WO2021142606A1

WO2021142606A1 - Artificial intelligence-based leather inspection method and leather product production method

Info

Publication number: WO2021142606A1
Application number: PCT/CN2020/071902
Authority: WO
Inventors: 张育斌
Original assignee: 卓峰智慧生态有限公司
Priority date: 2020-01-14
Filing date: 2020-01-14
Publication date: 2021-07-22

Abstract

An artificial intelligence-based leather inspection method and a leather product production method. The method comprises: acquiring leather data of a leather raw material (10); and inputting the leather data into an artificial intelligence module (20), performing calculation, and determining a defective region (22) and a non-defective region (24) of the leather raw material (10). In the subsequent production method, regional data of the leather raw material (10) is established and used to define a maximum quantity of reserved regions (32) in the non-defective region (24), such that the leather raw material (10) can be cut into leather components corresponding to the respective reserved regions (32). The invention can significantly reduce the time required for leather inspection, establish consistent and universal leather quality inspection criteria, realize an automated production process of leather products, and improve overall production efficiency.

Description

Artificial intelligence-based leather detection method and leather product production method

Technical field

The invention relates to leather, in particular to a leather detection method based on artificial intelligence and a leather product production method.

Background technique

Leather can be used in all kinds of people's livelihood products, such as clothing, leather bags, suitcases or decorative accessories, etc., which are everyday items that are often used. And because natural leather (also known as genuine leather) has good touch and durability, high-priced and high-value products often use natural leather as the main material.

Natural leather is susceptible to factors such as the original source environment or manufacturing process, such as animal injury, mold growth, pests, rupture, or transportation scratches that cause damage and defects to the leather surface or internal tissues. In order for leather products to be able to detect the aforementioned defects before production, most of the time, when the leather is still in the raw material state, the leather is first inspected in detail by human visual or hand inspection, and then the aforementioned defects are marked on the leather surface. For the defective parts, according to the demand characteristics of the finished product, the marked leather cutting is classified into usable leather parts for subsequent production.

However, the aforementioned inspection methods using human visual inspection or manual inspection not only take time, but also must have sufficient experience to judge defects. The training and development process of inspectors is long and difficult, because the judgment method is more dependent. Subjective visual or touch inspections are vulnerable to factors such as personal emotions, environment, or time and space, and cannot establish more consistent and universal quality standards.

Invention Disclosure

Therefore, the main purpose of the present invention is to provide an artificial intelligence-based leather inspection method and leather product production method, which can greatly reduce the leather inspection time, establish a consistent and universal leather quality inspection standard, and realize the automated production process of leather products. , Improve overall production efficiency.

In order to achieve the above objective, the artificial intelligence-based leather detection method provided by the present invention includes obtaining leather data of the leather raw material first, and then inputting the leather data to an artificial intelligence module to determine the defective area and non-defective area of the leather raw material.

The present invention provides a leather detection method based on artificial intelligence, including: a. laying leather raw materials flat on a leather detection platform; b. using a detection component to produce a relative displacement with the leather raw materials set on the leather detection platform, so that the The detection component obtains the leather data of the leather raw material; and c. Inputs the leather data to the artificial intelligence module to determine the defective area and the non-defective area of the leather raw material.

More preferably, first obtain the partial leather data of the leather raw material in different parts, and then integrate all the partial leather data to become the complete leather data of the leather raw material.

More preferably, the area data of the leather raw material is established, and the area data is used to define at least one reserved area of the non-defective area.

More preferably, light is projected toward the leather raw material, and the lighting characteristics of the light can be adjusted correspondingly according to the material characteristics of the leather raw material.

More preferably, the artificial intelligence module includes a deep learning model.

More preferably, the detection component is used to obtain the image or surface state of the leather raw material to form the leather data.

More preferably, the detection component first obtains the partial leather data of the leather raw material at different locations, and then integrates all the partial leather data to become the leather data of the leather raw material.

More preferably, light is projected toward the leather raw material, and the lighting characteristics of the light are adjusted correspondingly according to the material characteristics of the leather raw material.

More preferably, the detection component includes a plurality of linear arrays and moves along the top of the leather detection platform, and then scans the leather raw material line by line to form leather data.

The present invention provides a leather product production method, comprising: a. using the above leather detection method; and b. cutting the leather raw material to produce leather parts that can correspond to the defective area and the non-defective area.

More preferably, it further includes cutting the non-defective area to produce a plurality of leather parts.

The detailed features, steps or application methods provided by the present invention will be described in the following detailed description of the embodiments in conjunction with the accompanying drawings. However, those with ordinary knowledge in the field of the present invention should be able to understand that the detailed description and the specific embodiments listed for implementing the present invention are only used to illustrate the present invention, not as a limitation to the present invention.

Brief description of the drawings

Fig. 1 is a structural diagram of a preferred embodiment of the present invention.

Figure 2 is a schematic diagram of a preferred embodiment of the present invention, which mainly shows a leather data collection device.

Fig. 3 is a schematic diagram of a preferred embodiment of the present invention, which mainly shows another implementation aspect of the leather data collection device.

Figure 4 is a schematic diagram of a preferred embodiment of the present invention, which mainly shows a leather detection platform.

Figure 5 is similar to Figure 4 and mainly shows that the leather raw materials are set on the leather inspection platform.

Fig. 6 is a schematic diagram of a preferred embodiment of the present invention, which mainly shows the defective area of the leather raw material.

FIG. 7 is a schematic diagram of a preferred embodiment of the present invention, which mainly shows the state of layout of the non-defective areas of the leather raw material.

Figure 8 is a schematic diagram of a preferred embodiment of the present invention, which mainly shows the state of the leather raw material after being cut.

Fig. 9 is a schematic diagram of a preferred embodiment of the present invention, which mainly shows another implementation aspect of the leather data collection device.

Among them, the reference signs:

10Leather raw materials 12Leather testing platform

14 Leather data collection device 15 rails

16 Inspection components 17 light source

18 Leather data processing device 20 Artificial intelligence module

22 Defective Area 24 Non-Defective Area

30 Typesetting module 32 Reserved area

40 Cutting device 50 Leather parts

The best way to implement the invention

The first thing to note is that the artificial intelligence-based leather detection method and leather product production method provided by the present invention can be widely used to detect various types or surface treatments of natural leather or synthetic leather. Those skilled in the art should be able to understand this In the implementation, the artificial intelligence, operating instructions, and operating steps belong to the upper-level description that does not limit the specific calculation model, technical field, or operating sequence, and the quantitative term "one" includes one and more than one. Number of components.

Please refer to Figures 1 to 4, the artificial intelligence-based leather detection method provided by the present invention mainly includes the following steps:

1. Collecting data: A leather raw material 10 is set on a leather testing platform 12, and then the leather data collection device 14 provided on the leather testing platform 12 is used to obtain the leather data of the leather raw material 10.

In this preferred embodiment, the leather material 10 is natural cowhide as an example. Of course, it can also be applied to other types of leather. The leather data collection device 14 of this preferred embodiment includes a device that can capture surface images of the leather material 10 The optical detection component is taken as an example. The leather data collection device 14 photographs the surface of the leather material 10 to obtain a digital image of the leather surface to form leather data for judging the edge and surface condition of the leather material 10.

As shown in FIG. 2, the leather data collection device 14 may include a plurality of detection components 16 evenly arranged in an array above the leather raw material 10, and each detection component 16 obtains partial leather data of the leather raw material 10 at different positions. The leather data collection device 14 additionally includes a light source 17 that projects light toward the leather material 10 so that the detection component 16 can obtain complete and clear digital image leather data. The light source 17 can be a point light source or an array light source, and in order to match different types, different material processing methods, or different surface textures of the leather raw materials 10, the light source 17 can also project different lighting according to the material characteristics of the different leather raw materials 10 The characteristics, that is, the intensity of light, illuminance, or brightness can be adjusted in accordance with the leather material 10.

2. Data processing: As shown in Figures 4 and 5, when the leather raw material 10 is placed flat on the leather detection platform 12, and the leather data collection device 14 above the leather detection platform 12 obtains the image leather data, the leather data will be Input the leather data processing device 18 to perform calculation procedures. The leather data processing device 18 of the preferred embodiment at least includes an image processing module, which can splice and integrate the partial leather data obtained by the leather data collection device 14 into complete leather data .

The leather data processing device 18 further includes an artificial intelligence module 20. In the preferred embodiment, the artificial intelligence module 20 (Artificial Intelligence Model) includes a Deep Learning Model as an example to calculate and determine the leather raw materials. 10 Defective areas 22 and non-defective areas 24 on the surface.

3. Generate process data: As shown in Figure 1, Figure 6 to Figure 8, after the complete leather data is judged by the artificial intelligence module 20 of the leather data processing device 18, the defective area 22 and the non-defective area 24, a typesetting module is used 30 Establish the area data of the leather raw material 10. The area data is used to define the non-defective area 24 to define at least one reserved area 32 to provide a subsequent leather product production method. A cutting device 40 can be used to cut the leather raw material 10 to generate corresponding reserved areas. 32 of the leather parts 50.

With the above leather detection method and leather product production method, the present invention has at least the following technical effects:

1. The use of an artificial intelligence module with a deep learning model eliminates the need for humans to judge the defects of the leather, which greatly reduces the inspection time of the leather.

2. There is no need to consider the detection environment, time or human factors, the present invention can quickly complete the detection, and establish a consistent and universal leather quality inspection standard.

3. The leather detection method combined with the subsequent typesetting and cutting process can more effectively use leather raw materials and increase the utilization rate of leather raw materials.

4. The present invention can integrate the quality inspection steps of leather raw materials to the subsequent cutting process, and realize the automatic production process of leather products.

It is worth mentioning that the above leather data collection device can also be a device that uses transmission or mechanical force to produce a kneading effect on the leather raw material to obtain the internal organization or material state of the leather raw material, for example, by irradiating the leather raw material with an X-ray device X-rays can obtain the signal change state of the X-rays after passing through the leather raw materials through X-rays, which can be used to learn characteristic data such as the internal organization of the leather raw materials.

As shown in FIG. 3, the leather data collection device 14 can also scan the leather surface line by line with the leather material 10 moving along the conveyor belt 13 on the leather inspection platform 12 to form leather data, which can further improve the overall inspection and production efficiency.

Or as shown in FIG. 9, the leather data collection device 14 includes a plurality of optical detection components 16 arranged at intervals on a linear guide 15. All the detection components 16 of the leather data collection device 14 are along the top of the leather detection platform 12 at the same time. Moving, you can gradually scan the leather surface line by line for the leather raw materials 10 to form leather data, which will further improve the resolution and wide depth of the leather data, which will help determine more subtle defects and increase the learning results and efficiency of artificial intelligence. To achieve the purpose of the invention.

In addition, in addition to the deep learning model, the artificial intelligence module can also use other machine learning models such as neural network models, convolutional network models, or recurrent neural network models to enhance the accuracy and accuracy of artificial intelligence judgments. The purpose of each invention.

Of course, the present invention can also have various other embodiments. Without departing from the spirit and essence of the present invention, those skilled in the art can make various corresponding changes and modifications according to the present invention, but these corresponding All changes and deformations shall belong to the protection scope of the appended claims of the present invention.

Industrial applicability

The invention discloses a leather detection method based on artificial intelligence and a leather product production method. The leather data of the leather raw material is first obtained, and then the leather data is input to the artificial intelligence module to calculate and judge the defective area and the non-defective area of the leather raw material, and then It is used in subsequent production methods to establish area data of leather raw materials in non-defective areas. The area data can define most reserved areas, which can be used as cutting leather raw materials to produce leather parts corresponding to each reserved area. The invention can greatly reduce the leather inspection time, establish a consistent and universal leather quality inspection standard, realize the automated production process of leather products, and improve the overall production efficiency.

Claims

An artificial intelligence-based leather detection method, which is characterized in that it includes:

a. Lay the leather raw materials flat on the leather testing platform;

b. Use the detection component to generate a relative displacement with the leather raw material set on the leather detection platform, so that the detection component obtains the leather data of the leather raw material; and

c. Input the leather data to the artificial intelligence module to determine the defective area and non-defective area of the leather raw material.
The artificial intelligence-based leather detection method according to claim 1, wherein the step a is to first obtain partial leather data of the leather raw material at different locations, and then integrate all the partial leather data to become the leather data of the leather raw material .
The artificial intelligence-based leather inspection method according to claim 1, further comprising establishing area data of the leather raw material, and the area data is used to typeset the non-defective area to define at least one reserved area.
The artificial intelligence-based leather detection method according to claim 1, wherein the artificial intelligence module includes a deep learning model.
The artificial intelligence-based leather detection method according to claim 1, wherein the step b uses the detection component to obtain an image or surface state of the leather raw material to form the leather data.
The artificial intelligence-based leather detection method of claim 5, wherein the detection component first obtains partial leather data of the leather raw material at different locations, and then integrates all the partial leather data to become the leather data of the leather raw material .
The artificial intelligence-based leather detection method according to claim 5, wherein a light is projected toward the leather raw material, and the lighting characteristics of the light are adjusted correspondingly according to the material characteristics of the leather raw material.
The artificial intelligence-based leather detection method according to claim 1, wherein the detection component is arranged in a linear manner and moves along the top of the leather detection platform, and then scans the leather raw materials step by step to form leather data.
A method for producing leather products, characterized in that it comprises:

a. Using the leather detection method of any one of claims 1 to 7; and

b. Cut the leather raw material to produce leather parts that can correspond to the defective area and the non-defective area.
9. The leather product production method of claim 9, further comprising cutting the non-defective area to produce a plurality of leather parts.