Earth pressure balance shield discharge volume estimation method

Abstract

The application provides a soil pressure balance shield slag discharge volume estimation method, which estimates the soil pressure balance shield tunneling process slag discharge volume according to the depth image of the discharged slag soil on the belt conveyor, and comprises the following steps: firstly, a depth camera is installed above the belt conveyor, which is responsible for collecting the slag soil depth image; then, in the no slag soil discharge stage, the slag soil volume calculation parameters are calculated or determined according to the background depth image, and the image detection area for slag soil identification and volume calculation is demarcated; after the shield machine starts to advance, the depth camera collects the depth image in real time, the slag soil target in the image is identified by using the computer vision technology, and the slag soil volume is calculated; after the advance of each ring is completed, the discharged slag soil volume calculated in the advance process is accumulated to obtain the final discharged slag soil volume of each ring. The method is simple and direct, has few calculation assumptions, is not dependent on artificial experience and operation, has low equipment cost, and is suitable for use.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of earth pressure balance tunnel construction, and designs a method for real-time estimation of shield residue volume based on computer vision technology. BACKGROUND

[0002] With the continuous acceleration of urbanization in China, the mileage and scale of subways have also rapidly expanded, which has brought great challenges to the safety of tunnel construction and the control of the impact on the surrounding environment. Shield method is widely used in the construction of subway tunnels due to its fast construction speed, small impact on the surrounding environment, and high degree of mechanization. However, during large-scale shield construction, due to the influence of uncertain factors such as complex geology and unstable construction level, accidents frequently occur during urban subway construction, causing serious economic losses, casualties, and social adverse effects. Ground loss caused by shield tunneling is an important reason for ground subsidence or collapse, and the value of this important parameter at the construction site still mainly relies on experience. The method for reasonable, rapid, and accurate determination of ground loss needs further exploration and research. Ground loss is directly related to the volume of excavated soil during tunneling, and the residue volume of the shield machine can reflect the amount of soil excavation, which can be used to estimate ground loss and ground subsidence, and then adjust the shield construction parameters, reduce the construction risk, improve the construction quality, and more effectively ensure the safety of construction. However, the estimation of residue volume at the construction site is very rough, and generally the number of residue transport vehicles or the weight of residue transport vehicles is directly counted to roughly determine the amount of residue. However, the residue in the residue transport vehicle is not completely compacted, and there are a lot of voids; and the density of the residue is difficult to accurately determine, and the estimation of volume by weight also has a large deviation, so such estimation results cannot be used for fine construction adjustment. Residue is discharged from the screw soil removal machine to the belt conveyor, and direct estimation of residue volume during the discharge process will obtain more accurate results. Therefore, a method for real-time and accurate estimation of residue volume during residue discharge can provide important construction information for construction personnel to assist them in adjusting construction decisions and ensuring safe and stable tunneling.

[0003] During the process of shield tunneling, the amount of external additive and the soil chamber pressure maintaining mode will change with the change of geological conditions. In addition, the adjustment of the soil chamber pressure maintaining mode is also related to the technical level of the shield driver. The use of the average value of the external additive quality and the average value of the additional discharge quality in a certain section to correct the estimated value of the discharge amount has the problems of large error and low generalization. The calculation of the discharge amount by establishing a theoretical equation has many calculation assumptions, and it is difficult to determine the value of many parameters in the equation, which has the problem of strong experience in parameter value selection. The discharge amount can only be roughly estimated, and the equation calculation is complex and the calculation speed is difficult to guarantee. Real-time scanning and volume calculation of the transported spoil on the belt conveyor by using a laser scanner can directly obtain the estimated value of the spoil volume. This process has fewer assumptions and simple and direct calculation, but this method needs to install a laser scanner, which has the problem of high cost of expensive equipment.

[0004] Currently, the following solutions are mainly used to estimate the volume of the spoil discharged by the shield machine:

[0005] A method for estimating and correcting the discharge amount of an earth pressure balance shield is proposed in Chinese patent application No. 202111459641.X. This method calculates the theoretical discharge quality of the shield using the tunneling parameters of the normal construction section after cleaning, and then calculates the additional discharge quality of the shield by combining the actual discharge quality of the shield. The average value of the additional discharge quality is used to optimize and correct the estimated value of the discharge amount of the shield. This method takes into account the changes in the discharge quality of the shield caused by the adjustment of the external additive and the soil chamber pressure maintaining mode, as well as factors such as groundwater. The calculated value of the predicted discharge quality of the shield is close to the actual discharge quality, and the prediction accuracy and reliability of the discharge amount of the shield per ring are further improved through cyclic correction.

[0006] A method for quickly calculating the discharge amount of an earth pressure balance shield machine is proposed in Chinese patent application No. 202010186622.3. This method proposes an energy relationship equation for the screw conveyor by analysis, and establishes an equation set that can numerically solve the discharge amount of the earth pressure balance shield machine by combining the motion condition equation and the force state equation of the screw conveyor. The Newton iteration method is used to determine the value range and value of the unique redundant unknown parameter shear stress proportion coefficient under the relevant domain conditions of the discharge amount. The determined shear stress proportion coefficient is substituted into the equation to solve the discharge amount of the earth pressure balance shield machine.

[0007] A dynamic measurement method for the discharge amount of an earth pressure balance shield machine is proposed in Chinese patent application No. 201911033576.7. This method requires installing a laser scanner directly above the spoil conveying belt climbing section. The scanning angle of the laser scanner forms a fan-shaped profile area. First, calculate the difference between the fan-shaped profile areas scanned per unit time when the belt is idling and when conveying spoil. Then, by combining the spoil conveying speed, integrate in time to obtain the discharge amount. SUMMARY

[0008] The present application aims to solve the problem that the earthwork volume in the shield tunneling process is difficult to be accurately estimated in the engineering site, and the stratum loss information cannot be accurately obtained. The present application provides a soil pressure balance shield discharge volume estimation method, which can realize real-time and accurate estimation of the discharge volume of spoil in the shield tunneling process based on the spoil depth image captured by the depth camera, thereby providing an important calculation basis for engineers to subsequently predict the stratum loss caused by shield tunneling and the surface subsidence amount that may be caused.

[0009] Technical scheme:

[0010] A soil pressure balance shield discharge volume estimation method estimates the discharge volume in the soil pressure balance shield tunneling process according to the depth image of the discharged spoil on the belt conveyor, and includes the following steps:

[0011] Step 1, first, a depth camera is installed at a suitable position above the belt conveyor, which is responsible for collecting the spoil depth image.

[0012] Step 2, then, in the spoil discharge-free stage, the calculation parameters required in the spoil volume calculation formula are calculated or determined according to the background depth image, and the image detection area for spoil recognition and volume calculation is demarcated. If the range of the image detection area is not suitable, the camera installation position needs to be adjusted, and the adjustment is repeated until the detection area is demarcated in a suitable spoil flow area.

[0013] Step 3, after the shield machine starts advancing, the depth camera collects the depth image in real time, and the spoil target in the image is recognized and the spoil volume is calculated by using computer vision technology.

[0014] Step 4, after the advance of each ring is completed, the discharged spoil volume calculated in real time during the advance is accumulated, and the final discharged spoil volume of each ring can be obtained.

[0015] Compared with the prior art, the present application has the following advantages and beneficial effects:

[0016] The present application is a soil pressure balance shield discharge volume estimation method, which is based on computer vision technology to recognize the discharged spoil depth image, extracts the spoil depth information in the continuous discharge process, and estimates the discharge volume. Using the method of the present application, spoil image data can be collected in real time during the soil pressure balance shield tunneling process, and the discharge volume can be estimated and accumulated in real time by using computer vision technology. The method is simple and direct, has few calculation assumptions, does not depend on artificial experience and operation, and has low equipment cost and strong applicability. The characteristics are low cost, automation, and applicability to various soil strata in the soil pressure balance shield tunneling tunnel. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1Method flow chart of the present application;

[0018] Figure 2 Depth camera installation schematic diagram;

[0019] Figure 3 Residual volume formula calculation parameter determination flow;

[0020] Figure 4 Detection area demarcation schematic diagram;

[0021] Figure 5 Identify the target flow of slag. DETAILED DESCRIPTION

[0022] Terminology and general knowledge:

[0023] Tunnel: a linear engineering building embedded in the stratum, which is a form of human utilization of underground space.

[0024] Subway: a rapid, large-capacity, electric traction tunnel track transportation built underground in a city.

[0025] Shield method: a tunnel construction method suitable for soft soil, using a shield machine to excavate the stratum and assemble the tunnel segment.

[0026] Shield machine: a construction machine composed of a shell, a cutter head, a pushing device, an assembly device and other supporting devices. The shell is a cylinder that plays a protective role, and the other devices are inside the shell.

[0027] Earth pressure balance shield: a type of shield machine that uses the soil cut by the rotating cutter head at the front end of the shield to fill the soil compartment when the shield is advancing. The passive earth pressure is balanced with the earth pressure and water pressure on the excavation face, so that the excavation face and the shield face are in a balanced state.

[0028] Earth pressure balance shield slag: the discarded soil excavated when the earth pressure balance shield is advancing.

[0029] Spiral earth removal machine: a machine that transports the soil cut by the cutter head to the belt conveyor through the rotation of the spiral member. The rotation speed of the spiral member determines the speed of the earth removal.

[0030] Belt conveyor: a belt device that transports the soil from the spiral earth removal machine to the slag car.

[0031] Shield construction parameters: various parameters that need to be set during the construction of the shield machine, such as cutter head speed, etc. Whether the parameters are set reasonably determines the safety and quality of the shield construction.

[0032] Strata loss: strata loss is the difference between the actual excavated soil volume and the tunnel volume after construction in shield construction.

[0033] Image enhancement: By means of some methods, additional information or transformed data is added to the original image, and the features of interest in the image are selectively highlighted or some unwanted features in the image are suppressed (masked), so that the image is matched with the visual response characteristics.

[0034] Computer vision: Computer vision is a science that studies how to make machines "see". More specifically, it refers to using cameras and computers to replace human eyes to identify, track and measure targets, and further perform image processing to make the computer processing more suitable for human observation or transmission to the instrument detection.

[0035] Depth camera: Depth camera (3D camera) can detect the distance information of the shooting space. By means of depth camera, the distance from each point in the image to the camera is obtained, and then the two-dimensional coordinates of the point in the 2D image are added, so that the three-dimensional space coordinates of each point in the image can be obtained.

[0036] Depth image: Depth image, also known as distance image, refers to the image taking the distance (depth) from the image collector to each point in the scene as the pixel value, which directly reflects the geometric shape of the visible surface of the scene.

[0037] The technical solutions provided by the present application will be further described below in combination with specific embodiments and their accompanying drawings. The advantages and features of the present application will be more apparent in combination with the following description.

[0038] Embodiments

[0039] The overall flow of the embodiment of the present application is shown in Figure 1

[0040] A soil pressure balance shield slag discharge volume estimation method estimates the soil pressure balance shield tunneling process slag discharge volume according to the depth image of the discharged slag soil on the belt conveyor.

[0041] Step 1, first install a depth camera at a suitable position above the belt conveyor, which is responsible for collecting the slag depth image.

[0042] Step 2, then in the no slag discharge stage, calculate or determine the calculation parameters required in the slag volume calculation formula according to the background depth image, and demarcate the image detection area for slag identification and volume calculation. If the image detection area range is not suitable, the camera installation position needs to be adjusted, and the adjustment is repeated until the detection area is demarcated in a suitable slag flow area.

[0043] Step 3, after the shield machine starts to advance, the depth camera collects depth images in real time, and uses computer vision technology to identify the slag target in the image and calculate the slag volume.

[0044] ​Step 4, after the end of each ring advance, the volume of the slag discharged is accumulated in real time, and the final volume of the slag discharged in each ring is obtained.

[0045] Step 1, install the depth camera:

[0046] The installation diagram of the depth sensor depth camera is shown in Figure 2 .

[0047] The depth camera should be installed directly above the belt conveyor, ensuring that the camera optical axis is perpendicular to the belt surface of the belt conveyor, and the camera centerline is as consistent as possible with the centerline of the belt conveyor. The installation height should ensure that the horizontal viewing angle of the camera can cover the width of the belt conveyor, and the vertical length of the detection area calculated according to the installation height in Step 2 should be within the vertical viewing angle range. The installation point of the depth camera should be kept at a certain distance from the soil outlet of the screw conveyor, ensuring that the slag soil has been completely broken on the belt conveyor when it enters the image detection range, and maintaining the same motion speed as the belt conveyor.

[0048] Pay attention to details:

[0049] Do not place high-intensity concentrated light sources near the belt conveyor to avoid affecting the quality of the depth image collected by the depth camera.

[0050] Keep the camera lens surface clean and dust-free to ensure that the depth image captured is clear, complete, and easy to identify.

[0051] Step 2, determine the calculation parameters:

[0052] In the calculation formula of the slag volume, two important calculation parameters need to be determined in advance, as shown in the flowchart Figure 3 .

[0053] Step 2.1, before each ring of shield cutting and advancing, determine the parameters:

[0054] First, use the depth camera to collect the depth image of the belt surface of the belt conveyor, i.e. the background depth image, and save the background depth matrix D b . According to this, the distance D0 between the belt surface and the camera can be directly determined, i.e. the background depth D0. Then calculate the real-world area S0 represented by each pixel in the depth image on the plane at a distance of background depth D0. The parameters required for this calculation process include the camera horizontal viewing angle α, the vertical viewing angle β, the total number of horizontal pixels W, and the total number of vertical pixels L. According to the following calculation formulas (1) and (2), the horizontal width w0 and the vertical length l0 of the real world corresponding to each pixel in the image on the plane at depth D0 can be calculated respectively.

[0055] Then, the area S0 of each pixel in the real world can be obtained by multiplying the length and width of the pixel in the real world, as shown in equation (3).

[0056]

[0057]

[0058] S0 = w0 x l0 (3)

[0059] Step 2.2: Define the detection area and perform calculations:

[0060] The schematic diagram of image detection area is shown in Figure 4 All subsequent image processing, slag target recognition, and volume calculation are only considered within the detection area. The slag discharge during shield tunneling is a continuous and dynamic process. In order to ensure that no slag targets are missed and no repeated calculations of slag targets are performed, a slag detection area needs to be defined. Only the slag entering this area will be identified and the volume calculated.

[0061] The horizontal width W d of the detection area should be slightly larger than the width of the belt conveyor. The vertical length L d of the detection area needs to be determined based on the movement speed v of the slag on the belt conveyor, the frame interval time t0 of the slag depth picture, and the number of slag volume statistics K, as shown in equation (4).

[0062] wherein,

[0063] The movement speed v of the slag on the belt conveyor is consistent with the belt conveyor speed v, which can be directly obtained from the automatic recording of the construction data of the shield machine.

[0064] The number of slag volume statistics K represents that the slag volume of each pixel is calculated K times during the process of passing through the detection area. The total slag volume in the ring passing through the detection area can be obtained by summing the slag volume calculated for each frame of picture in the ring and dividing by the number of slag volume statistics K.

[0065] The product of the movement speed of the slag and the frame interval time represents the distance of the slag moving forward between two picture collections. Multiplying by the number of statistics K represents the forward distance of the slag during K consecutive frame collections. Dividing by the vertical length l0 of each pixel in the real world can obtain the number of pixels in the vertical direction of the detection area, i.e., the vertical length L d The detection area is selected by human in the middle of the camera field of view.

[0066] L d = (v x K x t0) / l0 (4)

[0067] If the vertical length of the image detection area calculated by the current installation height exceeds the visual angle range of the muck moving on the belt conveyor at the common speed, the camera installation position needs to be adjusted until the picture of the muck discharged by the belt conveyor at the same speed is completely in the detection area.

[0068] Step 3, on-site configuration computer and run program for identifying muck target:

[0069] The identification process of the muck target is shown in Figure 5 .

[0070] Step 3.1 When the shield starts to excavate in a certain ring, the muck volume calculation program is started, and the depth camera starts to collect depth images.

[0071] Step 3.2 The program first performs muck movement detection on each collected depth image, that is, it judges whether there is moving muck in the current detection area.

[0072] This step has two reasons:

[0073] First, it is to determine whether there is muck in the detection area. If there is no muck, there is no need to perform subsequent muck image processing, muck target identification, and muck volume calculation operations, thereby improving the calculation efficiency.

[0074] Second, during the pushing process of each ring of the shield, when it is necessary to replace the muck car or supplement the slurry, there will be a time to pause the pushing, at which time the belt conveyor also stops working, and part of the muck may be left in the detection area. At this time, if the muck volume in the area is always calculated and accumulated, the muck discharge amount of the ring will be greatly overestimated.

[0075] Step 3.3 Each depth image needs to be compared with the previous depth image to detect the depth values of the pixels in the detection area. If the sum of the depth difference values of the pixels in the detection area of the adjacent two depth images is less than a certain threshold, it can be considered that the muck is not moving at this time, and there is no need to perform subsequent target identification and volume calculation operations. On the contrary, if the sum of the depth difference values of the pixels in the detection area of the adjacent two depth images is greater than a certain threshold, it can be considered that the muck is in the process of being discharged, and the subsequent operation can be performed on the depth image.

[0076] Step 3.4 For the muck depth image that passes the muck movement detection, first save its depth matrix D, then calculate the difference D b -D between the background depth matrix and the muck depth matrix, to obtain the difference image, which highlights the muck part in the image.

[0077] Step 3.5 Image enhancement is performed on the difference image, mainly for image denoising and hole filling. For active imaging depth cameras, the light source emitted may be absorbed by some medium or reflected in other directions, so that the camera cannot receive the returned light, resulting in holes in the depth image. Reasonable filling of image holes is very important for subsequent calculation of the volume of the muck, and appropriate filling methods can be adopted according to the environmental characteristics of the construction site and the characteristics of the collected depth image holes.

[0078] Step 3.6 Finally, image segmentation is performed on the enhanced depth image to extract the muck edge, determine the pixel set belonging to the muck, and save the pixel position of the muck.

[0079] Step 4 Calculate the volume of the muck

[0080] Step 4.1 After identifying the muck target and saving the position information of all muck pixel points, the volume of the muck is calculated. For each frame of the difference image processed in step 3, the pixel value of all muck pixel 3 points, i.e., the muck height value H i (background depth value D0-muck depth value D i ), is extracted using the double integration idea, where i represents the i-th muck pixel point. The pixel value of each pixel point, i.e., the muck height value H i , multiplied by the area of a unit pixel S0, can obtain the volume of the muck represented by each pixel, and the volume of all muck pixel points is accumulated to obtain the volume of the muck contained in the frame image. The specific calculation formula is shown in equation (5) as follows.

[0081]

[0082] In the formula, V represents the volume of the muck in a frame of muck depth image, i represents the i-th muck pixel point, x i , y i represent the position coordinates of the i-th pixel point, M represents a total of M muck pixel points, v i represents the muck volume value of the i-th muck pixel point, and H i represents the muck height value of the i-th muck pixel point in the frame of muck depth image.

[0083] Step 4.2 After the shield advances for one ring or a period of time, the volume of the muck calculated according to all the collected muck depth images during the period is accumulated and divided by the number of muck statistics, to obtain the total volume of the muck discharged in the ring or in the period of time. The specific calculation formula is shown in equation (6) as follows.

[0084]

[0085] In the formula, Vol is the total volume of the discharged muck in a ring or in a certain period of time, j represents the jth muck depth image collected, n represents the total number of muck depth images collected in the ring or in the certain period of time, V j represents the muck volume calculated in the jth muck depth image, and K represents the muck counting times.

[0086] Advantages of the present application:

[0087] The present application provides a method for estimating the muck volume discharged in earth pressure balance shield construction based on computer vision technology. The method uses the muck depth images collected by a depth camera to identify the muck target and calculate the discharged muck volume, and has the characteristics of non-contact, low cost and convenient application.

[0088] The present application provides a method for estimating the muck volume in a dynamic muck discharge process based on depth images. The method includes how to use the double integral idea to estimate the muck volume of a single image and how to realize the volume accumulation calculation of the muck movement process by delimiting the detection area. The method has the characteristics of accuracy, speed and simple calculation.

[0089] The method for estimating the muck volume based on computer vision technology provided by the present application directly identifies the muck volume based on the muck depth image, without considering the construction parameters such as shield tunneling parameters and modifier addition amount for inference. Therefore, the method is not limited by the construction experience of a certain region or a certain project, and can be applied to any earth pressure balance shield tunneling project in any region, and has strong applicability.

[0090] The muck volume calculation method based on depth images provided by the present application mainly uses the double integral idea, and realizes the accumulation estimation of the dynamic flowing muck volume by delimiting the detection range. The whole calculation process does not involve complex formula derivation and calculation assumptions, and there is no calculation parameter that is difficult to determine or needs to rely on experience for value selection. The calculation process is simple and direct, fast and accurate.

[0091] The muck volume estimation method provided by the present application only needs to install a depth camera and an industrial computer for data processing and calculation at the construction site. Compared with other volume measurement devices or laser scanners, the installation is simple, the cost is low, and there is no need to directly contact the muck, which has no effect on the construction process. The reasoning speed of the calculation algorithm itself is also sufficient to keep up with the picture collection speed of the camera, so the method has the characteristics of real-time and low cost.

Claims

1. An earth pressure balance shield muck volume estimation method, characterized by, According to the depth image of the discharged soil on the belt conveyor, the discharged soil volume in the earth pressure balance tunneling process is estimated, comprising: Step 1, first install a depth camera above the belt conveyor, responsible for collecting the depth image of the discharged soil; Step 2, then in the stage of no discharged soil, calculate or determine the calculation parameters required in the calculation formula of the discharged soil volume, and demarcate the image detection area for discharged soil recognition and volume calculation; if the range of the image detection area is not suitable, the camera installation position needs to be adjusted, and the adjustment is repeated until the detection area is demarcated in a suitable discharged soil flowing area; Step 3, after the shield machine starts advancing, the depth camera collects the depth image in real time, and uses computer vision technology to recognize the discharged soil target in the image and calculate the discharged soil volume; Step 4, after each ring advancing is completed, the discharged soil volume calculated in real time during the advancing process is accumulated to obtain the final discharged soil volume of each ring; The step 3 specifically comprises: Step 3.1 when the shield starts advancing, the depth camera starts collecting the depth image after starting the discharged soil volume calculation program; Step 3.2 the program first performs the discharged soil movement detection operation on each collected depth image, that is, judges whether there is moving discharged soil in the current detection area; Step 3.3 each depth image needs to compare the depth values of the pixel points in the detection area with the previous depth image, if the sum of the depth difference values of the pixel points in the detection area of the adjacent two depth images is less than a certain threshold value, it is considered that the discharged soil is not moving at this time, and no subsequent target recognition and volume calculation operation is needed; otherwise, if the sum of the depth difference values of the pixel points in the detection area of the adjacent two depth images is greater than a certain threshold value, it is considered that the discharged soil is in the process of being discharged, and the subsequent operation is performed on the depth image; Step 3.4 The slag depth image detected by slag movement detection first saves its depth matrix D, and then calculates the difference D between the background depth matrix and the slag depth matrix b - D, get the difference image, highlight the slag part in the image; Step 3.5 the difference image is enhanced, the main purpose of which is image denoising and hole filling; Step 3.6 finally, the enhanced depth image is segmented to extract the discharged soil edge, determine the pixel set belonging to the discharged soil, and save the pixel position of the discharged soil; The step 4 specifically comprises: Step 4.1 After identifying the slag target and saving the position information of all slag pixels, the slag volume is calculated; for each processed difference image, the pixel value of all slag pixel positions, i.e. the slag height value H is extracted by using the double integration idea i , the slag height value H i = background depth value D0- slag depth value D i , where i represents the i-th slag pixel; the pixel value of each pixel, i.e. the slag height value H i , is multiplied by the area of a unit pixel S0 to obtain the volume of the slag represented by each pixel, and the volume of all slag pixels is accumulated to obtain the slag volume value contained in the image frame; Step 4.2 after the shield advances for a ring or a period of time, the discharged soil volume calculated according to all the collected discharged soil depth images during the accumulation is divided by the number of discharged soil statistics to obtain the total discharged soil volume discharged in the ring or the period of time.

2. The method of claim 1, wherein, The step 1: The depth camera is installed at the position directly above the belt conveyor, the optical axis of the camera is perpendicular to the surface of the belt conveyor, the center line of the camera is consistent with the center line of the belt conveyor as much as possible, the installation height needs to ensure that the horizontal viewing angle of the camera can cover the width of the belt conveyor, the vertical length of the detection area calculated according to the installation height should be within the vertical viewing angle range; the installation point of the depth camera should be kept away from the soil outlet of the screw conveyor to ensure that the discharged soil has been completely broken on the belt conveyor when it enters the image detection range, and the movement speed is consistent with the belt conveyor.

3. The method of claim 1, wherein, The step 2 specifically comprises: Step 2.1 before each ring of the shield cutting and advancing, the parameters are determined: First, the depth image of the surface of the belt conveyor is collected by using the depth camera, i.e. the background depth image, and the background depth matrix D is saved b According to the above, the distance D0 between the surface of the belt and the camera is directly determined, i.e. the background depth D0; then, the area S0 of the real world represented by each pixel in the depth image on the plane with the distance of the background depth D0 is calculated, and the parameters of the depth camera required in the calculation process include the horizontal view angle α, the vertical view angle β, the total number of horizontal pixels W, and the total number of vertical pixels L; according to the formulas (1) and (2), the horizontal width w0 and the vertical length l0 of the real world corresponding to each pixel in the image on the plane with the depth D0 are calculated respectively; Then, the area S0 of each pixel in the real world is obtained by multiplying the length and width of the pixel in the real world, and the calculation formula is shown as formula (3); （1） （2） （3） Step 2.2: defining the detection area and performing calculation: All image processing, slag target recognition and volume calculation are only considered in the detection area; The slag discharge is a continuous and dynamic process during the shield tunneling process. In order to ensure that no slag target is missed and no slag target is repeatedly calculated, a slag detection area needs to be defined, and only the slag entering the area is identified and calculated for volume; The horizontal width W of the detection area d Slightly greater than the width of the belt conveyor; the vertical length L of the detection area d The calculation formula is shown in the following formula (4) according to the movement speed v of the slag on the belt conveyor, the frame interval time t0 of the slag depth picture, and the statistical times K of the slag volume. Wherein, The movement speed v of the slag on the belt conveyor is consistent with the belt conveyor transmission speed v, which is directly obtained from the construction data recorded automatically by the shield machine; The slag volume statistical frequency K represents that the volume of each pixel of the slag is calculated K times in the process of passing through the detection area. The total slag volume in the detection area is obtained by summing up the volume of the slag calculated in each frame of picture in the ring and dividing by the slag volume statistical frequency K. The product of the slag movement speed and the frame interval time represents the distance of the slag moving forward between the two picture acquisition intervals, and the product of the distance and the statistical times K represents the forward distance of the slag in the process of K continuous frame acquisition, and the forward distance is divided by the vertical length l0 corresponding to each pixel in the real world to obtain the number of pixels in the vertical direction of the detection area, that is, the vertical length L d The detection area is selected by a person and is selected in the middle of the camera field of view （4） If the vertical length range of the image detection area calculated at the current installation height exceeds the visual angle range of the slag moving at the same speed on the belt conveyor, the camera installation position needs to be adjusted until the picture in the detection area is completely the picture of the slag discharged with the belt conveyor at the same speed.

4. The method of claim 1, wherein, The step 4.1 is specifically calculated according to the formula (5) as shown below; （5） In the formula, V represents the volume of the sludge in a frame of sludge depth image, i represents the i th sludge pixel point, x i , y i Represent the position coordinates of the i th pixel point, M represents a total of M sludge pixel points, v i Represent the sludge volume value of the i th sludge pixel point, H i Represent the sludge height value of the i th sludge pixel point in the frame of sludge depth image. The step 4.2 is specifically calculated according to the formula (6) as shown below; （6） In the formula, Vol is the total volume of the dredged spoil in a ring or in a certain period of time, j represents the jth dredged spoil depth image collected, n represents the total number of dredged spoil depth images collected in the ring or in the certain period of time, V j represents the calculated volume of the dredged spoil in the jth dredged spoil depth image, and K represents the number of times of dredged spoil statistics.

Images