Animal skin cutting method and system

Abstract

The invention discloses an animal skin cutting method and system, and the method comprises the steps: obtaining a tiled animal skin image, and carrying out the preprocessing of the image; obtaining an animal skin contour and key features from the preprocessed image through a deep learning target detection algorithm; determining a cutting path according to the key features; and the cutting equipment carries out cutting according to the cutting path. The cutting route is judged by obtaining the contour of the animal skin, then accurate cutting is conducted through the cutting equipment according to the cutting route, the cutting efficiency and the cutting precision are improved, and the cutting quality is guaranteed.

Description

Technical Field

[0001] The invention relates to the field of animal hide processing, in particular to an animal hide cutting method and system. Background Art

[0002] In the animal skin processing industry, the pre-cutting work is a crucial link. The core goal of this process is to accurately cut the limbs and neck of the animal, laying a good foundation for the subsequent animal skin shaping and processing steps. The accuracy and efficiency of cutting directly affect the quality of the final product and the overall efficiency of the production line.

[0003] However, the most commonly used cutting method is still manual operation. Although manual cutting is flexible to a certain extent, it faces many challenges in actual operation. First of all, manual cutting is prone to over-cutting or under-cutting, which is caused by the difficulty in maintaining the accuracy and consistency of manual operation for a long time. This uneven cutting quality not only affects the subsequent processing and forming effect of animal skins, but also may cause waste of raw materials. At the same time, the labor intensity of workers is high and the efficiency is low. Summary of the invention

[0004] Therefore, in order to solve the above-mentioned shortcomings, the present invention provides an animal skin cutting method and system. The present invention determines the cutting route by obtaining the contour of the animal skin, and then accurately cuts according to the cutting route through a cutting device, thereby improving the cutting efficiency and cutting accuracy and ensuring the cutting quality.

[0005] In one aspect, the present invention provides an animal skin cutting method, comprising the following steps: Obtain the tiled animal skin image and pre-process the image; The animal skin contour and key features are obtained from the preprocessed images through deep learning object detection algorithm; Determine cutting paths based on key features; The cutting device performs cutting according to the cutting path.

[0006] Alternatively, the method for obtaining the tiled animal skin image is: The animal skin is spread flat on a working platform, and an image of the animal skin is captured by taking a photo with one or more cameras; Methods for preprocessing images include: Distort the image or stitch it after distortion; Then, the distorted or spliced ​​image is subjected to noise reduction.

[0007] Preferably, the key feature is one or more of the left forelimb, right forelimb, left hind limb, right hind limb, neck, tail, and designated parts on the animal skin, and the designated parts are artificially marked parts, such as unnecessary parts of the abdomen of the animal skin.

[0008] Alternatively, a method for determining the cutting path based on key features is: A proposed cutting path is formed according to the cutting sequence of each key feature, wherein the cutting sequence is pre-set or calculated; after the proposed cutting path is formed, the path can be manually confirmed, reviewed or corrected to obtain the cutting path; Determine the cutting coordinates and actual cutting line based on the cutting path.

[0009] Optionally, the cutting sequence is calculated by: All cutting sequences that can be formed by each key feature are obtained, a path with the shortest stroke is found in each cutting sequence, and the path sequence with the shortest stroke is determined as the cutting sequence, and the path with the shortest stroke is determined as the cutting path.

[0010] Optionally, the method for determining the cutting coordinates and the actual cutting line according to the cutting path is: The pixel points on the cutting path are multiplied by the distortion coefficient to obtain the actual cutting line on the animal skin; wherein the cutting coordinates are the coordinates of the pixels passed through in the actual cutting path, and the coordinates are obtained by multiplying the pixel points in the proposed cutting path by the distortion coefficient.

[0011] Optionally, the cutting device is a laser cutting machine or a shearing machine or a water jet cutting machine or an existing tool capable of cutting animal skins, wherein the method in which the cutting device performs cutting according to the cutting path is: After determining the cutting coordinates and the actual cutting route, the cutting device is positioned by the coordinates, and the cutting device then follows the actual cutting route to cut the animal skin.

[0012] In another aspect, the present invention provides an animal hide cutting system, comprising: An image acquisition unit, used to acquire the flattened animal skin image and pre-process the image; An extraction unit, for obtaining animal skin contours and key features from the preprocessed image using a deep learning object detection algorithm; A cutting path forming unit, used for determining a cutting path according to key features; The cutting unit is used for the cutting device to perform cutting according to the cutting path.

[0013] Optionally, the path determination module includes: A simulated cutting path unit, used to form a simulated cutting path according to the cutting sequence of each key feature, wherein the cutting sequence is pre-set or calculated; A review unit is used to manually confirm, review or modify the proposed cutting path after the proposed cutting path is formed, so as to obtain the cutting path; Determination unit, used to determine the cutting coordinates and actual cutting line according to the cutting path Optionally, the animal skin cutting system further comprises an identification module, which is used for identifying the user, and is used for entering the system only after inputting correct identification information; the identification information is a user name and a password.

[0014] The present invention has the following advantages: The present invention provides a method and system for cutting animal skins, which determines the cutting route by acquiring the contour of the animal skin, and then accurately cuts according to the cutting route through a cutting device, thereby improving the cutting efficiency and cutting accuracy and ensuring the cutting quality; compared with traditional manual cutting, the efficiency is better and the quality is more stable. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 is a schematic flow chart of the animal skin cutting method of the present invention; Figure 2 is a systematic block diagram of the animal skin cutting system of the present invention; Figure 3 is a screenshot of the interface of the animal skin cutting system of the present invention; Figure 4 This is a screenshot of the user interface of the animal skin cutting system of the present invention. DETAILED DESCRIPTION

[0016] Embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.

[0017] In this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprises", or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or apparatus that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article, or apparatus.

[0018] As described in the background technology, the most commonly used cutting method is still manual operation. Although manual cutting is flexible to a certain extent, it faces many challenges in actual operation. First of all, manual cutting is prone to over-cutting or under-cutting, which is caused by the difficulty in maintaining the accuracy and consistency of manual operation for a long time. This uneven cutting quality not only affects the subsequent processing and forming effect of animal skins, but also may cause waste of raw materials. At the same time, the labor intensity of the staff is high and the efficiency is low.

[0019] Based on the above reasons, this embodiment provides an animal skin cutting method, such as Figure 1 As shown, the method includes: Step S100, obtaining a tiled animal skin image and preprocessing the image; Step S200, obtaining the animal skin contour and key features from the preprocessed image through a deep learning target detection algorithm; Step S300, determine the cutting path according to the key features, wherein the key features are one or more of the left forelimb, right forelimb, left hind limb, right hind limb, neck, tail, and designated parts on the animal skin, wherein the designated parts are manually calibrated parts, such as unnecessary parts of the abdomen of the animal skin.

[0020] Step S400: The cutting device performs cutting according to the cutting path.

[0021] The above technology is characterized in that it can realize automatic cutting of animal skins, reduce the labor intensity of workers and improve work efficiency.

[0022] In one embodiment, in step S100, the tiled animal skin image is obtained, and the image is preprocessed in the following manner: The animal skin is spread flat on a work platform, and an image of the animal skin is collected by taking photos with one or more cameras; the image is then preprocessed, and the preprocessed image is subjected to noise reduction; the preprocessing includes distorting or stitching the image. When two or more cameras take photos, stitching is required. In this method, the preferred number of cameras is 2.

[0023] In the above technical features, since the principle of the camera is not to reduce the image to an imaging system equidistantly, the farther away from the focus point, the greater the reduction ratio. The camera focuses on the center. The farther away from the center, the greater the difference between the distance between pixels and the actual distance, and the image is distorted. In order to ensure the accuracy of the cutting trajectory, it is necessary to set the distortion parameters of the image. The distortion can be eliminated through image distortion processing to ensure that the distance between pixels measured on the image is consistent with the actual distance.

[0024] The process of setting the image distortion parameters is: During the equipment debugging stage, the image is calibrated with positioning paper. There are multiple small dots with fixed positions on the positioning paper. The pixel positions of multiple small dots in the photo are found by shooting with the camera, and then the distortion coefficient of each pixel is determined by comparing with the actual position of the small dots. This step is performed during the equipment installation and debugging, and is no longer used in subsequent normal operation.

[0025] At the same time, image distortion processing can also be achieved through other existing technologies.

[0026] The above-mentioned image stitching process is because some animal skins are large, such as cowhide and horsehide. In order to ensure a higher resolution, one camera cannot fully image them, so multiple cameras are required to work together. At this time, there will be multiple photos. In order to facilitate processing, multiple photos need to be stitched in advance, and the common parts of multiple photos need to be overlapped to form a picture consistent with the real object. The image stitching process is implemented and processed by existing technology, that is, when two cameras are used, the photos taken by the two cameras are overlapped by the distance of the relative positions through the relative positions of the two cameras.

[0027] The purpose of denoising an image is to remove noise such as blood stains inside the animal skin and highlight the color difference between the animal skin and the background, so that the subsequent deep learning model can accurately find the outline of the animal skin.

[0028] In one embodiment, the deep learning target detection algorithm described in step S200 is based on a deep neural network, and can be implemented, for example, by deep learning models such as Faster R-CNN, YOLO and SSD. Before the preprocessed image is calculated by the deep learning model, the deep learning model needs to be trained. The training method is to form a target data set, which includes a certain number of target images, and the target image is an image with an animal skin contour and annotated with key features, wherein the key feature is the left forelimb, or the right forelimb, or the left hind limb, or the right hind limb, or the neck, or the tail of the animal skin; and then send the target data set to the deep learning model for training. When the deep learning model has recognition capability, the denoised image is sent to the trained deep learning model for calculation.

[0029] The Faster R-CNN algorithm is one of the classic algorithms in the field of target detection. It consists of two parts: Region Proposal Network (RPN) and target classification network. RPN is used to generate bounding boxes of candidate targets, and then classify and locate the candidate targets through the target classification network. The YOLO (You Only Look Once) algorithm is a target detection algorithm based on single-stage detection. It converts the target detection problem into a regression problem and directly predicts the category and location information of the target through a convolutional neural network. The YOLO algorithm has the advantages of real-time and high efficiency, and is suitable for scenarios with high real-time requirements. The SSD (Single Shot MultiBox Detector) algorithm is a multi-scale target detection algorithm that achieves effective detection of targets of different scales by detecting on feature maps at different levels. The SSD algorithm has a simple structure and high detection accuracy and efficiency.

[0030] In one embodiment, the method for determining the cutting path according to the key features in step S300 is: A proposed cutting path is formed according to the cutting sequence of each key feature, and the cutting sequence is pre-set or calculated; after the proposed cutting path is formed, the path can be manually confirmed, reviewed or corrected to obtain the final cutting path.

[0031] The method of forming the cutting sequence by calculation is: obtaining all cutting sequences that can be formed by each key feature, finding the path with the shortest stroke in each cutting sequence, and determining the path sequence with the shortest stroke as the final cutting sequence, and the corresponding path is the proposed cutting path.

[0032] Determine the cutting coordinates and actual cutting line according to the proposed cutting path.

[0033] Exemplarily, the pixel points on the intended cutting path are multiplied by the distortion coefficient to obtain the actual cutting line on the animal skin; wherein the cutting coordinates are the coordinates of the pixels passed through in the actual cutting path, which are obtained by multiplying the pixel points in the intended cutting path by the distortion coefficient.

[0034] In one embodiment, in step S400, the cutting device is a laser cutting machine or a shearing machine or other cutting mechanism capable of cutting animal skin. After determining the cutting coordinates and the actual cutting line, the cutting device is positioned by the coordinates, and the cutting device then follows the actual cutting line to split the animal skin.

[0035] The animal skin cutting method of the present invention determines the cutting route by obtaining the contour of the animal skin, and then accurately cuts according to the cutting route through a cutting device, thereby improving the cutting efficiency and cutting accuracy and ensuring the cutting quality; compared with traditional manual cutting, the efficiency is better and the quality is more stable.

[0036] In another embodiment, if Figure 2-Figure 4 As shown, an animal hide cutting system is provided, comprising: An image acquisition unit, used to acquire the flattened animal skin image and pre-process the image; An extraction unit, for obtaining animal skin contours and key features from the preprocessed image using a deep learning object detection algorithm; A cutting path forming unit, used for determining a cutting path according to key features; The cutting unit is used for the cutting device to perform cutting according to the cutting path.

[0037] Furthermore, the path determination module includes: A simulated cutting path unit, used to form a simulated cutting path according to the cutting sequence of each key feature, wherein the cutting sequence is pre-set or calculated; A review unit is used to manually confirm, review or modify the proposed cutting path after the proposed cutting path is formed, so as to obtain the cutting path; A determination unit, used to determine the cutting coordinates and the actual cutting line according to the cutting path; Furthermore, the animal skin cutting system also includes an identification module, which is used for identifying the user, and is used for entering the system only after the correct identification information is input; the identification information is a user name and a password.

[0038] The animal skin cutting system is used to implement the above-mentioned animal skin cutting method, and the specific embodiment steps are not repeated here.

[0039] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may 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 the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for cutting animal skins, comprising the following steps: Obtain the tiled animal skin image and pre-process the image; The animal skin contour and key features are obtained from the preprocessed images through deep learning object detection algorithm; Determine cutting paths based on key features; The cutting device performs cutting according to the cutting path.

2. The animal skin cutting method according to claim 1, characterized in that: The way to get the tiled animal skin image is: The animal skin is spread flat on a working platform, and an image of the animal skin is captured by taking a photo with one or more cameras; Methods for preprocessing images include: Distort the image or stitch it after distortion; Then, the distorted or spliced ​​image is subjected to noise reduction.

3. The animal skin cutting method according to claim 1, characterized in that: The key features are one or more of the left forelimb, right forelimb, left hind limb, right hind limb, neck, tail, and designated parts on the animal skin.

4. The animal skin cutting method according to claim 1, characterized in that: The method to determine the cutting path based on key features is: Forming a proposed cutting path according to the cutting sequence of each key feature, wherein the cutting sequence is pre-set or calculated; After the proposed cutting path is formed, the cutting path is obtained by manually confirming, reviewing or correcting the path; Determine the cutting coordinates and actual cutting line based on the cutting path.

5. The animal skin cutting method according to claim 4, characterized in that: The cutting sequence is calculated by: All cutting sequences that can be formed by each key feature are obtained, a path with the shortest stroke is found in each cutting sequence, and the path sequence with the shortest stroke is determined as the cutting sequence, and the path with the shortest stroke is determined as the cutting path.

6. The method for cutting animal skin according to claim 4, characterized in that: The method to determine the cutting coordinates and the actual cutting line according to the cutting path is: The pixel points on the cutting path are multiplied by the distortion coefficient to obtain the actual cutting line on the animal skin; wherein the cutting coordinates are the coordinates of the pixels passed through in the actual cutting path, and the coordinates are obtained by multiplying the pixel points in the proposed cutting path by the distortion coefficient.

7. The animal skin cutting method according to claim 1, characterized in that: The method in which the cutting device performs cutting according to the cutting path is: After determining the cutting coordinates and the actual cutting route, the cutting device is positioned by the coordinates, and the cutting device then follows the actual cutting route to cut the animal skin.

8. An animal skin cutting system, characterized in that: include: An image acquisition unit, used to acquire the flattened animal skin image and pre-process the image; An extraction unit that obtains the animal skin contour and key features from the preprocessed image using a deep learning object detection algorithm; A cutting path forming unit, used for determining a cutting path according to key features; The cutting unit is used for the cutting device to perform cutting according to the cutting path.

9. The animal skin cutting system according to claim 7, characterized in that: The path determination module comprises: A simulated cutting path unit, used to form a simulated cutting path according to the cutting sequence of each key feature, wherein the cutting sequence is pre-set or calculated; A review unit is used to manually confirm, review or modify the proposed cutting path after the proposed cutting path is formed, so as to obtain the cutting path; The determination unit is used to determine the cutting coordinates and the actual cutting line according to the cutting path.

Images