with the visual perception of the human eye.

[0018] A color matching unit performs the final recognition and matching of the aforementioned color features. The specific steps are as follows:

[0019] First, in order to obtain a single quantitative value representing the color of the entire cone yarn, the average value of the color features of all pixel points in the mask region in the CIE-Lab color space is calculated. If the color value of any pixel point in the mask region is: Let the total number of pixel points in the region be N, then the quantitative color value of the cone yarn is The calculation is as follows:

[0020]

[0021] In the formula, C i represents the color value vector of the i-th pixel point in the mask region; represents the brightness component of the i-th pixel point in the CIE-Lab color space; represents the green-red color component of the i-th pixel point in the CIE-Lab color space; represents the blue-yellow color component of the i-th pixel point in the CIE-Lab color space; i represents the index of the pixel point, which takes integer values from 1 to N; N represents the total number of pixel points contained in the specific cone yarn mask region output by the instance segmentation unit; represents the quantitative color value vector finally calculated for the entire cone yarn; represents the average brightness component of the entire cone yarn; represents the average green-red color component of the entire cone yarn; represents the average blue-yellow color component of the entire cone yarn.

[0022] Subsequently, the obtained quantitative color value is compared with the color difference distance of each standard color number in a preset color number library. Preferably, the CIEDE color difference formula is used for calculation, which can simulate the perception of the human eye to color difference. The calculated 2000 color difference distance is compared with a preset threshold value, and the standard color number with the smallest color difference distance and less than the threshold value is selected as the final color number of the cone yarn.

[0023] In particular, the preset color number library is a configurable database, and an operator can conveniently add, delete or modify the standard color number and the corresponding CIE-Lab value in the library through a man-machine interaction interface according to specific needs of different production batches or customers.

[0024] Subsequently, after determining the final color number of the cone yarn, the sorting control module generates an instruction to drive the equipment to perform physical sorting. The machine arm first moves above the cone yarn to be grabbed, and then the cylinder extends its output end to push the connecting plate to move downward. The connecting plate extrudes the inner side of the plurality of clamps through lever action, so that the plurality of clamps are synchronously retracted to the center, thereby stably clamping the cone yarn. Then, the machine arm transports the cone yarn to a target area determined by the color number of the cone yarn. After reaching the target position, the cylinder extends its output end, the connecting plate is retracted, and the clamps are pulled through the connecting relationship to be synchronously opened, thereby completing the release of the cone yarn.

[0025] The application provides a cone yarn color recognition and sorting equipment based on machine vision. The equipment has the following advantages:

[0026] 1. The Mask-R-CNN model is used to segment the cone yarn, which can separate the target from the complex background. Combined with the channel attention mechanism, the visual noise interference caused by the physical characteristics such as feathers and winding texture on the surface of the cone yarn can be effectively suppressed. The CIE-Lab color space is used to realize the color recognition ability with high accuracy and stability under different illumination conditions and surface states of the cone yarn, especially when dealing with similar color numbers with extremely small color difference.

[0027] 2. The camera is used for visual recognition, and the machine arm and the clamps are used for cooperation. The whole process automation from image acquisition, target recognition, color number matching to final physical grabbing and placing can be realized, which replaces the traditional manual sorting operation. The errors and inconsistencies caused by subjective judgment are fundamentally eliminated, and the sorting speed is improved by orders of magnitude, thereby significantly improving the overall efficiency of the textile production line.

[0028] 3. The clamps, the camera and the fill light are integrated on the execution end of the mechanical arm. The fill light can provide imaging conditions for the camera, so that the device is easy to deploy and maintain in a complex industrial environment. By setting a preset color number library that can be dynamically configured by the user, the equipment can flexibly adapt to different color standards of production batches, and can quickly switch sorting tasks without retraining the algorithm model, thereby showing strong industrial practicability and flexibility. BRIEF DESCRIPTION OF DRAWINGS

[0029] Figure 1 It is a schematic diagram of the three-dimensional structure of the application;

[0030] Figure 2This is a schematic diagram of an embodiment of the present invention;

[0031] Figure 3 This is a schematic diagram of the gripper portion of the present invention;

[0032] Figure 4 This is a schematic diagram of the support structure distribution according to the present invention;

[0033] Figure 5 This is a schematic diagram of the support structure of the present invention;

[0034] Figure 6 This is a schematic diagram of the system architecture of the present invention.

[0035] Among them, 1. robotic arm; 101. cylinder; 2. gripper; 3. connecting plate; 4. camera; 5. bracket. Detailed Implementation

[0036] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0037] To better understand the present invention, the above content will be described in detail below with reference to specific embodiments.

[0038] Example 1, please refer to the appendix. Figure 1 -Appendix Figure 5 This invention provides a machine vision-based yarn color recognition and sorting equipment, comprising: a robotic arm 1, with grippers 2 and a cylinder 101 mounted on its movable end, the grippers 2 being circumferentially distributed outside the cylinder 101; a connecting plate 3, one side of which is connected to the output end of the cylinder 101, and the edge of which is connected to the grippers 2, the connecting plate 3 being used to drive the grippers 2 to move along with the cylinder 101; a camera 4, which is mounted on the other side of the connecting plate 3, for collecting yarn color; and the grippers 2 being located outside the camera 4; and a bracket 5, which is mounted on the outer wall of the connecting plate 3 and is staggered with the grippers 2, the bracket 5 having a supplementary light installed inside, for providing illumination to the collection area of ​​the camera 4.

[0039] In this embodiment, the robotic arm can be used to operate the gripper 2 and the cylinder 101. The cylinder 101 can drive the gripper 2 to operate, thereby causing the gripper 2 to open and close, thus achieving the clamping and release of the yarn package. Subsequently, in conjunction with the movement of the robotic arm, and with the camera 4 collecting and analyzing the color of the yarn package, the robotic arm can be controlled to move yarn packages of different colors to different positions, thereby achieving the sorting action of the yarn package.

[0040] When the machine arm 1 and the cylinder 101 are in operation, the machine arm 1 is used to drive the cylinder 101 and the clamping jaw 2 to move, the cylinder 101 retracts the output end to pull the connecting plate 3 back, at this time the connecting plate 3 pulls the clamping jaw 2 to make the clamping jaw 2 close and clamp the bobbin; the cylinder 101 extends the output end to push the connecting plate 3 out, at this time the connecting plate 3 presses the clamping jaw 2 to make the clamping jaw 2 open and release the bobbin.

[0041] In the embodiment, the connecting plate 3 is pushed by the cylinder 101 to move, and the connecting plate 3 presses or pulls the clamping jaw 2 to realize the opening and closing of the clamping jaw 2.

[0042] The supports 5 are arranged in multiple numbers and are arranged in a ring shape outside the camera 4 to provide a ring-shaped light source, and the supports 5 are coaxially arranged around the lens of the camera 4.

[0043] In the embodiment, multiple supports 5 are arranged, and the supports 5 are used to realize the coaxial arrangement of multiple light supplementing lamps around the lens of the camera 4, so as to provide a ring-shaped light source for the camera 4, and provide good lighting conditions for the camera 4 to improve the accuracy of the image acquisition of the bobbin.

[0044] Working principle: when in use, the cylinder 101 and the clamping jaw 2 are driven by the mechanical arm to move, when the bobbin needs to be sorted, first, the mechanical arm moves the cylinder 101 and the clamping jaw 2 above the bobbin, at this time the camera and the light supplementing lamp are in operation, the light supplementing lamp is used to illuminate the bobbin, which is conducive to the image acquisition of the bobbin by the camera 4, when the acquisition is completed, the controller can analyze the image of the acquired bobbin, and drive the mechanical arm and the cylinder 101 to operate according to the analysis result.

[0045] Specifically, first, the mechanical arm drives the clamping jaw 2 to approach the bobbin and wrap the bobbin with the clamping jaw 2, then drives the cylinder 101 to retract the output end to drive the connecting plate 3 to approach the mechanical arm, at this time the connecting plate 3 pulls the clamping jaw 2 to drive the clamping jaw 2 to move, then the clamping jaw 2 closes to clamp the bobbin, then moves the mechanical arm to pick up the bobbin and move to a specified position. Finally, drive the cylinder 101 to extend the output end to drive the connecting plate 3 away from the mechanical arm and press the clamping jaw 2, at this time the clamping jaw 2 opens to release the bobbin. In this way, one sorting action is completed, and after the bobbin is placed, the next sorting can be performed.

[0046] Embodiment 2, based on embodiment 1, please refer to the attached Figure 6 In the embodiment,

[0047] The controller is installed in the inside of the machine arm 1, and the controller comprises:

[0048] An image data acquisition module is used to receive the image of the bobbin collected by the camera 4.

[0049] In this embodiment, to realize the automatic operation and intelligent identification of the above-mentioned equipment, specific information of a controller is further provided. The controller is used for receiving external sensor information, executing an algorithm, and issuing a control instruction to the mechanical arm. The controller mainly comprises an image data acquisition module, a signal processing and color identification module, and a sorting control module.

[0050] Specifically, the image data acquisition module is connected with the camera 4 and receives the original image data of the cheese containing the cheese collected by the camera 4.

[0051] Specifically, the module is responsible for initializing various parameters of the camera 4 and establishing a communication link with the camera 4. In a possible implementation manner, the image acquisition process can be triggered by an event. For example, when the machine arm 1 moves to above the preset shooting position of the area to be sorted. The camera 4 collects the cheese image. Alternatively, the acquisition process can also be continuous. The camera 4 continuously transmits a video stream to the image data acquisition module at a preset frame rate (for example, 30 frames per second).

[0052] Generally, the module receives a digital image, for example, an RGB format image with a resolution of 1920x1080 pixels and 24-bit true color coding. After receiving the original image data, the module does not simply forward the data. First, the integrity of the data frame is checked to ensure that a complete image without damaged or missing data is received.

[0053] In addition, to ensure the accuracy and repeatability of the measurement results, the module will call the camera 4 calibration program after the system is started for the first time or the camera 4 is replaced. The program is used to correct the geometric distortion of the camera 4 itself, such as barrel distortion or pillow distortion, to ensure that the straight line in the final image is also a straight line in the real world, providing an accurate image basis for subsequent instance segmentation and size measurement.

[0054] Finally, the image data acquisition module stores the complete image data that has been checked and possibly pre-processed, with a time stamp, in a specific high-speed memory buffer or data queue for the signal processing and color identification module to call at any time. This modular design ensures clear and efficient data flow, providing a stable and reliable data source for a series of complex image processing and recognition algorithms.

[0055] The signal processing and color identification module is used to receive and process the cheese image to determine the final color number of the cheese:

[0056] In this embodiment, when the image data acquisition module transmits stable and clear cheese image data into the controller, the signal processing and color identification module begins to take over the processing process. The purpose is to accurately and stably extract the final color number information of each cheese from the complex original image.

[0057] Specifically, the functions of the signal processing and color recognition module are refined into multiple sequentially executed units, including an instance segmentation unit, a feature enhancement unit, a color feature extraction unit, and a color matching unit. These units work together to form a complete processing pipeline from target positioning, feature purification, feature conversion to final decision-making.

[0058] Among them, the instance segmentation unit

[0059] First, the image data is sent to the instance segmentation unit. It is used to accurately locate and separate each target object of interest in a complex scene.

[0060] Specifically, this unit uses a Mask-R-CNN (Masked Region-based Convolutional Neural Network) model. This model receives the entire cheese image transmitted by the image data acquisition module and performs in-depth analysis on it. Its output is not a simple rectangular box, but can outline the outline of each cheese in the image, thereby completely separating it from the background, other adjacent cheeses, and any debris at the data level.

[0061] In one possible implementation, the Mask-R-CNN model specifically includes:

[0062] The backbone network uses a mature convolutional neural network structure such as ResNet (Residual Network) or EfficientNet. It is responsible for performing preliminary feature extraction on the input cheese image to generate an initial feature map containing low-level to high-level information (such as edges, textures, components, etc.) in the image.

[0063] The feature enhancement unit is set after the backbone network. It is used to receive the initial feature map output by the backbone network and process it using a channel attention mechanism. This mechanism enables the model to adaptively learn which regions (spatial dimension) and features (channel dimension) in the image are most important for identifying cheese color. For example, it enhances attention to the main body area of the cheese, while suppressing the response to noise features such as the fluff on the surface of the cheese due to the fluffiness of the yarn, and the subtle texture due to uneven winding. In this way, the core features related to the target color of the cheese are greatly enhanced, providing higher quality data for subsequent processing.

[0064] The segmentation head receives the enhanced feature map processed by the feature enhancement unit and, based on these feature maps, finally generates independent and accurate binary mask regions for each cheese instance in the image. The data of the mask region will then be passed to the subsequent color feature extraction unit.

[0065] Color feature extraction unit

[0066] After obtaining the accurate mask region of each cone yarn, the color feature extraction unit extracts color features that are stable and reliable and can resist the interference of changes in the workshop ambient light.

[0067] Generally, the image data directly output by the industrial camera 4 is based on the RGB (red, green, and blue) color space. However, the RGB values are extremely sensitive to changes in the intensity and color temperature of light, and slight fluctuations in light can cause huge changes in the RGB values, thereby causing color misjudgment.

[0068] Therefore, in this embodiment, the color feature extraction unit first performs a key color space conversion on the image data in the mask region, i.e., converts it from the RGB color space to the CIE-Lab color space. The CIE-Lab color space is a color model that is more consistent with human visual perception, and after separating the brightness from the color information, even if the ambient light changes (mainly affecting the component), the component remains relatively stable, thereby enabling the extraction of pure color features that are not sensitive to changes in light.

[0069] Color matching unit

[0070] When the stable and reliable color features are extracted, the final decision-making stage is entered, i.e., the color matching unit is executed. The unit matches the extracted color features with the preset standard color number through a classifier to determine the final color number of the cone yarn.

[0071] Specifically, the steps of matching the color features with the preset color number by the color matching unit include:

[0072] First, in order to obtain a single quantitative value that can represent the overall color of the cone yarn from a mask region containing thousands of pixels, the unit calculates the arithmetic mean of the color features of all the pixels in the mask region.

[0073] In the CIE-Lab color space, the color value of any pixel in the mask region can be represented as:

[0074]

[0075] At this time, let the total number of pixels in the region be N, then the quantitative color value of the cone yarn is The calculation is as follows:

[0076]

[0077] In the formula, C i represents the color value vector of the i-th pixel in the mask region; represents the brightness component of the i-th pixel in the CIE-Lab color space; G(i) represents the green-red color component of the i-th pixel in CIE-Lab color space; B(i) represents the blue-yellow color component of the i-th pixel in CIE-Lab color space; i represents the index of the pixel, which takes integer values from 0 to N; N represents the total number of pixels contained in the specific cheese mask region output by the instance segmentation unit; Q(i) represents the quantized color value vector of the i-th pixel in CIE-Lab color space; L represents the average brightness component of the entire cheese; G represents the average green-red color component of the entire cheese; B represents the average blue-yellow color component of the entire cheese.

[0078] In some embodiments, a preset color difference threshold can also be set. Only when the minimum color difference distance is less than the threshold, the corresponding standard color number is confirmed as the final color number of the cheese. If the minimum color difference is also greater than the threshold, the cheese can be determined as unknown color or unqualified product.

[0079] wherein the preset color number library is a configurable database, which allows users or on-site engineers to conveniently add new standard color numbers, delete obsolete color numbers or modify the CIE-Lab standard values of existing color numbers according to the needs of the current production batch through a human-computer interaction interface, without the need to retrain the core algorithm model, greatly improving the adaptability and practicality of the equipment.

[0080] The sorting control module is used to generate control instructions to drive the robot arm 1 and the air cylinder 101 to operate according to the final color number of the cheese, so as to sort the cheese to the corresponding target position.

[0081] In this embodiment, the sorting control module receives the color number of the cheese and its position information in the image coordinate system output by the signal processing and color recognition module, and generates control instructions to drive the robot arm 1 and the air cylinder 101 to complete the complete sorting operation from grabbing to placing according to these inputs.

[0082] Specifically, the module maintains a color number-target position mapping table inside. This table can be a configurable data structure, such as a lookup table (Look-up-Table) or an associative array, which associates each preset cheese color number with one or more specific three-dimensional coordinate points. These coordinate points correspond to different collection bins, conveyors or packaging areas in the production site.

[0083] In one possible implementation, when the sorting control module receives a determined cheese color number (such as "positive red-01") and its position information, its workflow is as follows:

[0084] Firstly, the module queries the mapping table to obtain the three-dimensional coordinates of the target placement position corresponding to "positive red-01".

[0085] Subsequently, the module initiates a path planning algorithm. The algorithm calculates a trajectory for the robot arm 1 to move from the current position to a position directly above the to-be-grabbed cheese. The generation of the trajectory needs to take into account the kinematic model of the robot arm 1, and the angles at which the joints need to be rotated are calculated by inverse kinematics.

[0086] In some embodiments, the path planning algorithm can also load information about static obstacles in the workspace. When generating the trajectory, the algorithm actively avoids these obstacles to generate a collision-free safe path.

[0087] When the end effector of the robot arm 1 reaches the predetermined grabbing position, the sorting control module generates and sends a grabbing action instruction. The instruction first drives the retracting end of the air cylinder 101 to extend, and then drives the connecting plate 3 to press the clamping jaw 2 to retract, so as to complete the clamping action of the cheese.

[0088] Alternatively, the control signal sent to the air cylinder 101 can include parameterized control of the force or speed. In this way, flexible grabbing of the cheese can be achieved, avoiding damage to the yarn or deformation of the cheese due to excessive clamping force.

[0089] After confirming that the cheese is stably clamped, the module again calls the path planning algorithm to calculate a trajectory for the robot arm 1 to move from the current grabbing position to the target placement position obtained by querying the table.

[0090] Generally, the robot arm 1 transports the cheese to the position above the target position according to the trajectory. At this time, the module generates and sends a release action instruction. The instruction drives the output end of the air cylinder 101 to extend, the connecting plate 3 is thereby pushed out and presses the clamping jaw 2 to open, so as to release the cheese into the specified collection bin.

[0091] During the entire execution process, the sorting control module continuously receives and processes feedback signals from the joint encoders of the robot arm 1, the position sensors of the air cylinder 101, or the pressure sensors of the clamping jaw 2, etc. This closed-loop control ensures that each instruction is accurately executed. If an abnormality is detected at a certain link (e.g. grabbing), such as insufficient pressure feedback from the sensor, indicating that the grabbing has failed, the module will immediately suspend the current task process and execute a pre-set exception handling program, such as re-attempting to grab or sending an alarm to the central monitoring system.

[0092] Although embodiments of the present application have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, replacements and variations can be made to these embodiments without departing from the principles and spirit of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. A machine vision-based yarn color recognition and sorting equipment, characterized in that, include: The robotic arm (1) has a gripper (2) and a cylinder (101) installed at its movable end. The gripper (2) is circumferentially distributed on the outside of the cylinder (101). The connecting plate (3) has its center connected to the output end of the cylinder (101) and its edge connected to the gripper (2). The connecting plate (3) is used to drive the gripper (2) to move along with the cylinder (101). A camera (4) is mounted on the other side of the connecting plate (3) in the middle for collecting the color of the yarn package; and the gripper (2) is located outside the camera (4); A bracket (5) is installed on the outer wall of the connecting plate (3) and is staggered with the gripper (2). A supplementary light is installed inside the bracket (5) to provide illumination to the acquisition area of ​​the camera (4).

2. The machine vision-based yarn color recognition and sorting equipment according to claim 1, characterized in that, The robotic arm (1) is also equipped with a controller, which includes: The image data acquisition module is used to receive images of yarn bobbins captured by the camera (4); A signal processing and color recognition module is used to receive and process the yarn package image to determine the final color code of the yarn package. The sorting control module is used to generate control commands based on the final color number of the yarn package to drive the robotic arm (1) and cylinder (101) to sort the yarn package to the corresponding target position.

3. The machine vision-based yarn color recognition and sorting equipment according to claim 2, characterized in that, The signal processing and color recognition module includes: An instance segmentation unit is used to process the image using a Mask-R-CNN model to generate a mask region for locating the yarn package and separating it from the background. The feature enhancement unit is used to enhance the feature attention of the mask region and suppress background noise through a channel attention mechanism; The color feature extraction unit is used to perform color space conversion on the image data within the mask area output by the feature enhancement unit in order to extract color features that are not sensitive to changes in illumination. The color matching unit is used to identify the extracted color features through a classifier and match them with preset color numbers to determine the final color number of the yarn package.

4. The machine vision-based yarn color recognition and sorting equipment according to claim 3, characterized in that, When the robotic arm (1) and cylinder (101) are running, the robotic arm (1) is used to drive the cylinder (101) and gripper (2) to move. After the cylinder (101) extends its output end to push out the connecting plate (3), the connecting plate (3) squeezes the gripper (2) to open the gripper (2) and release the yarn. After the cylinder (101) retracts its output end and retracts the connecting plate (3), the connecting plate (3) pulls the clamp (2) to close the clamp (2) and clamp the yarn.

5. The machine vision-based yarn color recognition and sorting equipment according to claim 1, characterized in that, The bracket (5) is configured as a plurality of brackets, which are arranged in a ring around the outside of the camera (4) to provide a ring light source, and the brackets (5) are coaxially arranged around the lens of the camera (4).

6. The machine vision-based yarn color recognition and sorting equipment according to claim 3, characterized in that, The color feature extraction unit is used to convert the image data within the mask area from the RGB color space to the CIE-Lab color space in order to extract color features that are not sensitive to changes in illumination.

7. The machine vision-based yarn color recognition and sorting equipment according to claim 3, characterized in that, The Mask-R-CNN model specifically includes: The backbone network is used to extract the initial feature map of the yarn package image; The feature enhancement unit, integrated after the backbone network, is used to receive the initial feature map and process the initial feature map through a spatial channel attention mechanism to enhance features related to the yarn target and suppress background noise. A segmentation head is used to generate the mask region for locating the yarn bobbin based on the enhanced feature map output by the feature enhancement unit.

8. The machine vision-based yarn color recognition and sorting equipment according to claim 3, characterized in that, The step of the color matching unit matching the color feature with a preset color number includes: First, the average value of the color features of all pixels in the mask area output by the color feature extraction unit is calculated to obtain the quantized color value of the current yarn package. Subsequently, the color difference distance between the quantified color value and each standard color number in the preset color number library is calculated, and the standard color number with the smallest color difference distance and less than the preset threshold is selected as the final color number of the yarn package. The preset color code library is a configurable database that allows users to adjust the standard color codes according to the needs of production batches.

Images