Gantry machine tool magazine safety detection system

Abstract

The invention discloses a safety detection system for a tool magazine of a gantry machine tool, relates to the technical field of gantry machine tools, and solves the problems that an object entering a blind area cannot be quickly recognized to belong to a human body, and the comprehensiveness of recognition needs to be improved. By means of two-dimensional coordinate system analysis, central point overlapping and zooming operation, calculating an overlapping area proportion so as to judge whether the area is a human body area or not; the scientific judgment mode improves the accuracy and reliability of recognition, reduces the false alarm rate, ensures the accurate transmission of early warning information, and can quickly send out an early warning signal through the early warning end to remind an operator to take measures in time to avoid danger once the system judges that a change area belongs to a human body area; the real-time response mechanism can effectively reduce the probability of accidents and kill potential risks in the bud.

Description

Technical Field

[0001] The invention relates to the technical field of gantry machine tools, in particular to a tool magazine safety detection system for gantry machine tools. Background Art

[0002] Gantry machine tools are a type of large, high-precision machine tools that play a key role in the field of mechanical processing. They are named after their gantry-like frame structure. Their iconic gantry frame is composed of beams, columns and other components. The structure is stable and can withstand large cutting forces, ensuring accuracy when processing large workpieces. For example, when processing large ship engine cylinders, gantry machine tools can complete complex processing with their stable structure. They are large in size and have a strong load-bearing capacity, so large workpieces can be placed stably. Some workbenches can be linked by multiple axes to achieve complex movements and meet diverse processing needs.

[0003] The tool magazine of the gantry machine tool is huge. When performing maintenance and tool management, relevant personnel have to enter the work area. Due to the large size of the tool magazine, there are blind spots in the operation process. In order to avoid the operational risks caused by this, safety identification is required to promote safe production in the enterprise.

[0004] During the safety monitoring process of the gantry machine tool magazine, because no image analysis is performed on the relevant objects entering the blind spot, it is unable to quickly and effectively determine whether the objects entering the blind spot are related human bodies. Often, when an object enters, an early warning is processed, but this early warning method cannot quickly identify that the objects entering the blind spot are human bodies, and the comprehensiveness of its recognition needs to be improved. Summary of the invention

[0005] In view of the deficiencies in the prior art, the present invention provides a gantry machine tool tool magazine safety detection system, which solves the problem that an object entering a blind spot cannot be quickly identified as a human body and the comprehensiveness of the identification needs to be improved.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: a gantry machine tool magazine safety detection system, comprising:

[0007] The visual monitoring end monitors the blind area of ​​the gantry machine tool and transmits the real-time blind area video to the real-time monitoring video processing end;

[0008] The real-time processing end of the monitoring video processes the blind area video in real time. Based on the comparison difference between adjacent frames in the blind area video, the non-overlapping area is locked. Then, based on the change of the non-overlapping area, the changed area and the frames to be processed associated with the changed area are locked. The specific method is as follows:

[0009] Based on the blind spot video monitored in real time, the adjacent frames in the blind spot video are compared to identify whether the previous and next sets of frames are consistent. If they are consistent, the comparison is continued. If they are not consistent, the next set of frames is marked as the first different frames, and the inconsistent non-overlapping areas are marked;

[0010] Based on the marked initial different frame and the internal marked non-overlapping area, identify whether the next set of frames and the initial different frame are consistent with each other. If they are consistent, the non-overlapping area marked this time is recorded as the changed area. If they are not consistent, continue to confirm the non-overlapping area, and re-record the entire area of ​​the non-overlapping area confirmed this time and the marked non-overlapping area as the non-overlapping area;

[0011] The process is repeated in this way until the subsequent frame is consistent with the previous frame, the non-overlapping area is locked, and the locked non-overlapping area is marked as a changed area, and the frame to which the changed area belongs is recorded as a frame to be processed;

[0012] The change area confirmation end confirms the pixel values ​​associated with the pixel points inside the change area based on the marked frame to be processed and the marked change area inside, and locks the gradient pixel points based on the confirmed pixel values, confirms the area contour of the change area based on the continuously appearing gradient pixel points, and transmits the confirmed area contour to the human body confirmation processing end, specifically in the following manner:

[0013] Based on the confirmed frame to be processed and the associated change area, the pixel values ​​associated with different pixel points in the change area are calibrated as X i , where i represents different pixels;

[0014] Sobel is used to confirm the horizontal gradient and vertical gradient associated with the specified pixel point, and the confirmed horizontal gradient is calibrated as H i , the confirmed vertical gradient is calibrated as S i ;

[0015] use: Determine the comprehensive gradient ZH associated with the corresponding pixel point i ;

[0016] The confirmed comprehensive gradient ZH i Check with the preset value Y1, where the specific value of Y1 is determined by the operator based on experience. i ≥Y1, then mark this pixel as a gradient pixel. If ZH i <Y1, no calibration is performed on this pixel;

[0017] Confirm adjacent gradient pixel points in sequence, lock the area contours associated with several groups of adjacent gradient pixel points, mark the locked area contours as the area contours of the change area, and transmit the confirmed area contours to the human body confirmation processing end;

[0018] The human body confirmation processing end checks the area contour confirmed by the corresponding change area with the preset human body template to assess whether such change area belongs to the human body area. If it does, it directly controls the warning end to perform warning processing. If it does not, it continues to monitor. The specific method is as follows:

[0019] Based on the confirmed regional contour, the regional contour is placed in a two-dimensional coordinate system, and the associated two-dimensional coordinates are confirmed based on the positions of different contour points in the two-dimensional coordinate system, and then the associated groups of two-dimensional coordinates are averaged to confirm the average point, and the confirmed average point is marked as the center point of the regional contour;

[0020] Based on the preset human body template, confirm the preset center point in the human body template, move the human body template and the regional outline, and stop when the center point of the human body template coincides with the center point of the regional outline. Then, scale the human body template according to the determined center point. During the scaling process, keep the regional outline in the human body template, and determine the maximum overlap between the regional outline and the human body template. During the scaling process, record the overlap area between the regional outline and the human body template, and determine the proportion of the overlap area in the human body template, which = the total area of ​​the overlap area / the total area of ​​the human body template;

[0021] The confirmed proportion value is checked against the preset value Y2: if the proportion value is <Y2, it means that the contour of this area does not belong to the human body contour, and continuous monitoring is performed; if the proportion value is ≥Y2, it means that the contour of this area belongs to the human body contour, and an early warning signal is generated through the early warning terminal for display to alert relevant external operators.

[0022] Preferably, the blind area of ​​vision is set in advance by relevant personnel, and safety monitoring is performed on relevant areas of the gantry machine tool to prevent relevant personnel from entering the dangerous area of ​​the gantry machine tool.

[0023] The present invention provides a gantry machine tool magazine safety detection system. Compared with the prior art, it has the following beneficial effects:

[0024] The present invention can accurately lock the changed area and its corresponding frames to be processed by carefully comparing the adjacent frames in the blind area video, and then determine the position and state of the person after entering the blind area; in the change area confirmation link, the gradient pixel points are locked based on pixel value analysis, and the contour of the change area is accurately outlined, providing accurate basic data for subsequent human body confirmation;

[0025] The outline of the changed area is compared with the preset human body template, and the proportion of the overlapping area is calculated with the help of two-dimensional coordinate system analysis, center point overlap and scaling operations to determine whether it is a human body area. This scientific judgment method improves the accuracy and reliability of recognition, reduces the false alarm rate, and ensures the accurate transmission of warning information. Once the system determines that the changed area belongs to the human body area, it can quickly send a warning signal through the warning terminal to remind the operator to take timely measures to avoid danger. This real-time response mechanism can effectively reduce the probability of accidents and nip potential risks in the bud. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 It is a schematic diagram of the principle framework of the present invention. DETAILED DESCRIPTION

[0027] 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.

[0028] See also Figure 1 , the present application provides a gantry machine tool magazine safety detection system, including a visual monitoring end, a monitoring video real-time processing end, a change area confirmation end, a human body confirmation processing end and an early warning end, wherein the visual monitoring end, the monitoring video real-time processing end, the change area confirmation end, the human body confirmation processing end and the early warning end are all electrically connected from the output node to the input node in sequence;

[0029] Among them, the visual monitoring end monitors the blind area of ​​the gantry machine tool and transmits the real-time monitored blind area video to the real-time monitoring video processing end. The blind area of ​​the gantry machine tool is set in advance by relevant personnel, and the relevant areas of the gantry machine tool are safely monitored to prevent relevant personnel from entering the dangerous area of ​​the gantry machine tool;

[0030] Among them, the monitoring video real-time processing end processes the blind spot video monitored in real time in real time, locks the non-overlapping area based on the comparison difference between adjacent frames in the blind spot video, and then locks the changed area and the frames to be processed associated with the changed area based on the change of the non-overlapping area. Specifically, it is planned that a person walks into this blind spot of vision. In the process of gradual entry, there are several different frames of frames, and there is a gradually increasing non-overlapping area in each frame. When the corresponding person completely enters this blind spot of vision, the area associated with the gradually increasing non-overlapping area will not change at all, and then the confirmed non-overlapping area belongs to the confirmed changed area;

[0031] The specific method of locking the changed area is as follows:

[0032] Based on the blind spot video monitored in real time, the adjacent frames in the blind spot video are compared to identify whether the previous and next sets of frames are consistent. If they are consistent, the comparison is continued. If they are not consistent, the next set of frames is marked as the first different frames, and the inconsistent non-overlapping areas are marked;

[0033] Based on the marked initial different frame and the internal marked non-overlapping area, identify whether the next set of frames and the initial different frame are consistent with each other. If they are consistent, the non-overlapping area marked this time is recorded as the changed area. If they are not consistent, continue to confirm the non-overlapping area, and re-record the entire area of ​​the non-overlapping area confirmed this time and the marked non-overlapping area as the non-overlapping area;

[0034] The process is repeated in this way until the subsequent frame is consistent with the previous frame, the non-overlapping area is locked, and the locked non-overlapping area is marked as a changed area, and the frame to which the changed area belongs is recorded as a frame to be processed;

[0035] Specifically, the changed area confirmed here has been calibrated in the corresponding frame. The confirmed changed area is the state of the area associated with the corresponding person completely entering the blind spot, and the non-overlapping area about such person is locked. Based on the determined non-overlapping area, the associated coverage area of ​​the specified human body can be fully confirmed, and then the overall outline of the changed area is confirmed to ensure the specific accuracy of the confirmation of this frame and improve the specific confirmation effect of the frame.

[0036] Among them, the change area confirmation end confirms the pixel values ​​associated with the pixel points inside the change area based on the marked frame to be processed and the marked change area inside, and locks the gradient pixel points based on the confirmed pixel values, confirms the regional contour of this change area based on the continuously appearing gradient pixel points, and transmits the confirmed regional contour to the human body confirmation processing end. Specifically, there are several different pixel points in the change area, and different pixel points have different pixel values. At the edge contour of the corresponding area, there are points where the pixel value changes more obviously. Such points belong to gradient pixel points. Based on the continuously appearing gradient pixel points, the regional contour of the corresponding change area can be confirmed;

[0037] Based on the confirmed frame to be processed and the associated change area, the pixel values ​​associated with different pixel points in the change area are calibrated as X i , where i represents different pixels;

[0038] Sobel is used to confirm the horizontal gradient and vertical gradient associated with the specified pixel point, and the confirmed horizontal gradient is calibrated as H i, the confirmed vertical gradient is calibrated as S i Specifically, when confirming the horizontal gradient and the vertical gradient, it is necessary to confirm the eight groups of pixels adjacent to the pixel point, and assign weights associated with the pixel values ​​of each different pixel point, and then lock the horizontal gradient and the vertical gradient associated with the corresponding pixel point based on the assigned weights. Since this type of processing method is relatively common in the prior art, it will not be described in detail here;

[0039] use: Determine the comprehensive gradient ZH associated with the corresponding pixel point i ;

[0040] The confirmed comprehensive gradient ZH i Check with the preset value Y1, where the specific value of Y1 is determined by the operator based on experience. i <Y1, no calibration is performed on this pixel. If ZH i ≥Y1, then the pixel is marked as a gradient pixel;

[0041] Adjacent gradient pixel points are confirmed in sequence, and the regional contours associated with several groups of adjacent gradient pixel points are locked (a line is composed of points, so the edge contour can be determined based on several adjacent points), the locked regional contour is marked as the regional contour of this change area, and the confirmed regional contour is transmitted to the human body confirmation processing end.

[0042] Among them, the human body confirmation processing end checks the regional contour with the preset human body template based on the regional contour confirmed by the corresponding change area, and evaluates whether such change area belongs to the human body area. If it does, it directly controls the warning end to perform warning processing. If it does not, it continues to monitor. The specific way of evaluation is:

[0043] Based on the confirmed regional contour, the regional contour is placed in a two-dimensional coordinate system, and the associated two-dimensional coordinates are confirmed based on the positions of different contour points in the two-dimensional coordinate system, and then the associated groups of two-dimensional coordinates are averaged to confirm the average point, and the confirmed average point is marked as the center point of the regional contour;

[0044] Based on the preset human body template, confirm the preset center point in the human body template, move the human body template and the regional outline, and stop when the center point of the human body template coincides with the center point of the regional outline. Then, scale the human body template according to the determined center point. During the scaling process, keep the regional outline in the human body template, and determine the maximum overlap between the regional outline and the human body template. During the scaling process, record the overlap area between the regional outline and the human body template, and determine the proportion of the overlap area in the human body template, which = the total area of ​​the overlap area / the total area of ​​the human body template;

[0045] The confirmed proportion value is checked against the preset value Y2: if the proportion value is less than Y2, it means that the contour of this area does not belong to the human body contour, and continuous monitoring is performed; if the proportion value is greater than or equal to Y2, it means that the contour of this area belongs to the human body contour, and an early warning signal is generated and displayed through the early warning terminal to alert external relevant operators;

[0046] Specifically, the system can monitor the blind spots of the gantry machine tool in real time, set dangerous areas in advance to prevent people from entering, reduce safety accidents caused by people mistakenly entering blind spots, ensure the personal safety of operators, and create a safe and reliable working atmosphere for the production environment;

[0047] By carefully comparing adjacent frames in the blind spot video, the changed area and its corresponding frames to be processed can be accurately locked, and the position and status of the person after entering the blind spot can be determined. In the change area confirmation stage, the gradient pixel points are locked based on pixel value analysis, and the contour of the changed area is accurately outlined, providing accurate basic data for subsequent human body confirmation.

[0048] The human body confirmation processing end compares the contour of the changed area with the preset human body template, and uses two-dimensional coordinate system analysis, center point overlap and zoom operations to calculate the proportion of the overlap area to determine whether it is a human body area. This scientific judgment method improves the accuracy and reliability of recognition, reduces the false alarm rate, and ensures the accurate transmission of warning information;

[0049] Once the system determines that the changed area belongs to the human body area, it can quickly send out a warning signal through the early warning terminal to remind the operator to take timely measures to avoid danger. This real-time response mechanism can effectively reduce the probability of accidents and nip potential risks in the bud.

[0050] Some of the data in the above formulas are dimensionless and numerically calculated. Meanwhile, the contents not described in detail in this specification belong to the prior art known to those skilled in the art.

[0051] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A gantry machine tool magazine safety detection system, characterized in that: include: The visual monitoring end monitors the blind area of ​​the gantry machine tool and transmits the real-time blind area video to the real-time monitoring video processing end; The monitoring video real-time processing end processes the blind spot video in real time, and locks the non-overlapping area based on the comparison difference between adjacent frames in the blind spot video. Then, based on the change of the non-overlapping area, locks the changed area and the frames to be processed associated with the changed area; The change region confirmation end confirms the pixel values ​​associated with the pixel points inside the change region based on the marked frame to be processed and the marked change region inside, locks the gradient pixel points based on the confirmed pixel values, confirms the region outline of the change region based on the continuously appearing gradient pixel points, and transmits the confirmed region outline to the human body confirmation processing end; The human body confirmation processing end checks the area contour confirmed by the corresponding change area with the preset human body template to assess whether such change area belongs to the human body area. If so, it directly controls the early warning end to perform early warning processing; if not, it continues to monitor.

2. A gantry machine tool magazine safety detection system according to claim 1, characterized in that: The specific method of locking the changed area at the monitoring video real-time processing end is: Based on the blind spot video monitored in real time, the adjacent frames in the blind spot video are compared to identify whether the previous and next sets of frames are consistent. If they are consistent, the comparison is continued. If they are not consistent, the next set of frames is marked as the first different frames, and the inconsistent non-overlapping areas are marked; Based on the marked initial different frame and the internal marked non-overlapping area, identify whether the next set of frames and the initial different frame are consistent with each other. If they are consistent, the non-overlapping area marked this time is recorded as the changed area. If they are not consistent, continue to confirm the non-overlapping area, and re-record the entire area of ​​the non-overlapping area confirmed this time and the marked non-overlapping area as the non-overlapping area; The process is repeated in this way until the subsequent frames are consistent with the previous frame group, the non-overlapping area is locked, and the locked non-overlapping area is marked as a changed area, and the frame to which the changed area belongs is recorded as a frame to be processed.

3. A gantry machine tool magazine safety detection system according to claim 1, characterized in that: The specific method of the change area confirmation terminal confirming the area outline of the change area is: Based on the confirmed frame to be processed and the associated change area, the pixel values ​​associated with different pixel points in the change area are calibrated as X i , where i represents different pixels; Sobel is used to confirm the horizontal gradient and vertical gradient associated with the specified pixel point, and the confirmed horizontal gradient is calibrated as H i , the confirmed vertical gradient is calibrated as S i ; use: Determine the comprehensive gradient ZH associated with the corresponding pixel point i ; The confirmed comprehensive gradient ZH i Check with the preset value Y1, where the specific value of Y1 is determined by the operator based on experience. i ≥Y1, then the pixel is marked as a gradient pixel; Adjacent gradient pixel points are confirmed in sequence, the regional contours associated with several groups of adjacent gradient pixel points are locked, the locked regional contours are marked as the regional contours of the change area, and the confirmed regional contours are transmitted to the human body confirmation processing end.

4. A gantry machine tool magazine safety detection system according to claim 3, characterized in that: If ZH i <Y1, no calibration is performed on this pixel.

5. A gantry machine tool tool magazine safety detection system according to claim 1, characterized in that: The specific method of the human body confirmation processing end to assess whether such a changed area belongs to the human body area is: Based on the confirmed regional contour, the regional contour is placed in a two-dimensional coordinate system, and the associated two-dimensional coordinates are confirmed based on the positions of different contour points in the two-dimensional coordinate system, and then the associated groups of two-dimensional coordinates are averaged to confirm the average point, and the confirmed average point is marked as the center point of the regional contour; Based on the preset human body template, confirm the preset center point in the human body template, move the human body template and the regional outline, and stop when the center point of the human body template coincides with the center point of the regional outline. Then, scale the human body template according to the determined center point. During the scaling process, keep the regional outline in the human body template, and determine the maximum overlap between the regional outline and the human body template. During the scaling process, record the overlap area between the regional outline and the human body template, and determine the proportion of the overlap area in the human body template, which = the total area of ​​the overlap area / the total area of ​​the human body template; The confirmed proportion value is compared with the preset value Y2: if the proportion value is less than Y2, it means that the contour of this area does not belong to the human body contour and continues to be monitored.

6. A gantry machine tool tool magazine safety detection system according to claim 5, characterized in that: If the proportion value is ≥Y2, it means that the contour of this area belongs to the human body contour, and an early warning signal is generated through the early warning terminal for display to alert relevant external operators.

7. A gantry machine tool magazine safety detection system according to claim 1, characterized in that: The blind area of ​​vision is set in advance by relevant personnel to conduct safety monitoring on the relevant areas of the gantry machine tool to prevent relevant personnel from entering the dangerous area of ​​the gantry machine tool.

Images