Mineral water bottle processing and conveying device

Abstract

The invention relates to the technical field of conveying devices, and particularly discloses a mineral water bottle processing and conveying device which comprises a belt conveyor and a control cooperation system, the belt conveyor comprises a rack and a driving motor fixedly connected to the outer portion of the rack, and rotating rollers are rotationally connected to the two ends of the rack; the output end of the driving motor is fixedly connected with a rotating roller, the rotating roller is sleeved with a conveying belt, the outer portion of the rack is fixedly connected with a mounting frame, the mounting frame is fixedly connected with a visual camera, and a tension adjusting mechanism is further arranged below the rack. According to the invention, through the arrangement of the control cooperation system, in the aspect of positioning precision, multi-dimensional detection of the size, the posture and the position of the bottle body is realized by means of the visual camera, and the positioning accuracy is greatly improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of conveying devices, and in particular to a mineral water bottle processing conveying device. BACKGROUND

[0002] In the field of mineral water production, recycling and regeneration, the belt conveyor is the core equipment for realizing the continuous transfer, sorting and processing of mineral water bottles, and its running stability and processing precision directly affect the production efficiency and product quality.

[0003] The existing mineral water bottle processing belt conveyor relies on infrared transmission sensors or photoelectric switches to detect the position of the bottle, and can only determine whether the bottle exists or not, and cannot obtain detailed information such as the posture and size specification of the bottle, resulting in positioning deviation in subsequent sorting, labeling and other processes. SUMMARY

[0004] In view of the technical problems in the background art, the present application provides a mineral water bottle processing conveying device.

[0005] The technical scheme adopted by the present application is: a mineral water bottle processing conveying device, comprising a belt conveyor and a control coordination system, the belt conveyor comprising a rack and a driving motor fixedly connected to the outside of the rack, both ends of the rack being rotatably connected with a rotating roller, the output end of the driving motor being fixedly connected with the rotating roller, the outside of the rotating roller being sleeved with a conveying belt, the outside of the rack being fixedly connected with a mounting bracket, the mounting bracket being fixedly connected with a visual camera, the lower part of the rack being further provided with a tension adjusting mechanism, the control coordination system comprising a visual information acquisition and preprocessing module, a mineral water bottle feature extraction and positioning module, a dynamic speed adjusting module, a mineral water bottle quality detection module, a tension self-adaptive compensation module, a data encryption and remote monitoring module, a fault self-diagnosis and prediction module and an energy consumption optimization management module.

[0006] Further provided in the present application is that the tension adjusting mechanism comprises a mounting seat and an adjusting roller rotatably connected to the mounting seat, the outside of the rack being fixedly connected with a servo motor, the output end of the servo motor being fixedly connected with a lead screw, the outside of the lead screw being threadedly connected with a nut, the nut being fixedly connected with the mounting seat, the rack being further fixedly connected with a guide rod, the guide rod penetrating through the mounting seat.

[0007] Further provided in the present application is that the visual information acquisition and preprocessing module is based on the visual camera to acquire images, and automatically adjusts the exposure time and gain of the visual camera, and the formula is as follows:

[0008] Exposure time calculation:

[0009] ;

[0010] wherein: Camera exposure time Camera ISO Camera target output voltage Real-time ambient light intensity Camera photosensitive area

[0011] Gain adjustment formula

[0012] ;

[0013] Where: Camera actual gain Base gain Standard ambient light intensity Real-time ambient light intensity, consistent with exposure time calculation

[0014] Image preprocessing

[0015] The original image collected needs to go through three steps of preprocessing: noise reduction, distortion correction, and region of interest extraction:

[0016] Gaussian noise reduction: adaptive Gaussian filtering is used to eliminate noise caused by dust and electromagnetic interference. The filter kernel size is dynamically adjusted according to the image noise intensity, and the formula is as follows:

[0017] ;

[0018] Where: Gaussian filter standard deviation Image average gray value

[0019] Distortion correction: due to the radial distortion of the camera lens, it needs to be corrected by a polynomial correction model, and the formula is as follows:

[0020] ;

[0021] ;

[0022] Where: Pixel coordinates before correction Pixel coordinates after correction Distortion coefficient Distance of pixel to image center, , Image center coordinates

[0023] ROI extraction: only the image within the conveyor belt area is retained, and the ROI area is determined by conveyor belt edge detection, and the formula is as follows:

[0024] ;​

[0025] ;

[0026] wherein: : the gray value of a pixel ; : the upper boundary of the ROI region : the lower boundary of the ROI region

[0027] The application further provides that the mineral water bottle feature extraction and positioning module extracts multi-dimensional features of the mineral water bottle based on the pre-processed visual image, and realizes positioning of the mineral water bottle in a physical space through coordinate mapping.

[0028] Mineral water bottle size feature extraction

[0029] Through edge detection and contour fitting, the height H, diameter D and bottle mouth radius r of the mineral water bottle are calculated, and the formula is as follows:

[0030] Mineral water bottle height calculation

[0031] ;

[0032] wherein: H: the actual height of the mineral water bottle : the y-coordinate of the top of the mineral water bottle in the image : the y-coordinate of the bottom of the mineral water bottle in the image, obtained through Canny edge detection : the actual length of the visual camera field of view in the direction perpendicular to the conveying belt : the vertical resolution of the image

[0033] Mineral water bottle diameter calculation

[0034] ;

[0035] wherein: D: the maximum diameter of the mineral water bottle : the x-coordinate of the left side of the mineral water bottle in the image : the x-coordinate of the right side of the mineral water bottle in the image : the actual width of the visual camera field of view in the direction parallel to the conveying belt : the horizontal resolution of the image

[0036] Mineral water bottle posture and positioning

[0037] Through Hough transform to detect the central axis of the mineral water bottle, the rotation angle and center coordinates of the mineral water bottle are calculated to realize physical space positioning

[0038] Rotation angle calculation

[0039] ;

[0040] Wherein: : rotation angle of mineral water bottle around the axis perpendicular to the conveying belt;

[0041] Mapping of mineral water bottle center coordinates:

[0042] ;

[0043] ;

[0044] Wherein: : coordinates of mineral water bottle center in the physical coordinate system of conveying belt, is parallel to the direction of conveying belt, is perpendicular to the direction of conveying belt; : coordinates of mineral water bottle center in the image, , ; : coordinates of image center; , : horizontal and vertical resolution of image;

[0045] Mineral water bottle mark recognition;

[0046] The mark on the label of mineral water bottle is recognized by template matching algorithm, and the matching similarity formula is as follows:

[0047] ;

[0048] Wherein: : matching similarity, the value range is 0-1, then it is determined that the matching is successful; : gray value of ROI region of image to be detected; : gray value of template image; : region of interest where the mark is located, which is determined by label edge detection.

[0049] The further setting of the application is that the dynamic speed adjusting module dynamically adjusts the rotating speed of the driving motor, so as to avoid the stacking or idling of mineral water bottles.

[0050] Mineral water bottle spacing calculation first calculates the center spacing of two adjacent mineral water bottles, and the formula is as follows:

[0051] ;

[0052] Wherein: : center spacing between the th mineral water bottle and the th mineral water bottle; , : the th mineral water bottle and the The physical center coordinates of the mineral water bottle, parallel to the direction of the conveying belt;

[0053] Conveying belt target speed calculation;

[0054] Based on the mineral water bottle spacing, specifications, and subsequent process requirements, a weighted fusion algorithm is used to calculate the target speed, as follows:

[0055] ;

[0056] Where: : Conveying belt target speed; : Base speed; : Weighting coefficient; : Ideal mineral water bottle spacing; : Average diameter of the mineral water bottles on the current conveying belt, , Real-time mineral water bottle quantity; : Single processing time of the subsequent process;

[0057] Drive motor speed closed-loop control;

[0058] PID algorithm is used to achieve precise control of the drive motor speed, and the PID output formula is as follows:

[0059] ;

[0060] Where: : Drive motor control voltage; : Speed deviation, , Actual speed of the conveying belt; : Proportional coefficient; : Integral coefficient; : Differential coefficient;

[0061] Emergency deceleration control;

[0062] When the mineral water bottle posture is detected to be abnormal or the spacing is too small, emergency deceleration is triggered, and the deceleration acceleration formula is as follows:

[0063] ;

[0064] Where: : Deceleration acceleration, negative sign indicating deceleration; : Acceleration coefficient; : Absolute value of speed deviation, , Emergency speed.

[0065] The mineral water bottle quality detection module detects the appearance defects and content abnormalities of the mineral water bottle based on the gray scale, texture and shape features of a visual image.

[0066] Scratch detection: the gray scale gradient of the scratch area is significantly higher than that of the normal mineral water bottle, and the formula is as follows:

[0067] ;

[0068] If is the gradient threshold value, is taken as 30, and the number of continuous pixels is greater than or equal to 5, it is determined that there is a scratch.

[0069] wherein, : the gray scale gradient of a pixel; : the gray scale value of a pixel; : the gradient threshold value, which is determined by normal mineral water bottle sample statistics; Mineral water bottle deformation detection: the similarity of the contour of the mineral water bottle and the standard contour is compared, and the formula is as follows:

[0070]

[0071] ;

[0072] If , it is determined that there is deformation.

[0073] wherein: : the contour similarity; : the distance from the i th contour point of the detected mineral water bottle to the center; : the distance from the i th contour point of the standard mineral water bottle to the center; : the number of contour points, taken as 100; : the maximum contour distance of the standard mineral water bottle; Bottle opening damage detection: the bottle opening area is circular, and whether it is damaged is determined by circularity calculation, and the formula is as follows:

[0074]

[0075] ;

[0076] If , it is determined that the bottle opening is damaged; wherein: : the circularity, taken as 0-1; : the actual area of the bottle opening area; : the circumference of the bottle opening area;

[0077] Liquid level detection: the gray scale difference between the content of the mineral water bottle and the air is used to detect the liquid level height, and the formula is as follows:​​​​

[0078] ;

[0079] If ; wherein: : actual liquid level height; : y coordinate of liquid level line in image, obtained by gray level mutation detection; : y coordinate of mineral water bottle bottom;

[0080] Impurity detection: local gray level variance method is adopted, and the gray level variance of the impurity area is significantly higher than that of the normal water body, and the formula is as follows:

[0081] ;

[0082] If ; it is judged that there is impurity, is the variance threshold, and is 25;

[0083] wherein: : gray level variance of liquid level area; : pixel number of liquid level area; : pixel gray level value of liquid level area; : average gray level value of liquid level area; : region of interest where the liquid level is located.

[0084] The further setting of the application is that the tension self-adaptive compensation module step detects the conveying belt edge position through a visual camera, calculates the tension deviation, and controls the tension adjusting mechanism to realize self-adaptive compensation, and the specific steps are as follows:

[0085] The conveying belt edge position detection detects the x coordinate of the two sides of the conveying belt edge through a visual image, calculates the offset, and the formula is as follows:

[0086] ;

[0087] ;

[0088] wherein: : offset of the left side of the conveying belt; : offset of the right side of the conveying belt, positive for right offset, and negative for left offset; : physical x coordinate of actual edge, obtained by edge detection and coordinate mapping; : physical x coordinate of standard edge;

[0089] Tension deviation calculation;

[0090] ​​There is a linear relationship between the belt deviation and the tension deviation, and the tension deviation formula is as follows:

[0091] ;

[0092] Wherein: : Tension deviation, positive for right tension too large, negative for left tension too large; : Tension deviation coefficient; , : Left and right edge deviation of the conveying belt;

[0093] Tension compensation control;

[0094] The fuzzy PID algorithm is adopted to control the driving of the servo motor, so that the tension adjustment is realized, and the output formula of the fuzzy PID is as follows:

[0095] ;

[0096] Wherein: : Servo motor control voltage output by fuzzy PID; : Traditional PID output, , , , ; : Fuzzy correction coefficient, determined according to the fuzzy rule table of tension deviation And the change rate of deviation ( );

[0097] Temperature compensation environment;

[0098] Temperature change will cause the thermal expansion and contraction of the conveying belt material, which will affect the tension, so temperature compensation is needed, and the formula is as follows:

[0099] ;

[0100] Wherein: : Tension change caused by temperature; : Conveying belt temperature coefficient, rubber conveying belt ; : Real-time environmental temperature, collected by the built-in temperature sensor of the visual camera; : Standard temperature, taking 25 DEG C; : Effective length of the conveying belt.

[0101] The application further provides that the data encryption and remote monitoring module encrypts and transmits the visual image, control parameter and detection result data;

[0102] Image data compression and encryption: using JPEG2000 compression algorithm to reduce data, and then encrypted by AES-256 encryption algorithm, encryption key generation formula as follows:

[0103] ;

[0104] Where: : AES-256 encryption key; : device unique identification, solidified at the factory; : timestamp, accurate to seconds; : random number, generated by the device built-in random number generator; : SHA-256 hash function, which maps the input to a 256-bit key;

[0105] Control parameter encryption transmission: using RSA asymmetric encryption algorithm to encrypt speed, tension and other control parameters, public key is generated by cloud server and delivered, private key is stored in local device, encryption formula as follows:

[0106] ;

[0107] Where: : encrypted ciphertext; : control parameter to be encrypted; : public key exponent; : public key modulus;

[0108] Remote data transmission uses LoRa wireless communication technology to realize the communication between device and cloud, transmission power dynamic adjustment formula as follows: ;

[0109] Where: : LoRa module transmission power, value range-10 to 15dBm; : minimum transmission power, take-10dBm; : transmission distance; : system loss; : transmit and receive antenna gain; : LoRa signal wavelength; : receiving sensitivity;

[0110] Real-time monitoring: cloud platform real-time display of conveyor belt speed, mineral water bottle detection results, device status, data update frequency≤1s, support historical data query;

[0111] Remote parameter configuration: users modify control parameters through cloud platform, configuration instruction transmission delay formula as follows:

[0112] ;

[0113] wherein: : instruction transmission delay; : instruction data length; : transmission rate; : transmission distance; : speed of light;

[0114] If , trigger instruction retransmission, ensure successful configuration.

[0115] The application further provides that the fault self-diagnosis and prediction module realizes real-time diagnosis and progressive fault prediction of the equipment by analyzing visual images, driving motor currents and temperature data.

[0116] Driving motor fault diagnosis: judge whether the driving motor is locked or overloaded by driving motor current detection, the formula is as follows:

[0117] If , determine that the driving motor is locked.

[0118] If , determine that the driving motor is overloaded.

[0119] wherein: : actual current of the driving motor; : rated current of the driving motor; : actual speed of the conveying belt; : target speed of the conveying belt;

[0120] Conveying belt slip diagnosis: compare the theoretical speed corresponding to the driving motor speed with the actual speed detected by visual detection, the formula is as follows:

[0121] ;

[0122] If , determine that the conveying belt is slipping.

[0123] wherein: : slip rate; : theoretical speed corresponding to the driving motor speed, : driving motor speed;

[0124] Visual camera fault diagnosis: detect the image acquisition frame rate and definition, the formula is as follows:

[0125] If the frame rate or the average image gray value , underexposure, or , overexposure, determine that the visual camera is faulty.

[0126] The progressive fault prediction adopts an LSTM model to predict the aging of the driving motor and the progressive failure of the conveying belt, and the prediction formula is as follows:

[0127] ;

[0128] Wherein: : the failure probability in the future time, and the early warning is triggered; : a sigmoid activation function; : a weight matrix; : an LSTM hidden layer state; : current input data; : a bias term; : a prediction time window;

[0129] Local alarm: when the fault occurs, the buzzer and LED lamp are triggered, and the alarm duration formula is as follows:

[0130] ;

[0131] Wherein: : alarm duration; : basic duration; : a coefficient, taking 5s; : failure probability;

[0132] Remote alarm: send fault information to the cloud platform and recommend a processing scheme.

[0133] The beneficial effects of the present application are: in the present application, by controlling the setting of the cooperative system, in terms of positioning accuracy, the size, posture and position of the bottle are detected in multiple dimensions with the aid of a visual camera, greatly improving the positioning accuracy and meeting the needs of subsequent high-precision sorting and labeling processes. In terms of speed control, dynamic adjustment is made based on the distance and specifications of the bottles, which not only avoids the problem of bottle stacking congestion, but also reduces the waste caused by the idling of the conveying belt, while reducing the bottle breakage rate, achieving efficient and stable transfer. BRIEF DESCRIPTION OF DRAWINGS

[0134] Fig. 1 is a structural schematic diagram of the present application;

[0135] Fig. 2 is a structural schematic diagram of the rack in the present application.

[0136] In the figure, the marks are: 1, rack; 2, rotating roller; 3, driving motor; 4, mounting bracket; 5, visual camera; 6, servo motor; 7, mounting seat; 8, screw rod; 9, nut; 10, guide rod; 11, adjusting roller. DETAILED DESCRIPTION

[0137] In the description of the present application, it should be noted that the terms "front", "upper", "lower", "left", "right", "vertical", "horizontal" and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as limiting the present application.

[0138] The accompanying drawings will be described below. Figs. 1-2 The present application is further described.

[0139] In order to solve the problems in the background art, the present application proposes the following technical scheme: a mineral water bottle processing and conveying device, comprising a belt conveyor and a control coordination system, the belt conveyor comprising a rack 1 and a driving motor 3 fixedly connected outside the rack 1, both ends of the rack 1 are rotatably connected with a rotating roller 2, the output end of the driving motor 3 is fixedly connected with the rotating roller 2, the outside of the rotating roller 2 is sleeved with a conveying belt, the outside of the rack 1 is fixedly connected with a mounting bracket 4, the mounting bracket 4 is fixedly connected with a visual camera 5, and a tension adjusting mechanism is further arranged below the rack 1, the control coordination system comprises a visual information acquisition and preprocessing module, a mineral water bottle feature extraction and positioning module, a dynamic speed adjusting module, a mineral water bottle quality detection module, a tension self-adaptive compensation module, a data encryption and remote monitoring module, a fault self-diagnosis and prediction module and an energy consumption optimization management module;

[0140] The tension adjusting mechanism comprises a mounting seat 7 and an adjusting roller 11 rotatably connected to the mounting seat 7, a servo motor 6 is fixedly connected to the outside of the rack 1, the output end of the servo motor 6 is fixedly connected with a lead screw 8, the outside of the lead screw 8 is threadedly connected with a nut 9, the nut 9 is fixedly connected with the mounting seat 7, a guide rod 10 is further fixedly connected to the rack 1, and the guide rod 10 penetrates through the mounting seat 7.

[0141] The servo motor 6 drives the lead screw 8 to rotate, the mounting seat 7 is lifted, the adjusting roller 11 is driven to resist the conveying belt, and the tension of the conveying belt is adjusted.

[0142] The visual information acquisition and preprocessing module: this module uses an industrial-grade color visual camera 5 with a resolution of 2048x1536 and a frame rate of 30fps as the only sensor, is installed 1.2m above the conveying belt, covers a 1.5 times field of view range of 500mm width of the conveying belt, and realizes full-area dead-angle-free collection of the mineral water bottles on the conveying belt.

[0143] The core function of the module is to obtain high-quality mineral water bottle images, and to eliminate environmental interference through preprocessing to provide reliable data basis for subsequent feature extraction.

[0144] The visual camera 5 parameter dynamic configuration is adapted to different light environments such as workshop natural light changes and lamp light aging, and the exposure time and gain of the visual camera 5 are automatically adjusted based on real-time image brightness, and the formula is as follows:

[0145] Exposure time calculation:

[0146] ;

[0147] Wherein: : visual camera 5 exposure time unit: s, value range 0.001-0.01s; : visual camera 5 photosensitive coefficient determined by visual camera 5 model, Basler acA2040-180uc visual camera 5 is adopted in the embodiment ; : visual camera 5 target output voltage unit: V, taking 1.8V ensures that the image gray value is in the interval of 120-180, avoiding overexposure or underexposure; : real-time ambient light intensity unit: lx, collected by the built-in light intensity sensor of the visual camera 5, range 500-5000lx; : visual camera 5 photosensitive area unit: , the visual camera 5 photosensitive chip size is 1 / 1.8 inch, .

[0148] Effect: dynamically adjust the exposure time according to the ambient light intensity, avoid overexposure of mineral water bottle reflection under strong light and too large image noise under weak light, ensure that the edge, texture and other details of the mineral water bottle are clear and identifiable, and the image gray standard deviation is controlled within 15.

[0149] Gain adjustment formula:

[0150] ;

[0151] Wherein: : visual camera 5 actual gain unit: dB, value range 0-12dB; : basic gain unit: dB, taking 6dB as the optimal gain corresponding to the standard light intensity ; : standard ambient light intensity unit: lx, taking 2000lx; : real-time ambient light intensity unit: lx, consistent with the in the exposure time calculation.

[0152] Effect: appropriately increase the gain in low light environment to enhance the image signal intensity, while avoiding noise amplification caused by too high gain, and ensure that the image signal-to-noise ratio SNR≥30dB.

[0153] Image preprocessing;

[0154] The original image collected needs to be preprocessed by noise reduction, distortion correction, and ROI extraction:

[0155] Gaussian noise reduction: adaptive Gaussian filtering is used to eliminate noise caused by dust and electromagnetic interference. The filter kernel size is dynamically adjusted according to the image noise intensity, and the formula is as follows:

[0156] ;

[0157] Where: : Gaussian filter standard deviation, value range 0.5-2.0; : image average gray value, range 0-255, the lower the gray value, the more significant the noise impact, the greater.

[0158] Effect: On the premise of preserving the edge details of the mineral water bottle, reduce noise interference, and the image noise pixel ratio is reduced from 5%-8% before preprocessing to below 1%.

[0159] Distortion correction: due to the radial distortion of the visual camera 5 lens, it needs to be corrected by a polynomial correction model, the formula is as follows:

[0160] ;

[0161] ;

[0162] Where: : pixel coordinates before correction; : pixel coordinates after correction; : distortion coefficient measured by chessboard calibration experiment, the embodiment ; : distance from pixel to image center, unit: pixel, , is the image center coordinates (1024, 768).

[0163] Effect: Eliminate lens distortion caused by mineral water bottle shape distortion such as bottle body narrowing or widening, ensure that the subsequent size measurement error is ≤0.5mm.

[0164] ROI extraction: only the image in the conveyor belt area is retained to exclude background interference, and the ROI area is determined by conveyor belt edge detection, the formula is as follows:

[0165] ;

[0166] ;

[0167] Where: : gray value of pixel ; : Upper and lower boundary of ROI region Unit: Pixel

[0168] Effect: Reduce the amount of invalid image data, shorten the subsequent feature extraction calculation time by 40%, and avoid the interference of background objects such as workshop equipment and personnel on mineral water bottle detection.

[0169] Mineral water bottle feature extraction and positioning module; Based on the pre-processed visual image, this module extracts the multi-dimensional feature size, posture and identification of the mineral water bottle, and realizes the accurate positioning of the mineral water bottle in the physical space through coordinate mapping, providing a coordinate reference for subsequent sorting, labeling and other processes.

[0170] Mineral water bottle size feature extraction

[0171] Through edge detection and contour fitting, the height H, diameter D and bottle mouth radius r of the mineral water bottle are calculated, and the formulas are as follows:

[0172] Calculation of mineral water bottle height

[0173] ;

[0174] Where: H: actual height of mineral water bottle Unit: mm , : y coordinates of the top and bottom of the mineral water bottle in the image Unit: pixels, obtained by Canny edge detection

[0175] : Actual length of the visual camera 5 field of view perpendicular to the direction of the conveying belt Unit: mm, determined by the installation height and focal length of the visual camera 5 , Installation height Visual camera 5 vertical field of view angle, so ; : Image vertical resolution Unit: pixels pixels

[0176] Effect: Distinguish different specifications of mineral water bottles such as 500mL bottles , 1.5L bottles , provide basis for subsequent differentiated processing such as different sorting channel allocation, size measurement error ≤1mm.

[0177] Calculation of mineral water bottle diameter

[0178] ;

[0179] Where: D: maximum diameter of mineral water bottle Unit: mm , : x coordinates of the left and right sides of the mineral water bottle in the image Unit: pixels : Actual width of visual camera 5 field of view in parallel to the direction of the conveyor belt: mm, , Visual camera 5 horizontal field of view angle, so ; : Image horizontal resolution: pixels, pixels.

[0180] Function: Determine whether the mineral water bottle is deformed, such as a standard 500mL bottle , if the detection or is determined to be deformed, and the unqualified mineral water bottle is screened out in advance.

[0181] Mineral water bottle posture and positioning;

[0182] The central axis of the mineral water bottle is detected by the Hough transform, the rotation angle and the center coordinates of the mineral water bottle are calculated, and the physical space positioning is realized:

[0183] Rotation angle calculation:

[0184] ;

[0185] Where: : The rotation angle of the mineral water bottle around the axis perpendicular to the conveyor belt: °, the value range is -15° to 15°, and if it exceeds, it is determined to be an abnormal posture; , , , : The coordinate parameters in the same size calculation.

[0186] Function: Correct the positioning deviation caused by the rotation of the mineral water bottle, for example, in the labeling process, the label needs to be attached to the front of the mineral water bottle, if , the labeling mechanism is controlled to rotate 10° synchronously to ensure that the labeling position deviation is ≤2mm.

[0187] Mineral water bottle center coordinate mapping:

[0188] ;

[0189] ;

[0190] Where: : The coordinates of the center of the mineral water bottle in the physical coordinate system of the conveyor belt: mm, parallel to the direction of the conveyor belt, perpendicular to the direction of the conveyor belt; : The center coordinates of the mineral water bottle in the image: pixels, , ; : Image center coordinate unit: pixel, (1024, 768); , : Image horizontal and vertical resolution unit: pixel, 2048 and 1536 respectively.

[0191] Function: Convert image coordinates to physical coordinates, provide accurate position reference for conveyor speed adjustment and sorting mechanism action, positioning error ≤3mm.

[0192] Mineral water bottle identification recognition;

[0193] Identify the identification on the mineral water bottle label such as brand LOGO, production date through template matching algorithm, the matching similarity formula is as follows:

[0194] ;

[0195] Where: : Matching similarity, value range 0-1, Then it is determined that the matching is successful; : Gray value of ROI region of the image to be detected; : Gray value of template image such as brand LOGO; : The region of interest where the identification is located unit: pixel, determined by label edge detection.

[0196] Function: Realize the brand and batch differentiation of mineral water bottles, for example, when different brands of mineral water bottles are transported on the same conveyor belt, they can be distributed to different sorting channels according to the identification, and the recognition accuracy is ≥99.5%.

[0197] Dynamic speed regulation module: Based on the mineral water bottle position, spacing and specification information output by the mineral water bottle feature extraction module, this module uses a "multi-parameter fusion adaptive speed control algorithm" to dynamically adjust the speed of the driving motor 3, avoid mineral water bottle stacking or idling waste, and at the same time ensure the smooth transportation of mineral water bottles.

[0198] Mineral water bottle spacing calculation first calculates the center spacing of two adjacent mineral water bottles, the formula is as follows:

[0199]

[0200] Where: : The center spacing between the th and the th mineral water bottle unit: mm; , : The physical center coordinates of the th and the th mineral water bottle unit: mm, parallel to the conveyor belt direction.

[0201] Effect: judge whether there is a congestion risk for mineral water bottles, when the maximum mineral water bottle diameter, such as a 1.5L bottle , the threshold is 96mm, and the congestion warning is determined.

[0202] Conveyor target speed calculation;

[0203] Based on the distance between mineral water bottles, specifications, and subsequent process requirements such as sorting and labeling, a weighted fusion algorithm is used to calculate the target speed, as follows:

[0204] ;

[0205] Where: : Conveyor target speed, unit: m / s, value range 0.1-0.5m / s; : Basic speed, unit: m / s, take 0.3m / s corresponding to the optimal transfer speed of 500mL bottle; : Weight coefficient, determined by experimental fitting, distance weight, mineral water bottle specification weight, process time weight, and ; : Ideal mineral water bottle distance, unit: mm, take such as 120mm; : Average diameter of mineral water bottles on the current conveyor, unit: mm, , real-time mineral water bottle quantity; : Single processing time of subsequent processes such as labeling, unit: s, labeling process ; : Basic processing time, unit: s, take 0.5s.

[0206] Effect: dynamically balance mineral water bottle transfer efficiency and stability, for example when congestion warning, , the target speed is reduced; when , the speed is increased to avoid idling.

[0207] Drive motor 3 speed closed loop control;

[0208] PID proportional-integral-derivative algorithm is used to achieve precise control of the speed of drive motor 3, and the PID output formula is as follows:

[0209] ;

[0210] Where: : Drive motor 3 control voltage, unit: V, value range 0-24V; : Speed deviation unit: m / s, , : Actual speed of the conveyor belt collected by the motor encoder, unit: m / s; : Proportional coefficient, take 8.0 for fast response deviation; : Integral coefficient, take 0.5 to eliminate static deviation; : Differential coefficient, take 0.2 to suppress overshoot.

[0211] Effect: Ensure that the actual speed of the conveyor belt tracks the target speed, and the speed deviation is controlled within ±0.02 m / s, avoiding speed fluctuations that cause mineral water bottle position deviation or collision.

[0212] Emergency deceleration control;

[0213] When the posture of the mineral water bottle is detected to be abnormal, such as or the distance is too small , trigger emergency deceleration, and the deceleration acceleration formula is as follows:

[0214] ;

[0215] Where: : Deceleration acceleration unit: m / s², negative sign indicating deceleration; : Acceleration coefficient, take 0.8 to ensure smooth deceleration; : Speed deviation absolute value unit: m / s, , : Emergency speed unit: m / s, take 0.1 m / s.

[0216] Effect: Quickly reduce the speed of the conveyor belt to avoid stacking or damage caused by abnormal mineral water bottles, and the emergency deceleration response time is ≤0.1 s.

[0217] Mineral water bottle quality detection module; This module is based on the gray level, texture, shape features of the visual image, and synchronously detects the appearance defects and content abnormalities of the mineral water bottle without additional sensors, realizing "transfer-detection" integration.

[0218] Scratch detection: use gray level gradient threshold method, the gray level gradient of scratch area is significantly higher than that of normal mineral water bottle, the formula is as follows:

[0219] ;

[0220] If is the gradient threshold, take 30, and the number of continuous pixels is ≥5, it is judged as scratch.

[0221] Where, : Gray level gradient of pixel ; : Gray level gradient of pixel grayscale value; - The gradient threshold was determined through statistical analysis of normal mineral water bottle samples.

[0222] Mineral water bottle deformation detection: Compare the similarity between the mineral water bottle outline and the standard outline, using the following formula:

[0223] ;

[0224] like A deviation of 5% is considered deformation.

[0225] in: Outline similarity; The first step in testing mineral water bottles Distance from each contour point to the center, unit: mm; The first standard mineral water bottle Distance from each contour point to the center, unit: mm; Number of contour points, set to 100; The maximum circumference distance of a standard mineral water bottle is measured in mm.

[0226] Bottle neck damage detection: The bottle neck area is circular. Whether there is damage is determined by calculating the roundness, using the following formula:

[0227] ;

[0228] like If the roundness threshold is met, the bottle opening is considered broken. Where: Roundness, value range 0-1 for perfect circle ; The actual area of ​​the bottle neck region is represented by the following unit: ; Unit of perimeter for the bottle neck area: .

[0229] Liquid level detection: The liquid level is detected by utilizing the difference in grayscale between the water and air inside the mineral water bottle. The formula is as follows:

[0230] ;

[0231] like This refers to the standard liquid level height, such as in a 500mL bottle. If so, it is determined that the liquid level is insufficient. Among them: Actual liquid level height unit: ; The y-coordinate of the liquid level line in the image is in pixels, obtained through grayscale abrupt change detection. : The y-coordinate of the bottom of the mineral water bottle, in pixels;

[0232] Impurity detection: local gray variance method is used, and the gray variance of impurity area is significantly higher than that of normal water body. The formula is as follows:

[0233]

[0234] If is the variance threshold, take 25, then determine that there is impurity.

[0235] Wherein: : gray variance of liquid level area; : number of pixels in liquid level area; : pixel gray value of liquid level area; : average gray value of liquid level area; : the region of interest where the liquid level is located.

[0236] Detection result output;

[0237] Generate quality detection report for each mineral water bottle, including defect types such as "bottle mouth damage" and "low liquid level", defect location physical coordinates, confidence 0-1, ≥0.9 for effective determination, and feedback to the cloud background.

[0238] Conveyer belt tension adaptive compensation module: the conveyer belt will change in tension due to wear and temperature change during long-term operation, which will cause the conveyer belt to slip or deviate; this module detects the edge position of the conveyer belt through visual camera 5, calculates the tension deviation, and controls the tension adjusting mechanism such as tension roller to realize adaptive compensation, ensuring the smooth operation of the conveyer belt.

[0239] Conveyer belt edge position detection detects the x coordinate of the two sides of the conveyer belt through visual image detection, calculates the offset, and the formula is as follows:

[0240]

[0241]

[0242] Wherein: , : offset of left and right edges of conveyer belt, unit: mm, positive for right deviation, negative for left deviation; , : physical x coordinate of actual edge, unit: mm, obtained through edge detection and coordinate mapping; , : physical x coordinate of standard edge, unit: mm, 100mm and 600mm correspond to conveyer belt width 500mm.

[0243] Function: judge whether the conveyer belt is deviated, when or , it is determined that the deviation warning is given.​​​

[0244] Tension deviation calculation;

[0245] There is a linear relationship between the belt deviation and the tension deviation, and the tension deviation formula is as follows:

[0246]

[0247] Wherein: : Tension deviation unit: N, positive for right tension too large, negative for left tension too large; : Tension-deviation coefficient unit: N / mm, measured by experiment, the embodiment , : Belt left and right edge deviation unit: mm.

[0248] Effect: Quantify the degree of tension imbalance, for example , , then ; need to reduce the right tension or increase the left tension.

[0249] Tension compensation control;

[0250] The fuzzy PID algorithm is adopted to control the drive of the servo motor 6 to realize tension adjustment, and the output formula of the fuzzy PID is as follows:

[0251]

[0252] Wherein: : Fuzzy PID output servo motor 6 control voltage unit: V; : Traditional PID output unit: V, , , , : Fuzzy correction coefficient, determined according to the fuzzy rule table of tension deviation and the rate of change of deviation, the value range is 0.8-1.2.

[0253] Effect: Quickly eliminate tension deviation, control the belt deviation within ±2mm, avoid the belt slip, and prolong the service life of the belt.

[0254] Temperature compensation environment;

[0255] Temperature change will cause the thermal expansion and contraction of the belt material, affect the tension, and temperature compensation is needed, and the formula is as follows:

[0256]

[0257] Wherein:​​​​​ : Temperature-induced tension change unit: N; : Conveyor belt temperature coefficient unit: N / (℃·m), rubber conveyor belt ; : Real-time ambient temperature unit: ℃, collected by the temperature sensor built-in the visual camera 5; : Standard temperature unit: ℃, take 25℃; : Effective length of the conveyor belt unit: m, in the embodiment .

[0258] Effect: Correct the impact of temperature on tension, for example , then , the additional tension needs to be offset by the tension adjustment mechanism to ensure stable tension.

[0259] Data encryption and remote monitoring module: This module uses the "edge computing + cloud collaboration" architecture to encrypt the transmission of visual images, control parameters, detection results and other data, and realizes remote monitoring and parameter configuration, solving the problem of data insecurity and lack of remote management in existing systems.

[0260] Image data compression and encryption: JPEG2000 compression algorithm is used to compress the data by 10:1 to reduce the data volume, and then encrypted by AES-256 encryption algorithm. The encryption key generation formula is as follows:

[0261] ;

[0262] Among them: : AES-256 encryption key 256bit; : Device unique identifier 64bit, fixed at factory; : Timestamp 32bit, accurate to seconds; : Random number 128bit, generated by the device's built-in random number generator; : SHA-256 hash function, which maps the input to a 256-bit key.

[0263] Effect: Ensure that image data is not stolen or tampered with during transmission, and the difficulty of cracking encrypted data is ≥ Attempts.

[0264] Control parameter encryption transmission: RSA asymmetric encryption algorithm is used to encrypt speed, tension and other control parameters. The public key is generated by the cloud server and distributed, and the private key is stored locally. The encryption formula is as follows:

[0265] ;

[0266] Among them: : Encrypted ciphertext; : Control parameters to be encrypted ; : Public key exponent is 65537; : Public key modulus is generated by the product of two large prime numbers, , 2048-bit prime number.

[0267] Function: Prevent control parameters from being tampered with, ensure the security of remote configuration, and the parameter transmission error rate is ≤0.1%.

[0268] Remote data transmission adopts LoRa wireless communication technology with a working frequency of 433MHz to realize the communication between the device and the cloud. The transmission power is dynamically adjusted according to the following formula: ;

[0269] Among them: : LoRa module transmission power unit: dBm, value range -10 to 15 dBm; : Minimum transmission power unit: dBm, take -10 dBm; : Transmission distance unit: m, maximum 5000m; : System loss unit: dB, take 3dB cable + antenna loss; : Transmitting and receiving antenna gain unit: dBi, both take 2dBi; : LoRa signal wavelength unit: m, , , , therefore ; : Receiving sensitivity unit: dBm, take -120dBm.

[0270] Function: On the premise of ensuring reliable data transmission, reduce power consumption, transmission packet loss rate ≤0.5%, communication delay ≤1s.

[0271] Remote monitoring and configuration;

[0272] Real-time monitoring: The cloud platform displays the conveyor belt speed, mineral water bottle detection results, and device status such as motor current and tension in real time. The data update frequency is ≤1s, and the historical data query storage period is 1 year.

[0273] Remote parameter configuration: Users modify control parameters such as through the cloud platform. The configuration instruction transmission delay formula is as follows:

[0274] ;

[0275] Among them: : Instruction transmission delay unit: s; : Instruction data length unit: bit, average 1024 bits; : Transmission rate unit: bps, LoRa maximum rate 37500 bps; : Transmission distance unit: m; : Light speed unit: m / s, .

[0276] If , trigger instruction retransmission, ensure successful configuration.

[0277] Function: Realize the remote management of equipment, reduce the on-site maintenance workload, and shorten the fault response time from 4 hours to 1 hour or less.

[0278] Fault self-diagnosis and prediction module; This module realizes real-time diagnosis and progressive fault prediction of equipment failure by analyzing visual image, drive motor 3 current, temperature and other data, and gives early warning of potential failure to reduce downtime loss.

[0279] Drive motor 3 fault diagnosis: Determine whether the drive motor 3 is locked or overloaded by correlating the drive motor 3 current detection and visual speed data, as follows:

[0280] , determine that the drive motor 3 is locked;

[0281] , determine that the drive motor 3 is overloaded;

[0282] Among them: : Drive motor 3 actual current unit: A, collected by current sensor; : Drive motor 3 rated current unit: A, drive motor 3 ; : Conveyor belt actual speed, target speed unit: m / s.

[0283] Conveyor belt slip diagnosis: Compare the theoretical speed corresponding to the drive motor 3 speed with the actual speed detected by vision, as follows:

[0284] ;

[0285] If , determine that the conveyor belt is slipping.

[0286] Among them: : Slip rate unit: %; : Theoretical speed corresponding to drive motor 3 speed unit: m / s, , Drive motor 3 drum radius, drive motor 3 speed unit: rpm.

[0287] Vision camera 5 fault diagnosis: detect image acquisition frame rate and definition:

[0288] If the frame rate or the average gray value of the image underexposed or overexposed, it is determined that the vision camera 5 is faulty.

[0289] Progressive fault prediction uses an LSTM long short-term memory network model to predict progressive faults such as drive motor 3 aging and conveyor belt wear, and the prediction formula is as follows in simplified form:

[0290] ;

[0291] Where: : future probability of failure 01 within a certain time, then trigger an early warning; : sigmoid activation function; : weight matrix; : LSTM hidden layer state containing historical current, temperature, and speed data; : current input data motor current, conveyor belt tension, and temperature; : bias term; : prediction time window unit: h, take 24h.

[0292] Effect: 24 hours in advance to warn of progressive faults, such as motor insulation aging leading to slow increase in current, LSTM model can predict the probability of failure through historical current trend, avoid sudden shutdown, and reduce maintenance frequency by more than 3 times per year.

[0293] Fault handling and alarm:

[0294] Local alarm: when a fault occurs, trigger a 2000Hz frequency buzzer and a red flashing LED light with a frequency of 2Hz, the buzzer and LED light are installed outside the rack 1, and the alarm duration formula is as follows:

[0295] ;

[0296] Where: : alarm duration unit: s; : base duration unit: s, take 10s; : coefficient, take 5s; : probability of failure 01.

[0297] Remote alarm: send fault information type, location, and probability to the cloud platform, and recommend a handling scheme such as "motor locked-rotor, suggest checking if the conveyor belt is jammed", alarm information transmission delay ≤1s.

[0298] Another possible embodiment includes an energy consumption optimization management module. This module uses a "dynamic power consumption allocation" strategy based on the equipment's operating status, such as the number of mineral water bottles and processing procedures, to reduce unnecessary energy consumption while ensuring system performance and achieving energy-saving operation.

[0299] System power consumption analysis: The total power consumption of the system consists of the vision camera 5, drive motor 3, control unit, alarm unit, etc., as shown in the following formula:

[0300] ;

[0301] in: Total system power consumption (W) Visual camera 5 power consumption unit: W, 15W in working state, 2W in sleep state; Power consumption of drive motor 3 (unit: W) , The torque of the drive motor is measured in N·m. The efficiency of the drive motor is 0.85. Control unit power consumption unit: W, fixed at 5W; Alarm unit power consumption unit: W, 8W when alarming, 0W when normal.

[0302] Dynamic power consumption control:

[0303] Vision Camera 5 Sleep Control: Visual detection when there are no mineral water bottles on the conveyor belt Duration At this time, visual camera 5 enters sleep mode, and the wake-up condition is the detection of a mineral water bottle. Hibernation / wake-up response time .

[0304] Drive motor 3 power consumption reduction control: when the number of mineral water bottles At this time, reduce the motor torque using the following formula:

[0305] ;

[0306] in: Target torque unit: N·m; The basic torque unit is N·m, with 10 N·m corresponding to... One mineral water bottle; Real-time number of mineral water bottles; Threshold quantity. Function: Reduces motor idling energy consumption. hour, The motor power consumption is reduced to below 10W, and it is 50W during normal operation.

[0307] Control unit low-power mode: when the system is not in use, such as at night, the control unit enters a low-power mode, reducing power consumption from 5W to 1W. The wake-up conditions are remote instructions or power-on signals.

[0308] Energy consumption statistics and optimization:

[0309] Energy consumption statistics: daily total system energy consumption statistics, formula as follows:

[0310] ;

[0311] Where: : Daily total energy consumption unit: kWh.

[0312] Optimization suggestions: cloud platform generates optimization suggestions based on energy consumption data, such as "18:00-22:00 mineral water bottle quantity is small, suggest reducing motor base torque to 8N·m, expected to save 10% energy."

[0313] Effect: system energy consumption is reduced by 15%-20%, annual power consumption is reduced from 3000kWh to below 2400kWh, while not affecting processing efficiency and accuracy.

[0314] As other embodiments, it also includes a virtual-real collaborative control module based on digital twinning. This module constructs a digital twinning system for the belt conveyor and mineral water bottles. Through real-time data interaction between physical entities and virtual models, dynamic simulation, pre-control, and full-process optimization are achieved. This breaks through the limitations of traditional control systems with "single feedback at the physical layer" and forms an innovative control mode of "virtual-real closed loop", significantly improving the foresight and robustness of the system.

[0315] Digital twinning model construction:

[0316] Geometric modeling: based on CAD drawings of physical components such as conveyor belts, drive motors 3, and adjusting rollers 11, a three-dimensional virtual model with millimeter-level precision is constructed. Model parameters are one-to-one mapped with physical entities, formula as follows:

[0317] ;

[0318] Where: : Geometric model set at time t; : Three-dimensional coordinates of the i-th component, unit: mm, updated in real time with motion state; : Size parameters of the i-th component, unit: mm, such as conveyor belt length , motor roller diameter ; : Material properties of the i-th component, such as elastic modulus , friction coefficient ; : Total number of components, this embodiment Core components include conveyor belt, motor, visual camera 5, etc.

[0319] Function: Accurately reproduce the geometric shape of the physical system, provide the basis for subsequent simulation, the geometric deviation between model and entity ≤0.1%.

[0320] Physical behavior modeling: Based on the theory of multibody dynamics, the component motion and force transmission model is constructed, taking the conveyor belt motion as an example:

[0321]

[0322] Where: : Driving force of driving motor 3, unit: N; : Friction between conveyor belt and rotating roller 2, unit: N, , N for positive pressure, unit: N; : Conveyor belt tension, unit: N, taken from After compensation value; : Total mass of conveyor belt and mineral water bottle, unit: kg, , , The mass of the i-th mineral water bottle is; : Acceleration of conveyor belt, unit: ;

[0323] Function: Simulate the motion state of the conveyor belt under the action of driving force, tension and friction, speed simulation error ≤0.01 m / s, consistency with physical measured value ≥99%.

[0324] Mineral water bottle behavior modeling: modeling the collision and attitude change of mineral water bottle, the collision force formula of mineral water bottle is as follows:

[0325] ;

[0326] Where: : Collision force, unit: N; : Stiffness coefficient, unit: N / m, PET bottle ; : Collision deformation, unit: m; : Damping coefficient, unit: N·s / m, ; : Deformation velocity, unit: m / s.

[0327] Function: Rehearse the collision risk of mineral water bottle during conveying, when The critical force of mineral water bottle breakage is taken as 50 N, the speed adjustment is triggered in advance.

[0328] Real-time interaction between virtual and real data:

[0329] Physical to virtual data mapping: Real-time transmission of physical data such as mineral water bottle position, conveyor belt speed collected by visual camera 5 to virtual model, update frequency is 100Hz, mapping formula is as follows:

[0330]

[0331] Wherein: : The updated value of the virtual model at time t, such as mineral water bottle coordinates, speed; : The measured value of the physical entity at time t; : The predicted value of the virtual model at time t-1; : Weight coefficient, take 0.8 to prefer measured data.

[0332] Effect: Ensure the synchronization of virtual model and physical entity, synchronization delay ≤10ms, solve the limitations of pure physical system data sampling interval.

[0333] Virtual to physical control instruction generation: Based on the simulation results of the virtual model, generate optimized control instructions in advance, the formula is as follows:

[0334]

[0335] Wherein: : The optimized control instruction at future time, such as motor voltage; : Current basic control instruction; : Correction coefficient, take 0.3; : The system state predicted by the virtual model at time, such as mineral water bottle spacing; : Target state, such as ideal spacing ; : Prediction lead time, take 0.5s based on system response speed optimization.

[0336] Effect: Correct the control instruction in advance through virtual rehearsal, for example, predict that the mineral water bottle spacing will be less than the threshold after 0.5s, then reduce the conveyor belt speed in advance, avoid collision caused by physical layer lag, control response lead time ≥0.3s.

[0337] Dynamic simulation and rehearsal control;

[0338] Multi-scenario simulation: Simulate different working conditions in virtual model, such as mineral water bottle density 20 / m, 50 / m, environment temperature 10℃, 40℃, output system performance index energy consumption, failure rate, the deviation between simulation results and physical experiment is ≤5%.

[0339] Control strategy pre-verification: New control parameters such as PID coefficient The effect is verified in the virtual model before being applied to the physical system. The verification index formula is as follows:

[0340]

[0341] wherein, : simulation deviation integral unit: s·m / s or s·mm; : verification time unit: s, 10 s is taken; : simulation state; : target state.

[0342] If the threshold value is 0.1 s·m / s, it is determined that the parameter is effective, otherwise the parameter is re-optimized, and the parameter debugging period is shortened from 24 hours to 2 hours.

[0343] Fault injection simulation: injecting faults such as sudden increase of drive motor 3 current and tearing of conveying belt in the virtual model, simulating the fault evolution process, and generating the optimal emergency control strategy, for example, when the drive motor 3 is locked, the virtual simulation shows that "immediate power-off + braking" can reduce the damage to the equipment more than "simple deceleration", and the damage degree is reduced.

[0344] Whole-process optimization decision;

[0345] Based on the historical data and real-time simulation results of the digital twin system, the global control strategy is optimized by using the reinforcement learning algorithm, and the reward function formula is as follows:

[0346]

[0347] wherein, : reward value at time t, 0-1; : weight, respectively taken as 0.4, 0.3, 0.3, energy consumption, efficiency, and failure rate weight; : real-time energy consumption unit: W, ; : processing efficiency unit: pieces / s, pieces / s; : failure probability 0-1, .

[0348] The reinforcement learning agent maximizes the cumulative reward , and autonomously optimizes parameters such as speed regulation and tension compensation, so that the system comprehensive performance energy consumption-efficiency-stability weighted value is improved.

[0349] To sum up, in the positioning accuracy aspect, the multi-dimensional detection of the bottle size, posture and position is realized by means of the visual camera 5, which greatly improves the positioning accuracy and can meet the needs of subsequent high-precision sorting and labeling processes. In terms of speed control, dynamic adjustment is made based on the bottle spacing and specifications, which not only avoids the problem of bottle stacking congestion, but also reduces the waste caused by the empty running of the conveyor belt, while reducing the bottle breakage rate and achieving efficient and stable transfer. The quality detection realizes integration, and without additional detection equipment, the appearance defects and content abnormalities of the bottle can be detected synchronously, which not only improves the detection accuracy, but also reduces the equipment cost and the production line area. The fault self-diagnosis and prediction function can provide early warning of potential faults, shorten the fault handling time, reduce the annual maintenance frequency and reduce the loss caused by downtime. Remote management and energy saving features are prominent, using LoRa remote monitoring and AES-256 encryption technology to ensure data security and the convenience of remote configuration, and dynamic energy management effectively reduces annual energy consumption, balancing intelligence and economy. The system has strong compatibility and can support differentiated processing of different specifications of mineral water bottles without the need to replace sensors or adjust mechanical structures, and has high adaptability, can quickly access existing production lines and shorten the transformation cycle. The digital twin module realizes real-time interaction between physical entities and virtual models, responds to control requirements in advance, improves the accuracy of collision warning and solves the problem of physical system lag.

[0350] In addition, in the embodiment, a hybrid control architecture of "edge controller + PLC cooperation" is adopted to meet the dual needs of high-precision real-time control and complex algorithm operation. Among them, the edge controller selects an industrial embedded controller such as Advantech UNO-2484G, which is equipped with an Intel Core i5 processor with 4 cores and 3.0 GHz and 8 GB DDR4 memory, responsible for visual image preprocessing, bottle feature extraction, digital twin modeling, reinforcement learning optimization and other computationally intensive tasks, supporting Python / C++ algorithm development, and realizing 1ms-level data interaction with other modules through EtherCAT bus, ensuring real-time operation of complex models and collaborative scheduling of multiple modules. The programmable logic controller (PLC) selects Siemens S7-1214C, which is equipped with 16-point digital input / 16-point digital output and 4-way analog input, responsible for real-time control tasks such as conveyor motor speed closed-loop control, tension adjustment mechanism driving and fault alarm, and communicates data with the edge controller through the PROFINET protocol with a 10ms update period, and uses its high anti-interference to meet the EN61131-2 standard to ensure the stable operation of the bottom execution mechanism. When working together, the edge controller generates optimized control instructions such as target speed and tension compensation amount by analyzing visual data and virtual simulation results, and issues them to the PLC for execution.

[0351] In the description of the application, it is necessary to point out that, unless otherwise explicitly specified and limited, the terms "mounting", "connection", "linking" should be understood in a broad sense, for example, can be fixed connection, can also be detachable connection, or integrally connected; can be directly connected, or indirectly connected through an intermediate medium, can be the communication inside two elements. For those skilled in the art, the specific meaning of the above terms in the application can be understood according to the specific circumstances.

[0352] Although embodiments of the application have been shown and described, the scope of the application is defined by the appended claims and their equivalents.

Claims

1. A mineral water bottle handling and conveying device, comprising a belt conveyor and a control and coordination system, wherein the belt conveyor comprises a frame (1) and a drive motor (3) fixedly connected to the outside of the frame (1), and rollers (2) are rotatably connected to both ends of the frame (1), the output end of the drive motor (3) is fixedly connected to the rollers (2), and a conveyor belt is sleeved on the outside of the rollers (2), characterized in that, The frame (1) is externally fixedly connected to a mounting bracket (4), and a vision camera (5) is fixedly connected to the mounting bracket (4). A tension adjustment mechanism is also provided below the frame (1). The control and coordination system includes a visual information acquisition and preprocessing module, a mineral water bottle feature extraction and positioning module, a dynamic speed adjustment module, a mineral water bottle quality detection module, a tension adaptive compensation module, a data encryption and remote monitoring module, a fault self-diagnosis and prediction module, and an energy consumption optimization management module.

2. The mineral water bottle handling and conveying device according to claim 1, characterized in that, The tension adjustment mechanism includes a mounting base (7) and an adjustment roller (11) rotatably connected to the mounting base (7). A servo motor (6) is fixedly connected to the outside of the frame (1). A lead screw (8) is fixedly connected to the output end of the servo motor (6). A nut (9) is threaded onto the outside of the lead screw (8). The nut (9) is fixedly connected to the mounting base (7). A guide rod (10) is also fixedly connected to the frame (1). The guide rod (10) passes through the mounting base (7).

3. The mineral water bottle handling and conveying device according to claim 2, characterized in that, The visual information acquisition and preprocessing module steps are based on images acquired by the visual camera (5), and automatically adjust the exposure time and gain of the visual camera (5) according to the following formula: Exposure time calculation: ； in: Camera exposure time; Camera ISO (sensitivity) Camera target output voltage; Real-time ambient light intensity; : Camera photosensitive area; Gain adjustment formula: ； in: : Actual camera gain; Base gain; Standard ambient light intensity; Real-time ambient light intensity and exposure time calculation Consistent; Image preprocessing; The acquired raw images need to undergo three preprocessing steps: noise reduction, distortion correction, and region of interest extraction. Gaussian noise reduction: Adaptive Gaussian filtering is used to eliminate noise caused by dust and electromagnetic interference. The size of the filter kernel is dynamically adjusted according to the noise intensity of the image, as shown in the following formula: ； in: Gaussian filter standard deviation; Average grayscale value of the image; Distortion correction: Due to radial distortion in camera lenses, it needs to be corrected using a polynomial correction model, as shown in the following formula: ； ； in: Pixel coordinates before correction; Corrected pixel coordinates; : Distortion coefficient; : Distance from pixel to the center of image , The coordinates of the image center; ROI extraction: Only the image within the conveyor belt area is retained. The ROI area is determined by conveyor belt edge detection, as shown in the following formula: ； ； in: : pixel grayscale value; : The upper boundary of the ROI region; : The lower boundary of the ROI region.

4. The mineral water bottle handling and conveying device according to claim 3, characterized in that, The mineral water bottle feature extraction and localization module extracts multi-dimensional features of the mineral water bottle based on the preprocessed visual image and realizes the localization of the mineral water bottle in physical space through coordinate mapping. Extraction of dimensional features from mineral water bottles; The height H, diameter D, and mouth radius r of the mineral water bottle are calculated using edge detection and contour fitting, as shown in the following formula: Mineral water bottle height calculation: ； Where: H: actual height of the mineral water bottle; : The y-coordinate of the top of the mineral water bottle in the image; The y-coordinate of the bottom of the mineral water bottle in the image was obtained using Canny edge detection. The field of view of the vision camera (5) is the actual length perpendicular to the direction of the conveyor belt; Image vertical resolution; Mineral water bottle diameter calculation: ； Where: D: Maximum diameter of the mineral water bottle; The x-coordinate of the left side of the mineral water bottle in the image; : The x-coordinate of the right side of the mineral water bottle in the image; The field of view of the visual camera (5) is the actual width parallel to the direction of the conveyor belt; Image horizontal resolution; Posture and positioning of mineral water bottles; By detecting the central axis of the mineral water bottle using Hough transform, and calculating the bottle's rotation angle and center coordinates, physical spatial positioning is achieved. Rotation angle calculation: ； in: The angle of rotation of the mineral water bottle around a axis perpendicular to the conveyor belt; Mineral water bottle center coordinate mapping: ； ； in: The coordinates of the center of the mineral water bottle in the physical coordinate system of the conveyor belt. To be parallel to the direction of the conveyor belt, Perpendicular to the direction of the conveyor belt; : Coordinates of the center of the mineral water bottle in the image , ; Image center coordinates; , Image horizontal and vertical resolution; Mineral water bottle label identification; The template matching algorithm is used to identify the markings on mineral water bottle labels. The similarity formula is as follows: ； in: Matching similarity, with a value ranging from 0 to 1. Then the match is considered successful; : Gray value of the ROI region in the image to be detected; The grayscale value of the template image; The region of interest where the label is located is determined by label edge detection.

5. The mineral water bottle handling and conveying device according to claim 4, characterized in that, The dynamic speed adjustment module dynamically adjusts the speed of the drive motor (3) to avoid the mineral water bottles from stacking or running dry. To calculate the spacing between mineral water bottles, first calculate the center-to-center distance between two adjacent mineral water bottles, using the following formula: ； in: : No. The and the first The center-to-center distance between the mineral water bottles; , : No. The, the The physical center coordinates of each mineral water bottle are parallel to the direction of the conveyor belt; Calculation of target speed for conveyor belt; Based on the spacing and specifications of the mineral water bottles, as well as the requirements of subsequent processes, a weighted fusion algorithm is used to calculate the target speed, as shown in the following formula: ； in: Target speed of the conveyor belt; Base speed; Weighting coefficients; Ideal spacing between mineral water bottles; The average diameter of the mineral water bottles currently on the conveyor belt. , This represents the real-time number of mineral water bottles. : Single processing time for subsequent processes; Drive motor (3) speed closed-loop control; The PID algorithm is used to achieve precise control of the speed of the drive motor (3). The PID output formula is as follows: ； in: : Drive motor (3) control voltage; Speed ​​deviation, , This refers to the actual speed of the conveyor belt. : Proportional coefficient; Integral coefficient; Differential coefficients; Emergency deceleration control; When an abnormal posture or excessively small distance between mineral water bottles is detected, emergency deceleration is triggered. The deceleration acceleration formula is as follows: ； in: : deceleration / acceleration, the negative sign indicates deceleration; : Acceleration coefficient; : Absolute value of speed deviation , For emergency speed.

6. The mineral water bottle handling and conveying device according to claim 5, characterized in that, The mineral water bottle quality inspection module is based on the grayscale, texture, and shape features of visual images to simultaneously detect appearance defects and abnormal contents of mineral water bottles. Scratch detection: The gray-level gradient threshold method is used. The gray-level gradient of the scratched area is significantly higher than that of a normal mineral water bottle. The formula is as follows: ； like If the gradient threshold is set to 30 and the number of consecutive pixels is ≥5, then it is considered a scratch. in, : pixel The grayscale gradient; : pixel grayscale value; The gradient threshold was determined through statistical analysis of normal mineral water bottle samples. Mineral water bottle deformation detection: Compare the similarity between the mineral water bottle outline and the standard outline, using the following formula: ； like If so, it is determined to be a deformation; in: Outline similarity; The first step in testing mineral water bottles The distance from each contour point to the center; The first standard mineral water bottle The distance from each contour point to the center; Number of contour points, set to 100; : The maximum outline distance of a standard mineral water bottle; Bottle neck damage detection: The bottle neck area is circular. Whether there is damage is determined by calculating the roundness, using the following formula: ； like If so, it is determined that the bottle opening is damaged; among which: : Roundness, with a value range of 0-1; : The actual area of ​​the bottle neck region; : The perimeter of the bottle neck area; Liquid level detection: The liquid level is detected by utilizing the difference in grayscale between the contents of a mineral water bottle and the air. The formula is as follows: ； like ;in: Actual liquid level height; The y-coordinate of the liquid level line in the image is obtained through grayscale abrupt change detection; : The y-coordinate of the bottom of the mineral water bottle; Impurity detection: The local gray-scale variance method was used. The gray-scale variance of the impurity area was significantly higher than that of the normal water body. The formula is as follows: ； like If so, it is determined that impurities are present. The variance threshold is set to 25. in: : Gray-scale variance of the liquid level region; : Number of pixels in the liquid level area; : Grayscale value of pixels in the liquid level area; : Average gray value of the liquid level area; : The region of interest where the liquid level is located.

7. A mineral water bottle handling and conveying device according to claim 6, characterized in that, The tension adaptive compensation module steps involve detecting the edge position of the conveyor belt using a vision camera (5), calculating the tension deviation, and controlling the tension adjustment mechanism to achieve adaptive compensation, as detailed below: Conveyor belt edge position detection uses visual images to detect the x-coordinates of both sides of the conveyor belt and calculates the offset using the following formula: ； ； in: : The offset of the left side of the conveyor belt; : The offset of the right edge of the conveyor belt, positive for rightward offset and negative for leftward offset; , The physical x-coordinate of the actual edge is obtained through edge detection and coordinate mapping; , : Physical x-coordinate of the standard edge; Tension deviation calculation; There is a linear relationship between conveyor belt offset and tension deviation. The tension deviation formula is as follows: ； in: Tension deviation: positive indicates excessive tension on the right side, negative indicates excessive tension on the left side; Tension-offset coefficient; , : Offset of the left and right edges of the conveyor belt; Tension compensation control; The fuzzy PID algorithm is used to control the drive of the servo motor (6) to achieve tension adjustment. The output formula of the fuzzy PID is as follows: ； in: : Fuzzy PID output servo motor (6) control voltage; Traditional PID output, , , , ; : Fuzzy correction coefficient, based on tension deviation and the rate of change of deviation ( The fuzzy rule table is determined, with values ​​ranging from 0.8 to 1.2; Temperature-compensated environment; Temperature changes cause conveyor belt materials to expand and contract, affecting tension. Temperature compensation is necessary, as shown in the following formula: ； in: : Temperature-induced changes in tension; Temperature coefficient of conveyor belt, rubber conveyor belt ; Real-time ambient temperature is collected by the built-in temperature sensor of the vision camera (5); Standard temperature, taken as 25℃; Effective length of the conveyor belt.

8. A mineral water bottle handling and conveying device according to claim 7, characterized in that, The data encryption and remote monitoring module encrypts and transmits visual images, control parameters, and detection result data. Image data compression and encryption: The JPEG2000 compression algorithm is used to reduce the data size, and then the AES-256 encryption algorithm is used for encryption. The encryption key generation formula is as follows: ； in: AES-256 encryption key; Unique equipment identifier, fixed at the factory; Timestamp, accurate to the second; Random numbers are generated by the device's built-in random number generator. The SHA-256 hash function maps the input to a 256-bit key. Encrypted transmission of control parameters: RSA asymmetric encryption algorithm is used to encrypt control parameters such as speed and tension. The public key is generated and distributed by the cloud server, and the private key is stored locally on the device. The encryption formula is as follows: ； in: : The encrypted ciphertext; : Control parameters to be encrypted; Public key index; Public key modulus; Remote data transmission uses LoRa wireless communication technology to enable communication between the device and the cloud. The dynamic adjustment formula for transmission power is as follows: ; in: LoRa module transmission power, ranging from -10 to 15 dBm; Minimum transmission power, set to -10dBm; Transmission distance; System losses; : Transmit and receive antenna gain; LoRa signal wavelength; Receiver sensitivity; Real-time monitoring: The cloud platform displays the conveyor belt speed, mineral water bottle detection results, and equipment status in real time, with a data update frequency of ≤1 second and supports historical data query; Remote parameter configuration: Users modify control parameters through the cloud platform. The formula for the transmission delay of configuration commands is as follows: ； in: Command transmission delay; : Instruction data length; : Transmission rate; Transmission distance; Speed ​​of light; like If so, the command will be retransmitted to ensure successful configuration.

9. A mineral water bottle handling and conveying device according to claim 8, characterized in that, The fault self-diagnosis and prediction module analyzes visual images, drive motor (3) current and temperature data to realize real-time diagnosis and progressive fault prediction of equipment faults. Fault diagnosis of drive motor (3): The current detection of drive motor (3) is used to determine whether drive motor (3) is stalled or overloaded. The formula is as follows: Then the drive motor (3) is determined to be stalled; Then the drive motor (3) is determined to be overloaded; in: : Actual current of drive motor (3); : Rated current of drive motor (3); : Actual speed of the conveyor belt; Target speed of the conveyor belt; Conveyor belt slippage diagnosis: Compare the theoretical speed corresponding to the rotational speed of the drive motor (3) with the actual speed detected by vision, as shown in the following formula: ； like If so, it is determined that the conveyor belt is slipping; in: Slippage rate; The theoretical speed corresponding to the rotational speed of the drive motor (3), To drive the motor (3) speed; Visual camera (5) fault diagnosis: detect the image acquisition frame rate and sharpness, the formula is as follows: If frame rate Or the average gray value of the image Then underexposed, or If the image is overexposed, it is determined to be a malfunction of the visual camera (5); Progressive fault prediction uses an LSTM model to predict progressive faults in drive motor (3) aging and conveyor belt wear. The prediction formula is as follows: ； in: :future Failure probability over time This will trigger an alert; : sigmoid activation function; Weight matrix; LSTM hidden layer state; : Current input data; : Bias term; : Predicted time window; Local alarm: When a fault occurs, the buzzer and LED light are triggered. The alarm duration is calculated using the following formula: ； in: Alarm duration; Base duration; : Coefficient, taken as 5s; Failure probability; Remote alarm: Sends fault information to the cloud platform and recommends solutions.

Images