Woven fabric quality detection method and device

Abstract

The invention discloses a woven fabric quality detection method and device. The woven fabric quality detection method comprises the following steps: acquiring a cloth image of cloth in a target area; performing image recognition on the cloth image to obtain feature points in the cloth image; whether flaws exist in the cloth or not is judged through the feature points, if yes, preset operation is executed, and the preset operation comprises prompting and / or recording the positions of the flaws. According to the method, the flaws of the knitted cloth can be automatically detected, so that the loom is intelligent, the production efficiency and the product percent of pass are greatly improved, and meanwhile, the production cost can be reduced.

Description

Technical Field

[0001] The invention relates to a method and a device for detecting woven cloth quality. Background Art

[0002] A loom is a mechanical device used to weave cloth. It forms cloth by interweaving the warp yarns running lengthwise and the weft yarns running widthwise. The working principle of a loom is based on the basic principle of weaving, which is to interweave two sets of yarns to form cloth. Modern looms can be divided into many types, including manual looms, automatic looms, electronic looms, etc.

[0003] Traditional manual looms usually consist of a wooden frame and several main parts, such as the shuttle, warp frame, and weft frame. Modern looms are more complex and often use electric motors and computer control systems to improve production efficiency and cloth quality. They can produce a variety of different types of cloth, from simple cotton cloth to complex high-tech fabrics.

[0004] Existing looms have a low level of intelligence and require manual inspection of cloth defects. Summary of the invention

[0005] The technical problem to be solved by the present invention is to overcome the defects of the prior art that the loom has a low degree of intelligence and requires manual detection of cloth defects, and to provide a method and device for detecting cloth quality that can realize automatic detection of knitted cloth defects, make the loom intelligent, greatly improve production efficiency and product qualification rate, and at the same time reduce production costs.

[0006] The present invention solves the above technical problems through the following technical solutions:

[0007] A method for detecting the quality of woven fabrics, characterized in that the method comprises:

[0008] Acquire a cloth image of the cloth in the target area;

[0009] Perform image recognition on cloth images to obtain feature points in the cloth images;

[0010] The feature points are used to determine whether the cloth has defects. If so, a preset operation is performed, which includes making a prompt and / or recording the defect location.

[0011] There are three options for making prompts:

[0012] 1. Notify the mobile phone APP to make abnormal alarms through network communication;

[0013] 2. Alarm through external sound and light alarm;

[0014] 3. Access the MES system or computer through the network cable to display abnormal alarm pop-ups on the office operation interface.

[0015] Preferably, the weaving quality detection method is implemented by a weaving quality detection device, which includes a computing module, a camera module, a support module and a fill light module, wherein the support module is used to support the camera module and the fill light module, the camera is used to aim at the target area, the target area is the cloth outlet area of ​​the loom, the computing module is used to collect the cloth image sent by the camera module, and implement the weaving quality detection method.

[0016] Preferably, the fill light module includes at least one fill light, and the fabric quality detection method includes:

[0017] Controlling the at least one fill light to perform fill light at more than two fill light positions;

[0018] Collect the cloth image when the fill light is at each fill light position;

[0019] The defect position of the cloth is obtained by using the pixel difference value between the cloth images at each fill light position.

[0020] Preferably, the fabric quality detection method comprises:

[0021] Collecting cloth images when filling light at at least two filling light positions;

[0022] Aligning cloth images corresponding to at least two fill light positions according to the weaving speed of the loom;

[0023] Superimpose the pixel values ​​of all aligned cloth images;

[0024] Identify the feature points of defects in the cloth image after superimposing pixel values.

[0025] Preferably, the number of the fill lights is two, and the angle between the irradiation directions of the two fill lights is greater than zero degrees. The fabric quality detection method includes:

[0026] The feature points are used to determine whether the cloth has defects. If so, the position coordinates of the defect position on the cloth image are calculated according to the irradiation direction of the fill light;

[0027] The position of the defect on the cloth is obtained according to the weaving speed of the loom, the shooting time of the cloth image and the position coordinates.

[0028] Preferably, the fabric quality detection method comprises:

[0029] The position of the defect on the cloth, the frequency of the defect and the repetition of the defect are used to determine whether the loom meets the shutdown requirements. If so, the loom is shut down and a prompt is given.

[0030] Preferably, the fabric quality detection method comprises:

[0031] After fixing the fill light, camera module and loom position, add a calibration object on the cloth;

[0032] Collect at least two images of the cloth with calibration objects;

[0033] Acquire a cloth image of the calibration object and its shadow in the cloth image with the calibration object;

[0034] The positional relationship between the fixed fill light, the camera module and the loom is obtained by using the cloth image of the calibration object and its shadow;

[0035] The above positional relationship is used to calculate the position coordinates of the defect position on the cloth image.

[0036] Preferably, the fabric quality detection method includes an image training set, and the fabric quality detection method includes:

[0037] Marking the defect location of each training image in the image training set;

[0038] Training an artificial intelligence algorithm using training images with defect locations annotated;

[0039] Using an artificial intelligence algorithm to obtain feature points in the cloth image;

[0040] Wherein, the fabric quality detection method further comprises:

[0041] Obtain the category of defects based on the shape of feature points;

[0042] Determine whether the defect category meets the shutdown requirements, and if so, shut down the machine and give a prompt.

[0043] The present invention also provides a woven fabric quality detection device, which is characterized in that the woven fabric quality detection device is used to implement the woven fabric quality detection method as described above.

[0044] Preferably, the weaving quality detection device includes a computing module, a camera module, a support module and a fill light module, the support module is used to support the camera module and the fill light module, the camera is used to aim at the target area, the target area is the cloth output area of ​​the loom, the computing module is used to collect the cloth image sent by the camera module, and implement the weaving quality detection method.

[0045] On the basis of being in accordance with the common sense in the art, the above-mentioned preferred conditions can be arbitrarily combined to obtain the preferred embodiments of the present invention.

[0046] The positive and progressive effects of the present invention are:

[0047] The present invention can realize automatic detection of knitted fabric defects, make the loom intelligent, greatly improve production efficiency and product qualification rate, and also can reduce production costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 This is a flow chart of a method for detecting fabric quality according to Embodiment 1 of the present invention. DETAILED DESCRIPTION

[0049] The present invention is further described below by way of examples, but the present invention is not limited to the scope of the examples.

[0050] Example 1

[0051] This embodiment provides a woven fabric quality detection device.

[0052] The fabric quality detection device comprises a computing module, a camera module, a supporting module and a fill light module.

[0053] The supporting module is used to support the camera module and the fill light module.

[0054] The camera is used to aim at the target area, and the target area is the cloth output area of ​​the loom.

[0055] The computing module is used to collect the cloth images sent by the camera module and implement the fabric quality detection method.

[0056] Specifically, the camera module is used to obtain cloth images of cloth in the target area;

[0057] The computing module is used to perform image recognition on the cloth image and obtain feature points in the cloth image;

[0058] The operation module is also used to use the feature points to determine whether the cloth has defects, and if so, to execute a preset operation, which includes making a prompt and recording the defect location.

[0059] Furthermore, the fill light module includes at least one fill light.

[0060] The computing module is used to control the at least one fill light to perform fill light at more than two fill light positions;

[0061] The computing module is used to collect cloth images when the fill light is filling light at each fill light position;

[0062] The computing module is used for obtaining the defect position of the cloth by using the pixel difference value between the cloth images at each fill light position.

[0063] The computing module is used for:

[0064] Collecting cloth images when filling light at at least two filling light positions;

[0065] Aligning cloth images corresponding to at least two fill light positions according to the weaving speed of the loom;

[0066] Superimpose the pixel values ​​of all aligned cloth images;

[0067] Identify the feature points of defects in the cloth image after superimposing pixel values.

[0068] The number of the fill lights is two, and the angle between the irradiation directions of the two fill lights is greater than zero degree.

[0069] The computing module is used for:

[0070] The feature points are used to determine whether the cloth has defects. If so, the position coordinates of the defect position on the cloth image are calculated according to the irradiation direction of the fill light;

[0071] The position of the defect on the cloth is obtained according to the weaving speed of the loom, the shooting time of the cloth image and the position coordinates.

[0072] The computing module is used for:

[0073] The position of the defect on the cloth, the frequency of the defect and the repetition of the defect are used to determine whether the loom meets the shutdown requirements. If so, the loom is shut down and a prompt is given.

[0074] Fabric quality inspection device is used for:

[0075] After fixing the fill light, camera module and loom position, add a calibration object on the cloth;

[0076] The camera module is used to capture at least two cloth images with calibration objects.

[0077] The computing module is used for:

[0078] Acquire a cloth image of the calibration object and its shadow in the cloth image with the calibration object;

[0079] The positional relationship between the fixed fill light, the camera module and the loom is obtained by using the cloth image of the calibration object and its shadow;

[0080] The above positional relationship is used to calculate the position coordinates of the defect position on the cloth image.

[0081] The fabric quality detection device includes an image training set, and the computing module is used for:

[0082] Marking the defect location of each training image in the image training set;

[0083] Training an artificial intelligence algorithm using training images with defect locations annotated;

[0084] Using an artificial intelligence algorithm to obtain feature points in the cloth image;

[0085] Wherein, the fabric quality detection method further comprises:

[0086] Obtain the category of defects based on the shape of feature points;

[0087] Determine whether the defect category meets the shutdown requirements, and if so, shut down the machine and give a prompt.

[0088] See also Figure 1 , using the above-mentioned fabric quality detection device, this embodiment also provides a fabric quality detection method, including:

[0089] Step 100, obtaining a cloth image of the cloth in the target area;

[0090] Step 101: Perform image recognition on the cloth image to obtain feature points in the cloth image;

[0091] Step 102, using the feature points to determine whether the cloth has defects, if yes, execute step 103, if no, execute step 100 again.

[0092] Step 103: Execute preset operations, which include making prompts and recording defect locations.

[0093] The fill light module includes at least one fill light.

[0094] Step 100 specifically includes:

[0095] Controlling the at least one fill light to perform fill light at more than two fill light positions;

[0096] Collect the cloth image when the fill light is at each fill light position;

[0097] Step 103 includes: obtaining the defect position of the cloth by using the pixel difference value between the cloth images at each fill light position.

[0098] Step 100 specifically includes: collecting cloth images when filling light at at least two filling light positions;

[0099] Step 103 specifically includes:

[0100] Aligning cloth images corresponding to at least two fill light positions according to the weaving speed of the loom;

[0101] Superimpose the pixel values ​​of all aligned cloth images;

[0102] Identify the feature points of defects in the cloth image after superimposing pixel values.

[0103] The number of the fill lights is two, and the angle between the irradiation directions of the two fill lights is greater than zero degrees. The fabric quality detection method includes:

[0104] The feature points are used to determine whether the cloth has defects. If so, the position coordinates of the defect position on the cloth image are calculated according to the irradiation direction of the fill light;

[0105] The position of the defect on the cloth is obtained according to the weaving speed of the loom, the shooting time of the cloth image and the position coordinates.

[0106] The fabric quality detection method comprises:

[0107] The position of the defect on the cloth, the frequency of the defect and the repetition of the defect are used to determine whether the loom meets the shutdown requirements. If so, the loom is shut down and a prompt is given.

[0108] The fabric quality detection method further includes before step 100:

[0109] After fixing the fill light, camera module and loom position, add a calibration object on the cloth;

[0110] Collect at least two images of the cloth with calibration objects;

[0111] Acquire a cloth image of the calibration object and its shadow in the cloth image with the calibration object;

[0112] The positional relationship between the fixed fill light, the camera module and the loom is obtained by using the cloth image of the calibration object and its shadow;

[0113] Step 103 includes: using the above position relationship to calculate the position coordinates of the defect position on the cloth image.

[0114] The fabric quality detection method includes an image training set, and the fabric quality detection method includes:

[0115] Marking the defect location of each training image in the image training set;

[0116] Training an artificial intelligence algorithm using training images with defect locations annotated;

[0117] Using an artificial intelligence algorithm to obtain feature points in the cloth image;

[0118] Wherein, the fabric quality detection method further comprises:

[0119] Obtain the category of defects based on the shape of feature points;

[0120] Determine whether the defect category meets the shutdown requirements, and if so, shut down the machine and give a prompt.

[0121] Although the specific embodiments of the present invention are described above, those skilled in the art should understand that these are only examples, and the protection scope of the present invention is defined by the appended claims. Those skilled in the art may make various changes or modifications to these embodiments without departing from the principles and essence of the present invention, but these changes and modifications all fall within the protection scope of the present invention.

Claims

1. A method for detecting fabric quality, characterized in that: The fabric quality detection method comprises: Acquire a cloth image of the cloth in the target area; Perform image recognition on cloth images to obtain feature points in the cloth images; The feature points are used to determine whether the cloth has defects. If so, a preset operation is performed, which includes making a prompt and / or recording the defect location.

2. The method for detecting fabric quality according to claim 1, characterized in that: The weaving quality detection method is implemented through a weaving quality detection device, which includes a computing module, a camera module, a support module and a fill light module. The support module is used to support the camera module and the fill light module, the camera is used to aim at the target area, the target area is the cloth output area of ​​the loom, the computing module is used to collect the cloth image sent by the camera module, and implement the weaving quality detection method.

3. The method for detecting fabric quality according to claim 2, characterized in that: The fill light module includes at least one fill light, and the fabric quality detection method includes: Controlling the at least one fill light to perform fill light at more than two fill light positions; Collect the cloth image when the fill light is at each fill light position; The defect position of the cloth is obtained by using the pixel difference value between the cloth images at each fill light position.

4. The method for detecting fabric quality according to claim 3, characterized in that: The fabric quality detection method comprises: Collecting cloth images when filling light at at least two filling light positions; Aligning cloth images corresponding to at least two fill light positions according to the weaving speed of the loom; Superimpose the pixel values ​​of all aligned cloth images; Identify the feature points of defects in the cloth image after superimposing pixel values.

5. The method for detecting fabric quality according to claim 4, characterized in that: The number of the fill lights is two, and the angle between the irradiation directions of the two fill lights is greater than zero degrees. The fabric quality detection method includes: The feature points are used to determine whether the cloth has defects. If so, the position coordinates of the defect position on the cloth image are calculated according to the irradiation direction of the fill light; The position of the defect on the cloth is obtained according to the weaving speed of the loom, the shooting time of the cloth image and the position coordinates.

6. The method for detecting fabric quality according to claim 5, characterized in that: The fabric quality detection method comprises: The position of the defect on the cloth, the frequency of the defect and the repetition of the defect are used to determine whether the loom meets the shutdown requirements. If so, the loom is shut down and a prompt is given.

7. The method for detecting fabric quality according to claim 5, characterized in that: The fabric quality detection method comprises: After fixing the fill light, camera module and loom position, add a calibration object on the cloth; Collect at least two images of the cloth with calibration objects; Acquire a cloth image of the calibration object and its shadow in the cloth image with the calibration object; The positional relationship between the fixed fill light, the camera module and the loom is obtained by using the cloth image of the calibration object and its shadow; The above positional relationship is used to calculate the position coordinates of the defect position on the cloth image.

8. The method for detecting fabric quality according to claim 1, characterized in that: The fabric quality detection method includes an image training set, and the fabric quality detection method includes: Marking the defect location of each training image in the image training set; Training an artificial intelligence algorithm using training images with defect locations annotated; Using an artificial intelligence algorithm to obtain feature points in the cloth image; Wherein, the fabric quality detection method further comprises: Obtain the category of defects based on the shape of feature points; Determine whether the defect category meets the shutdown requirements, and if so, shut down the machine and give a prompt.

9. A woven fabric quality detection device, characterized in that: The woven fabric quality detection device is used to implement the woven fabric quality detection method as described in any one of claims 1 to 8.

10. The fabric quality detection device according to claim 9, characterized in that: The fabric quality detection device includes a computing module, a camera module, a support module and a fill light module. The support module is used to support the camera module and the fill light module. The camera is used to aim at the target area, which is the cloth outlet area of ​​the loom. The computing module is used to collect the cloth image sent by the camera module and implement the fabric quality detection method.

Images