Woodworking machine tool safety monitoring method, system and equipment and storage medium

Abstract

The invention discloses a woodworking machine tool safety monitoring method, system and device and a storage medium, and the method comprises the steps: obtaining a real-time video through a camera disposed above a cutting saw blade of a woodworking machine tool, and generating a first video stream; performing frame extraction processing on the first video stream to generate a second video stream; inputting the second video stream into the trained YOLOv3 model frame by frame, performing target detection on the hand of the operator and the cutting saw blade, selecting the positions of the hand of the operator and the cutting saw blade on each frame, respectively generating a first position detection frame and a second position detection frame, and outputting the first position detection frame and the second position detection frame; the distance between the first position detection frame and the second position detection frame is monitored in real time, whether a dangerous condition or an emergency dangerous condition exists currently is judged according to the distance, and a dangerous operation warning is controlled to be sent out or the woodworking machine tool is controlled to stop suddenly. According to the invention, safety monitoring can be carried out on the woodworking machine tool, rapid sudden stop can be carried out in emergency dangerous situations, and the method has the advantages of high monitoring precision and high response speed.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and in particular to a woodworking machine tool safety monitoring method, system, equipment and storage medium. Background Art

[0002] With the development of society, people's living standards are constantly improving, and the requirements for wooden products such as furniture are getting higher and higher. Nowadays, woodworking machine tools are often used in wood processing. Woodworking machine tools cut wood through cutting saw blade devices to process the required wood.

[0003] However, existing woodworking machine tools often cause safety accidents to operators, and it is common for the cutting saw blade device on the woodworking machine tool to cause damage to the operator's hands. Therefore, there is an urgent need for a method that can perform safety monitoring on the woodworking machine tool and can quickly stop the machine in the face of emergency and dangerous situations. In addition, the method must have the characteristics of high monitoring accuracy and fast response speed. Summary of the invention

[0004] In view of the above-mentioned technical deficiencies, the purpose of the present invention is to provide a woodworking machine safety monitoring method, system, equipment and storage medium, aiming to solve the safety accidents that often occur to operators of existing woodworking machines. The cutting saw blade device commonly seen on woodworking machine tools causes damage to the operator's hands. There is an urgent need for a method that can perform safety monitoring on woodworking machine tools and quickly stop the machine in the face of emergency and dangerous situations. In addition, there are still problems such as insufficient monitoring accuracy and insufficient response speed.

[0005] In view of the above problems, the present application provides a woodworking machine tool safety monitoring method, system, device and storage medium.

[0006] The first aspect disclosed in the present application provides a woodworking machine tool safety monitoring method, the method comprising the following steps: Step 1: obtaining a real-time video through a camera installed above a cutting saw blade of a woodworking machine tool to generate a first video stream; Step 2: Perform frame extraction processing on the first video stream to generate a second video stream; Step 3: Input the second video stream frame by frame into the trained YOLOv3 model, perform target detection on the operator's hands and the cutting saw blade, select the positions of the operator's hands and the cutting saw blade on each frame, and generate a first position detection frame and a second position detection frame respectively, wherein the first position detection frame is represented by (X1, Y1, W1, H1), and the second position detection frame is represented by (X2, Y2, W2, H2), X1 and X2 are the horizontal coordinates of the center points of the first position detection frame and the second position detection frame respectively, Y1 and Y2 are the vertical coordinates of the center points of the first position detection frame and the second position detection frame respectively, W1 and W2 are the widths of the first position detection frame and the second position detection frame respectively, and H1 and H2 are the heights of the first position detection frame and the second position detection frame respectively; Step 4: Monitor the distance between the first position detection frame and the second position detection frame in real time, determine whether there is a dangerous situation or an emergency dangerous situation based on the distance, and control the issuance of a dangerous operation warning or control the woodworking machine to stop urgently.

[0007] Preferably, the step 2 specifically includes the following steps: Step 2.1: Obtain a first video stream and determine a frame rate of the first video stream; Step 2.2: If the frame rate of the first video stream is less than or equal to 15fps, no operation is performed, and the first video stream is directly converted into the second video stream; Step 2.3: If the frame rate of the first video stream is greater than 15fps and less than or equal to 24fps, extract two frames out of every three frames, reassemble the extracted video frames and output them to generate a second video stream; Step 2.4: If the frame rate of the first video stream is greater than 24fps and less than or equal to 30fps, extract one frame out of every two frames, reassemble the extracted video frames and output them to generate a second video stream; Step 2.5: If the frame rate of the first video stream is greater than 30fps and less than or equal to 60fps, extract one frame out of every three frames, reassemble the extracted video frames and output them to generate a second video stream.

[0008] Preferably, the structure of the YOLOv3 model in step 3 is: Perform cropping, compression or supplementation on the input image so that the image dimension is 416×416×3 to generate the initial image; Perform convolution calculation on the initial image to generate a first feature map with a dimension of 416×416×32; Input the first feature map into the first residual network unit to generate a second feature map with a dimension of 208×208×64; Input the second feature map into the second residual network unit to generate a third feature map with a dimension of 104×104×128; Input the third feature map into the third residual network unit to generate a fourth feature map with a dimension of 52×52×256; Input the fourth feature map into the fourth residual network unit to generate a fifth feature map with a dimension of 26×26×512; Inputting the fifth feature map into the fifth residual network unit to generate a sixth feature map with a dimension of 13×13×1024, wherein the first to fifth residual network units all include three convolution calculations and residual connection addition calculations; Perform convolution calculation and upsampling calculation on the sixth feature map to generate a seventh feature map with a dimension of 26×26×256, and perform five convolution calculations on the sixth feature map after concatenating it with the fifth feature map to generate an eighth feature map with a dimension of 26×26×256; Perform convolution calculation and upsampling calculation on the eighth feature map to generate a ninth feature map with a dimension of 52×52×128, and perform five convolution calculations on the ninth feature map after concatenating it with the fourth feature map to generate a tenth feature map with a dimension of 52×52×128; After performing convolution calculation on the tenth feature map, perform another convolution calculation with a convolution kernel of 1×1 to generate a detection result with a dimension of 52×52×(3×(5+2)), where 3 in (3×(5+2)) represents 3 prior boxes, each of which is represented by 5+2 dimensions. The 5 dimensions are used to represent 5 parameters, namely the horizontal coordinate of the center point of the prior box, the vertical coordinate of the center point of the prior box, the width of the prior box, the height of the prior box, and the confidence level. 2 dimensions are used to represent 2 categories, (1,0) represents the operator's hand, and (0,1) represents the cutting saw blade; A decoding operation is performed on the detection result to generate a first position detection frame (X1, Y1, W1, H1) and a second position detection frame (X2, Y2, W2, H2) for respectively selecting the operator's hand and the cutting saw blade.

[0009] Preferably, the training step of the YOLOv3 model in step 3 includes: Collect 1500 images of the operator's hands and cutting saw blades, and crop the resolution of all images to 416×416 to generate an image set; All images in the image set are rotated 180° and flipped horizontally, and the number of images in the image set is doubled to generate an expanded image set; Use labelimg to annotate all images in the image set with the operator's hands and cutting saw blades to generate an image dataset. The image dataset is divided into training set and test set in a ratio of 8:2; Perform training and testing of the YOLOv3 model using the training set and test set.

[0010] Preferably, the step 4 specifically comprises the following steps: Step 4.1: When it is monitored that the distance between the first position detection frame and the second position detection frame is less than or equal to the first threshold and greater than the second threshold, it is determined that a dangerous situation currently exists, a dangerous situation signal is triggered, and a dangerous operation warning is issued, wherein the distance between the first position detection frame and the second position detection frame is, the first threshold is [(W1+W2) / 2+(H1+H2) / 2]×1.5, the second threshold is (W1+W2) / 2+(H1+H2) / 2, and the second threshold is less than the first threshold; Step 4.2: When it is detected that the distance between the first position detection frame and the second position detection frame is less than or equal to the second threshold, it is determined that an emergency danger situation exists, an emergency stop signal is triggered, and the woodworking machine tool is controlled to stop urgently.

[0011] The second aspect disclosed in the present application provides a woodworking machine tool safety monitoring system, the system is used for the above-mentioned woodworking machine tool safety monitoring method, the system comprises: A first video stream module, the first video stream module is used to obtain real-time video through a camera installed above a cutting saw blade of a woodworking machine tool to generate a first video stream; A second video stream module, wherein the second video stream module is used to perform frame extraction processing on the first video stream to generate a second video stream; A target detection module, wherein the target detection module is used to input the second video stream frame by frame into the trained YOLOv3 model, perform target detection on the operator's hand and the cutting saw blade, select the position of the operator's hand and the cutting saw blade on each frame, and generate a first position detection frame and a second position detection frame respectively, wherein the first position detection frame is represented by (X1, Y1, W1, H1), and the second position detection frame is represented by (X2, Y2, W2, H2), X1 and X2 are respectively the horizontal coordinates of the center points of the first position detection frame and the second position detection frame, Y1 and Y2 are respectively the vertical coordinates of the center points of the first position detection frame and the second position detection frame, W1 and W2 are respectively the widths of the first position detection frame and the second position detection frame, and H1 and H2 are respectively the heights of the first position detection frame and the second position detection frame; A safety monitoring module is used to monitor the distance between the first position detection frame and the second position detection frame in real time, determine whether there is a dangerous situation or an emergency dangerous situation based on the distance, and control the issuance of a dangerous operation warning or control the woodworking machine to stop urgently.

[0012] According to a third aspect of the present application, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented: Step 1: obtaining a real-time video through a camera installed above a cutting saw blade of a woodworking machine tool to generate a first video stream; Step 2: Perform frame extraction processing on the first video stream to generate a second video stream; Step 3: Input the second video stream frame by frame into the trained YOLOv3 model, perform target detection on the operator's hands and the cutting saw blade, select the positions of the operator's hands and the cutting saw blade on each frame, and generate a first position detection frame and a second position detection frame respectively, wherein the first position detection frame is represented by (X1, Y1, W1, H1), and the second position detection frame is represented by (X2, Y2, W2, H2), X1 and X2 are the horizontal coordinates of the center points of the first position detection frame and the second position detection frame respectively, Y1 and Y2 are the vertical coordinates of the center points of the first position detection frame and the second position detection frame respectively, W1 and W2 are the widths of the first position detection frame and the second position detection frame respectively, and H1 and H2 are the heights of the first position detection frame and the second position detection frame respectively; Step 4: Monitor the distance between the first position detection frame and the second position detection frame in real time, determine whether there is a dangerous situation or an emergency dangerous situation based on the distance, and control the issuance of a dangerous operation warning or control the woodworking machine to stop urgently.

[0013] According to a fourth aspect of the present application, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented: Step 1: obtaining a real-time video through a camera installed above a cutting saw blade of a woodworking machine tool to generate a first video stream; Step 2: Perform frame extraction processing on the first video stream to generate a second video stream; Step 3: Input the second video stream frame by frame into the trained YOLOv3 model, perform target detection on the operator's hands and the cutting saw blade, select the positions of the operator's hands and the cutting saw blade on each frame, and generate a first position detection frame and a second position detection frame respectively, wherein the first position detection frame is represented by (X1, Y1, W1, H1), and the second position detection frame is represented by (X2, Y2, W2, H2), X1 and X2 are the horizontal coordinates of the center points of the first position detection frame and the second position detection frame respectively, Y1 and Y2 are the vertical coordinates of the center points of the first position detection frame and the second position detection frame respectively, W1 and W2 are the widths of the first position detection frame and the second position detection frame respectively, and H1 and H2 are the heights of the first position detection frame and the second position detection frame respectively; Step 4: Monitor the distance between the first position detection frame and the second position detection frame in real time, determine whether there is a dangerous situation or an emergency dangerous situation based on the distance, and control the issuance of a dangerous operation warning or control the woodworking machine to stop urgently.

[0014] According to a fifth aspect of the present application, a computer program product is provided, including a computer program or instructions, wherein when the computer program or instructions are executed by a processor, the following steps are implemented: Step 1: obtaining a real-time video through a camera installed above a cutting saw blade of a woodworking machine tool to generate a first video stream; Step 2: Perform frame extraction processing on the first video stream to generate a second video stream; Step 3: Input the second video stream frame by frame into the trained YOLOv3 model, perform target detection on the operator's hands and the cutting saw blade, select the positions of the operator's hands and the cutting saw blade on each frame, and generate a first position detection frame and a second position detection frame respectively, wherein the first position detection frame is represented by (X1, Y1, W1, H1), and the second position detection frame is represented by (X2, Y2, W2, H2), X1 and X2 are the horizontal coordinates of the center points of the first position detection frame and the second position detection frame respectively, Y1 and Y2 are the vertical coordinates of the center points of the first position detection frame and the second position detection frame respectively, W1 and W2 are the widths of the first position detection frame and the second position detection frame respectively, and H1 and H2 are the heights of the first position detection frame and the second position detection frame respectively; Step 4: Monitor the distance between the first position detection frame and the second position detection frame in real time, determine whether there is a dangerous situation or an emergency dangerous situation based on the distance, and control the issuance of a dangerous operation warning or control the woodworking machine to stop urgently.

[0015] The beneficial effects of the present invention are: (1) A woodworking machine tool safety monitoring method, system, device and storage medium are provided, which can perform safety monitoring on a woodworking machine tool and quickly stop the machine tool in the face of an emergency or dangerous situation, thereby solving the problem that the cutting saw blade device on the existing woodworking machine tool often causes damage to the operator's hands.

[0016] (2) The YOLOv3 model is used to perform targeted model structure design, model training, and target detection, so that the monitoring accuracy of the present invention is high and scene-specific.

[0017] (3) Frame extraction is used on the video stream to shorten the monitoring delay and improve the response speed of emergency stop in emergency situations. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 The figure is a schematic diagram of the overall process of a woodworking machine tool safety monitoring method.

[0020] Figure 2 This is the structural diagram of the YOLOv3 model.

[0021] Figure 3 The figure is a schematic diagram of the overall structure of a woodworking machine tool safety monitoring system.

[0022] Description of reference numerals: initial image 1, first feature Figure 2 , the second feature Figure 3 , third feature map 4, fourth feature map 5, fifth feature map 6, sixth feature map 7, seventh feature map 8, eighth feature map 9, ninth feature map 10, tenth feature map 11, detection result 12. DETAILED DESCRIPTION

[0023] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0024] like Figure 1 As shown, the embodiment of the present application provides a woodworking machine tool safety monitoring method, the method comprising the following steps: Step 1: obtaining a real-time video through a camera installed above a cutting saw blade of a woodworking machine tool to generate a first video stream; Step 2: Perform frame extraction processing on the first video stream to generate a second video stream; Step 3: Input the second video stream frame by frame into the trained YOLOv3 model, perform target detection on the operator's hands and the cutting saw blade, select the positions of the operator's hands and the cutting saw blade on each frame, and generate a first position detection frame and a second position detection frame respectively, wherein the first position detection frame is represented by (X1, Y1, W1, H1), and the second position detection frame is represented by (X2, Y2, W2, H2), X1 and X2 are the horizontal coordinates of the center points of the first position detection frame and the second position detection frame respectively, Y1 and Y2 are the vertical coordinates of the center points of the first position detection frame and the second position detection frame respectively, W1 and W2 are the widths of the first position detection frame and the second position detection frame respectively, and H1 and H2 are the heights of the first position detection frame and the second position detection frame respectively; Step 4: Monitor the distance between the first position detection frame and the second position detection frame in real time, determine whether there is a dangerous situation or an emergency dangerous situation based on the distance, and control the issuance of a dangerous operation warning or control the woodworking machine to stop urgently.

[0025] Furthermore, the step 2 specifically includes the following steps: Step 2.1: Obtain a first video stream and determine a frame rate of the first video stream; Step 2.2: If the frame rate of the first video stream is less than or equal to 15fps, no operation is performed, and the first video stream is directly converted into the second video stream; Step 2.3: If the frame rate of the first video stream is greater than 15fps and less than or equal to 24fps, extract two frames out of every three frames, reassemble the extracted video frames and output them to generate a second video stream; Step 2.4: If the frame rate of the first video stream is greater than 24fps and less than or equal to 30fps, extract one frame out of every two frames, reassemble the extracted video frames and output them to generate a second video stream; Step 2.5: If the frame rate of the first video stream is greater than 30fps and less than or equal to 60fps, extract one frame out of every three frames, reassemble the extracted video frames and output them to generate a second video stream.

[0026] Furthermore, if Figure 2 As shown, the structure of the YOLOv3 model in step 3 is: The input image is cropped, compressed, or supplemented so that the image dimension is 416×416×3, generating an initial image 1; Perform convolution calculation on the initial image 1 to generate the first feature with a dimension of 416×416×32 Figure 2 ; The first feature Figure 2 Input the first residual network unit to generate the second feature with a dimension of 208×208×64 Figure 3 ; The second feature Figure 3 Input the second residual network unit to generate the third feature map 4 with a dimension of 104×104×128; Input the third feature map 4 into the third residual network unit to generate a fourth feature map 5 with a dimension of 52×52×256; Input the fourth feature map 5 into the fourth residual network unit to generate a fifth feature map 6 with a dimension of 26×26×512; Input the fifth feature map 6 into the fifth residual network unit to generate a sixth feature map 7 with a dimension of 13×13×1024, wherein the first to fifth residual network units all include three convolution calculations and residual connection addition calculations; Perform convolution calculation and upsampling calculation on the sixth feature map 7 to generate a seventh feature map 8 with a dimension of 26×26×256, and perform five convolution calculations on the sixth feature map 7 after concatenating it with the fifth feature map 6 to generate an eighth feature map 9 with a dimension of 26×26×256; Perform convolution calculation and upsampling calculation on the eighth feature map 9 to generate a ninth feature map 10 with a dimension of 52×52×128, and perform five convolution calculations on the eighth feature map 9 after concatenating it with the fourth feature map 5 to generate a tenth feature map 11 with a dimension of 52×52×128; After performing convolution calculation on the tenth feature map 11, a convolution calculation with a convolution kernel of 1×1 is performed again to generate a detection result 12 with a dimension of 52×52×(3×(5+2)), where 3 in (3×(5+2)) represents 3 priori boxes, each priori box is represented by 5+2 dimensions, and the 5 dimensions are used to represent 5 parameters, namely the horizontal coordinate of the center point of the priori box, the vertical coordinate of the center point of the priori box, the width of the priori box, the height of the priori box and the confidence level, and 2 dimensions are used to represent 2 categories, (1,0) represents the operator's hand, and (0,1) represents the cutting saw blade; A decoding operation is performed on the detection result to generate a first position detection frame (X1, Y1, W1, H1) and a second position detection frame (X2, Y2, W2, H2) for respectively selecting the operator's hand and the cutting saw blade.

[0027] Furthermore, the training steps of the YOLOv3 model in step 3 include: Collect 1500 images of the operator's hands and cutting saw blades, and crop the resolution of all images to 416×416 to generate an image set; All images in the image set are rotated 180° and flipped horizontally, and the number of images in the image set is doubled to generate an expanded image set; Use labelimg to annotate all images in the image set with the operator's hands and cutting saw blades to generate an image dataset. The image dataset is divided into training set and test set in a ratio of 8:2; Perform training and testing of the YOLOv3 model using the training set and test set.

[0028] Furthermore, the step 4 specifically includes the following steps: Step 4.1: When it is monitored that the distance between the first position detection frame and the second position detection frame is less than or equal to the first threshold and greater than the second threshold, it is determined that a dangerous situation currently exists, a dangerous situation signal is triggered, and a dangerous operation warning is issued, wherein the distance between the first position detection frame and the second position detection frame is, the first threshold is [(W1+W2) / 2+(H1+H2) / 2]×1.5, the second threshold is (W1+W2) / 2+(H1+H2) / 2, and the second threshold is less than the first threshold; Step 4.2: When it is detected that the distance between the first position detection frame and the second position detection frame is less than or equal to the second threshold, it is determined that an emergency danger situation exists, an emergency stop signal is triggered, and the woodworking machine tool is controlled to stop urgently.

[0029] In summary, the woodworking machine tool safety monitoring method provided by the embodiment of the present application has the following technical effects: (1) A woodworking machine safety monitoring method is provided, which can perform safety monitoring on a woodworking machine and quickly stop it in the face of an emergency or dangerous situation, thereby solving the problem that the cutting saw blade device on the existing woodworking machine often causes damage to the operator's hands.

[0030] (2) The YOLOv3 model is used to perform targeted model structure design, model training, and target detection, so that the monitoring accuracy of the present invention is high and scene-specific.

[0031] (3) Frame extraction is used on the video stream to shorten the monitoring delay and improve the response speed of emergency stop in emergency situations.

[0032] Based on the same inventive concept as the woodworking machine tool safety monitoring method in the aforementioned embodiment, Figure 3 As shown, the present application provides a woodworking machine tool safety monitoring system, the system comprising: A first video stream module, the first video stream module is used to obtain real-time video through a camera installed above a cutting saw blade of a woodworking machine tool to generate a first video stream; A second video stream module, wherein the second video stream module is used to perform frame extraction processing on the first video stream to generate a second video stream; A target detection module, wherein the target detection module is used to input the second video stream frame by frame into the trained YOLOv3 model, perform target detection on the operator's hand and the cutting saw blade, select the position of the operator's hand and the cutting saw blade on each frame, and generate a first position detection frame and a second position detection frame respectively, wherein the first position detection frame is represented by (X1, Y1, W1, H1), and the second position detection frame is represented by (X2, Y2, W2, H2), X1 and X2 are respectively the horizontal coordinates of the center points of the first position detection frame and the second position detection frame, Y1 and Y2 are respectively the vertical coordinates of the center points of the first position detection frame and the second position detection frame, W1 and W2 are respectively the widths of the first position detection frame and the second position detection frame, and H1 and H2 are respectively the heights of the first position detection frame and the second position detection frame; A safety monitoring module is used to monitor the distance between the first position detection frame and the second position detection frame in real time, determine whether there is a dangerous situation or an emergency dangerous situation based on the distance, and control the issuance of a dangerous operation warning or control the woodworking machine to stop urgently.

[0033] Through the above-mentioned detailed description of a woodworking machine tool safety monitoring method in this specification, technical personnel in this field can clearly understand a woodworking machine tool safety monitoring system in this embodiment. Since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.

[0034] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented: Step 1: obtaining a real-time video through a camera installed above a cutting saw blade of a woodworking machine tool to generate a first video stream; Step 2: Perform frame extraction processing on the first video stream to generate a second video stream; Step 3: Input the second video stream frame by frame into the trained YOLOv3 model, perform target detection on the operator's hands and the cutting saw blade, select the positions of the operator's hands and the cutting saw blade on each frame, and generate a first position detection frame and a second position detection frame respectively, wherein the first position detection frame is represented by (X1, Y1, W1, H1), and the second position detection frame is represented by (X2, Y2, W2, H2), X1 and X2 are the horizontal coordinates of the center points of the first position detection frame and the second position detection frame respectively, Y1 and Y2 are the vertical coordinates of the center points of the first position detection frame and the second position detection frame respectively, W1 and W2 are the widths of the first position detection frame and the second position detection frame respectively, and H1 and H2 are the heights of the first position detection frame and the second position detection frame respectively; Step 4: Monitor the distance between the first position detection frame and the second position detection frame in real time, determine whether there is a dangerous situation or an emergency dangerous situation based on the distance, and control the issuance of a dangerous operation warning or control the woodworking machine to stop urgently.

[0035] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented: Step 1: obtaining a real-time video through a camera installed above a cutting saw blade of a woodworking machine tool to generate a first video stream; Step 2: Perform frame extraction processing on the first video stream to generate a second video stream; Step 3: Input the second video stream frame by frame into the trained YOLOv3 model, perform target detection on the operator's hands and the cutting saw blade, select the positions of the operator's hands and the cutting saw blade on each frame, and generate a first position detection frame and a second position detection frame respectively, wherein the first position detection frame is represented by (X1, Y1, W1, H1), and the second position detection frame is represented by (X2, Y2, W2, H2), X1 and X2 are the horizontal coordinates of the center points of the first position detection frame and the second position detection frame respectively, Y1 and Y2 are the vertical coordinates of the center points of the first position detection frame and the second position detection frame respectively, W1 and W2 are the widths of the first position detection frame and the second position detection frame respectively, and H1 and H2 are the heights of the first position detection frame and the second position detection frame respectively; Step 4: Monitor the distance between the first position detection frame and the second position detection frame in real time, determine whether there is a dangerous situation or an emergency dangerous situation based on the distance, and control the issuance of a dangerous operation warning or control the woodworking machine to stop urgently.

[0036] In one embodiment, a computer program product is provided, including a computer program or instructions, which, when executed by a processor, implements the following steps: Step 1: obtaining a real-time video through a camera installed above a cutting saw blade of a woodworking machine tool to generate a first video stream; Step 2: Perform frame extraction processing on the first video stream to generate a second video stream; Step 3: Input the second video stream frame by frame into the trained YOLOv3 model, perform target detection on the operator's hands and the cutting saw blade, select the positions of the operator's hands and the cutting saw blade on each frame, and generate a first position detection frame and a second position detection frame respectively, wherein the first position detection frame is represented by (X1, Y1, W1, H1), and the second position detection frame is represented by (X2, Y2, W2, H2), X1 and X2 are the horizontal coordinates of the center points of the first position detection frame and the second position detection frame respectively, Y1 and Y2 are the vertical coordinates of the center points of the first position detection frame and the second position detection frame respectively, W1 and W2 are the widths of the first position detection frame and the second position detection frame respectively, and H1 and H2 are the heights of the first position detection frame and the second position detection frame respectively; Step 4: Monitor the distance between the first position detection frame and the second position detection frame in real time, determine whether there is a dangerous situation or an emergency dangerous situation based on the distance, and control the issuance of a dangerous operation warning or control the woodworking machine to stop urgently.

[0037] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0038] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those 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 application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A woodworking machine tool safety monitoring method, characterized in that: The method comprises the following steps: Step 1: obtaining a real-time video through a camera installed above a cutting saw blade of a woodworking machine tool to generate a first video stream; Step 2: Perform frame extraction processing on the first video stream to generate a second video stream; Step 3: Input the second video stream frame by frame into the trained YOLOv3 model, perform target detection on the operator's hands and the cutting saw blade, select the positions of the operator's hands and the cutting saw blade on each frame, and generate a first position detection frame and a second position detection frame respectively, wherein the first position detection frame is represented by (X1, Y1, W1, H1), and the second position detection frame is represented by (X2, Y2, W2, H2), X1 and X2 are the horizontal coordinates of the center points of the first position detection frame and the second position detection frame respectively, Y1 and Y2 are the vertical coordinates of the center points of the first position detection frame and the second position detection frame respectively, W1 and W2 are the widths of the first position detection frame and the second position detection frame respectively, and H1 and H2 are the heights of the first position detection frame and the second position detection frame respectively; Step 4: Monitor the distance between the first position detection frame and the second position detection frame in real time, determine whether there is a dangerous situation or an emergency dangerous situation based on the distance, and control the issuance of a dangerous operation warning or control the woodworking machine to stop urgently.

2. A woodworking machine tool safety monitoring method as claimed in claim 1, characterized in that: The step 2 specifically includes the following steps: Step 2.1: Obtain a first video stream and determine a frame rate of the first video stream; Step 2.2: If the frame rate of the first video stream is less than or equal to 15fps, no operation is performed, and the first video stream is directly converted into the second video stream; Step 2.3: If the frame rate of the first video stream is greater than 15fps and less than or equal to 24fps, extract two frames out of every three frames, reassemble the extracted video frames and output them to generate a second video stream; Step 2.4: If the frame rate of the first video stream is greater than 24fps and less than or equal to 30fps, extract one frame out of every two frames, reassemble the extracted video frames and output them to generate a second video stream; Step 2.5: If the frame rate of the first video stream is greater than 30fps and less than or equal to 60fps, extract one frame out of every three frames, reassemble the extracted video frames and output them to generate a second video stream.

3. A woodworking machine tool safety monitoring method as claimed in claim 1, characterized in that: The structure of the YOLOv3 model in step 3 is: Perform cropping, compression or supplementation on the input image so that the image dimension is 416×416×3 to generate the initial image; Perform convolution calculation on the initial image to generate a first feature map with a dimension of 416×416×32; Input the first feature map into the first residual network unit to generate a second feature map with a dimension of 208×208×64; Input the second feature map into the second residual network unit to generate a third feature map with a dimension of 104×104×128; Input the third feature map into the third residual network unit to generate a fourth feature map with a dimension of 52×52×256; Input the fourth feature map into the fourth residual network unit to generate a fifth feature map with a dimension of 26×26×512; Inputting the fifth feature map into the fifth residual network unit to generate a sixth feature map with a dimension of 13×13×1024, wherein the first to fifth residual network units all include three convolution calculations and residual connection addition calculations; Perform convolution calculation and upsampling calculation on the sixth feature map to generate a seventh feature map with a dimension of 26×26×256, and perform five convolution calculations on the sixth feature map after concatenating it with the fifth feature map to generate an eighth feature map with a dimension of 26×26×256; Perform convolution calculation and upsampling calculation on the eighth feature map to generate a ninth feature map with a dimension of 52×52×128, and perform five convolution calculations on the ninth feature map after concatenating it with the fourth feature map to generate a tenth feature map with a dimension of 52×52×128; After performing convolution calculation on the tenth feature map, perform another convolution calculation with a convolution kernel of 1×1 to generate a detection result with a dimension of 52×52×(3×(5+2)), where 3 in (3×(5+2)) represents 3 prior boxes, each of which is represented by 5+2 dimensions. The 5 dimensions are used to represent 5 parameters, namely the horizontal coordinate of the center point of the prior box, the vertical coordinate of the center point of the prior box, the width of the prior box, the height of the prior box, and the confidence level. 2 dimensions are used to represent 2 categories, (1,0) represents the operator's hand, and (0,1) represents the cutting saw blade; A decoding operation is performed on the detection result to generate a first position detection frame (X1, Y1, W1, H1) and a second position detection frame (X2, Y2, W2, H2) for respectively selecting the operator's hand and the cutting saw blade.

4. A woodworking machine tool safety monitoring method as claimed in claim 1, characterized in that: The training steps of the YOLOv3 model in step 3 include: Collect 1500 images of the operator's hands and cutting saw blades, and crop the resolution of all images to 416×416 to generate an image set; All images in the image set are rotated 180° and flipped horizontally, and the number of images in the image set is doubled to generate an expanded image set; Use labelimg to annotate all images in the image set with the operator's hands and cutting saw blades to generate an image dataset. The image dataset is divided into training set and test set in a ratio of 8:2; Perform training and testing of the YOLOv3 model using the training set and test set.

5. A woodworking machine tool safety monitoring method as claimed in claim 1, characterized in that: The step 4 specifically comprises the following steps: Step 4.1: When it is detected that the distance between the first position detection frame and the second position detection frame is less than or equal to the first threshold and greater than the second threshold, it is determined that there is a dangerous situation, a dangerous situation signal is triggered, and a dangerous operation warning is issued, wherein the distance between the first position detection frame and the second position detection frame is , the first threshold is [(W1+W2) / 2+(H1+H2) / 2]×1.5, the second threshold is (W1+W2) / 2+(H1+H2) / 2, and the second threshold is less than the first threshold; Step 4.2: When it is detected that the distance between the first position detection frame and the second position detection frame is less than or equal to the second threshold, it is determined that an emergency danger situation exists, an emergency stop signal is triggered, and the woodworking machine tool is controlled to stop urgently.

6. A woodworking machine tool safety monitoring system, the system comprising: A first video stream module, the first video stream module is used to obtain real-time video through a camera installed above a cutting saw blade of a woodworking machine tool to generate a first video stream; A second video stream module, wherein the second video stream module is used to perform frame extraction processing on the first video stream to generate a second video stream; A target detection module, wherein the target detection module is used to input the second video stream frame by frame into the trained YOLOv3 model, perform target detection on the operator's hand and the cutting saw blade, select the position of the operator's hand and the cutting saw blade on each frame, and generate a first position detection frame and a second position detection frame respectively, wherein the first position detection frame is represented by (X1, Y1, W1, H1), and the second position detection frame is represented by (X2, Y2, W2, H2), X1 and X2 are respectively the horizontal coordinates of the center points of the first position detection frame and the second position detection frame, Y1 and Y2 are respectively the vertical coordinates of the center points of the first position detection frame and the second position detection frame, W1 and W2 are respectively the widths of the first position detection frame and the second position detection frame, and H1 and H2 are respectively the heights of the first position detection frame and the second position detection frame; A safety monitoring module is used to monitor the distance between the first position detection frame and the second position detection frame in real time, determine whether there is a dangerous situation or an emergency dangerous situation based on the distance, and control the issuance of a dangerous operation warning or control the woodworking machine to stop urgently.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of a woodworking machine tool safety monitoring method described in any one of claims 1 to 5 are implemented.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of a woodworking machine tool safety monitoring method described in any one of claims 1 to 5 are implemented.

9. A computer program product comprising a computer program or instructions, characterized in that When the computer program or instruction is executed by a processor, the steps of a woodworking machine tool safety monitoring method described in any one of claims 1 to 5 are implemented.

Images