A water station information data processing method and system

Abstract

The present invention discloses a water station information data processing method, which obtains a bucket stacking image, takes an area with equal spacing between buckets as a bucket stacking area; identifies the number of buckets on the top layer of a pallet; identifies the number of pallets; counts the total number of buckets according to the number of bucket placement areas, the number of pallets, the number of buckets on the top layer of a pallet, and the number of buckets placed directly on the ground; identifies the outermost buckets on the ground through image edge detection technology to quickly identify the increase or decrease of buckets; and establishes a water station storage management stereo model according to the total number of buckets and the placement of buckets. Through image processing technology and algorithms, the bucket placement area, the number of pallets, the number of buckets on the top layer of a pallet, and the number of buckets placed directly on the ground can be automatically identified, so as to quickly and accurately count the total number of buckets; the established stereo model can reflect the changes in the number and position of buckets in real time, so that management personnel can understand the inventory of the water station at any time.

Description

Technical Field

[0001] The present invention belongs to the technical field of image recognition, and in particular relates to a water station information data processing method and system. Background Art

[0002] With the rapid development of modern water stations, accurate management of water bucket inventory has become particularly important. Traditional inventory management methods mostly rely on manual counting, which is not only time-consuming and labor-intensive, but also prone to errors. Especially when faced with a large number of water buckets, the management difficulty is significantly increased. In addition, since the placement of water buckets is often dense and there may be multiple layers of stacking, manual counting is even more difficult; and the backstage management staff cannot understand the storage of water buckets in the warehouse in real time and intuitively. Therefore, a technical solution is designed to use image recognition technology to identify the number of water buckets and three-dimensional modeling of warehouse water buckets. Summary of the invention

[0003] The purpose of the present invention is to provide a water station information data processing method and system to solve the problems raised in the above background technology.

[0004] To achieve the above object, the present invention provides the following technical solution: a water station information data processing method, comprising the following steps:

[0005] 1) Get the bucket stacking image, convert the coordinate calculation points in the camera coordinate system into plane coordinates to calculate the spacing between buckets. The area with equal spacing between buckets is a bucket stacking area;

[0006] 2) Remove pixels other than the ends to obtain an image of the ends of unconnected buckets, and identify the number of buckets on the top layer of the tray by counting the unconnected areas;

[0007] 3) Identify the number of pallets;

[0008] 4) Calculate the total number of buckets based on the number of bucket placement areas, the number of pallets, the number of buckets on the top of the pallets, and the number of buckets placed directly on the ground;

[0009] 5) Use image edge detection technology to identify the outermost bucket on the ground to quickly identify the increase or decrease of buckets;

[0010] 6) Establish a three-dimensional model of water station storage management based on the total amount of water buckets and their placement.

[0011] Preferably, the specific method of step 1) is to obtain a binary image through image processing, and convert the image pixel points from the coordinate calculation points in the camera coordinate system into plane coordinates (X, Y), as follows:

[0012]

[0013]

[0014] Where f is the focal length of the camera, , , is the point coordinate in the camera coordinate system;

[0015] After converting the image from the camera's perspective into plane coordinates, the bucket spacing is calculated, and the areas with the same spacing between the bucket ends are identified as bucket placement areas, and the number of bucket placement areas is obtained.

[0016] Preferably, the specific method of step 2) is to remove the pixels other than the ends to obtain an unconnected bucket end image, and use the opencv algorithm to count the pixel values ​​of each non-adjacent white area after the coordinates are converted, so as to obtain the number of buckets on the top layer of the tray.

[0017] Preferably, the specific method of step 4) is to calculate the total number of buckets T based on the placement of pallets through the number of pallet placement areas n, the number of pallets in each area m, and the number of top buckets k identified in each area, using the following formula:

[0018]

[0019] Among them, j is the number of buckets filled on the tray, is the number of pallets in the nth area, is the number of top buckets in the nth region;

[0020] By using the camera to recognize and count the buckets placed directly on the ground, the total number of buckets can be calculated.

[0021] Preferably, the specific method of step 5) is to obtain the outermost side image contour, and determine whether the images overlap according to the image contour, and if they overlap, predict the blocked part.

[0022] Preferably, the prediction method of the occluded part is as follows: compare and identify all adjacent edge line combinations of the image to be detected with the images in the single bucket data set, identify the bucket image closest to the camera, traverse the image, identify the vertical lines, and determine the number of buckets in this direction with the number of identified vertical lines, so as to know the number of buckets on the outermost side.

[0023] Preferably, a complete bucket image is compensated based on the known identified vertical lines of the occluded bucket, combined with the proportions of the bucket sides and the shooting angle of the camera, and the compensated complete bucket image is compared with the bucket image at the same position in the single bucket data set to determine the accuracy of the compensated image and calibrate the number of bucket recognitions.

[0024] Preferably, the buckets that are increased or decreased in real time in 6) are synchronized in real time through the established three-dimensional model.

[0025] A water station information data processing system, comprising:

[0026] Area recognition module: used to obtain bucket stacking images, convert coordinate calculation points in the camera coordinate system into plane coordinates to calculate the spacing between buckets. The area with equal spacing between buckets is a bucket stacking area.

[0027] The upper bucket quantity recognition module is used to remove the pixels outside the ends to obtain the unconnected bucket end images, and identify the number of buckets on the uppermost layer of the tray by counting the unconnected areas;

[0028] Pallet quantity identification module: used to identify the number of pallets;

[0029] Bucket total quantity calculation module: used to calculate the total quantity of buckets based on the number of bucket placement areas, the number of pallets, the number of buckets on the top of the pallets, and the number of buckets placed directly on the ground;

[0030] Bucket increase and decrease detection module: used to identify the outermost bucket on the ground through image edge detection technology to quickly identify the increase and decrease of buckets;

[0031] Three-dimensional model building module: used to build a three-dimensional model of water station storage management according to the total amount of water buckets and the placement of water buckets.

[0032] The technical effects and advantages of the present invention are as follows: 1. Through image processing technology and algorithms, the placement area of ​​buckets, the number of pallets, the number of buckets on the top of the pallet, and the number of buckets placed directly on the ground can be automatically identified, so as to quickly and accurately count the total number of buckets; greatly improving the efficiency and accuracy of bucket inventory management;

[0033] Through image edge detection technology, prediction method of the occluded part, and image comparison and judgment after compensation, the number of occluded buckets can be accurately identified and counted, and only the number of the outermost buckets needs to be identified, which greatly reduces the system's computing intensity and improves the system's computing efficiency. It is especially suitable for operations of taking a small number of buckets multiple times.

[0034] The established 3D model can reflect the changes in the number and location of water buckets in real time, allowing managers to understand the inventory status of the water station at any time and increase or decrease the number of water buckets in a timely manner. This helps the water station achieve more flexible and efficient inventory management;

[0035] By integrating multiple modules such as area identification, upper bucket quantity identification, pallet quantity identification, bucket increase and decrease detection, and three-dimensional model establishment, a complete water station information data processing system is formed, which can automatically complete various tasks of bucket inventory management, reduce the degree of manual intervention, and improve the level of intelligent management. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 is a flow chart of the present invention;

[0037] Figure 2 It is a system module diagram of the present invention. DETAILED DESCRIPTION

[0038] The specific embodiments of the present invention are further described below in conjunction with the accompanying drawings. It should be noted that the description of these embodiments is used to help understand the present invention, but does not constitute a limitation of the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0039] The present invention provides Figure 1 A water station information data processing method shown includes the following steps:

[0040] Step 1: Establish a three-dimensional model of water station storage management, build a three-dimensional model of bottled water based on the storage plant, and display images to facilitate intuitive understanding of the storage status of bottled water in the background;

[0041] The storage plant is divided into multiple areas, and each area is numbered, and the buckets in each area are sorted in a stacking manner. For example, buckets are stacked on pallets, and the layers are marked in order from bottom to top. Each layer can hold four rows and four columns, totaling sixteen buckets. The buckets are then numbered in sequence, and each bucket of water can be accurately located using the area code-layer number-serial number method.

[0042] Step 2: Automatically identify the number of areas. Obtain the image of the bucket stacking through the overhead camera, convert the image into a grayscale image, perform binarization on the grayscale image, set the background color pixels to 0 (black), set the bucket and tray pixel values ​​to 255 (white), and calculate the spacing between the buckets. Identify the spacing between the fixed position points on the buckets through the image, and determine whether the spacing between the buckets is equal by calculating the pixel points of the fixed position points. Since the end of the bucket has obvious features, the end of the bucket is used as the fixed position point; here, considering the difference in image pixel points due to different focal lengths of the same camera shooting at different positions, the coordinate calculation points in the camera coordinate system are now converted into plane coordinates (X, Y), as follows:

[0043]

[0044]

[0045] Where f is the focal length of the camera, , , is the point coordinate in the camera coordinate system;

[0046] Based on the fact that the distance between buckets is fixed after they are placed on the pallet, the number of bucket placement areas is determined on this basis. Specifically, the image from the camera's perspective is converted into plane coordinates and then the bucket spacing is calculated. The area with the same spacing between the bucket ends is the identified bucket placement area, and the number of bucket placement areas can be known; the pixels other than the ends are removed to obtain unconnected bucket end images, and the opencv algorithm is used to count the pixel values ​​of each non-adjacent white area after the coordinate conversion, so as to obtain the number of buckets on the top layer of the pallet; at the same time, white areas with pixel values ​​below 30 are filtered to exclude areas with incorrect identification.

[0047] Step 3: Pallet quantity recognition: Create a pallet dataset, use a camera to obtain images of pallets at all positions within the camera's field of view, and store them in the pallet dataset;

[0048] Since the features of the pallet are obvious, the acquired image to be detected is compared with the image in the pallet dataset to identify the pallets and count them.

[0049] Step 4: Count the number of buckets placed on pallets. Calculate the total number of buckets placed on pallets T by identifying the number of pallet placement areas n, the number of pallets in each area m, and the number of top buckets in each area k. The formula is as follows:

[0050]

[0051] Wherein, j is the number of buckets placed on the pallet. The present invention uses a pallet that can hold 16 buckets. is the number of pallets in the nth area, is the number of top buckets in the nth region.

[0052] The fifth step is to calculate the total number of buckets. Since there are not only pallets for buckets in the warehouse, but also a certain number of buckets placed directly on the ground, this part should also be recorded when counting the number of buckets. The counting method for this part is the same as the counting of the number of buckets on the top layer of each area. It can be identified and counted by the top camera. The number of buckets in this part is represented by b, so the total number of buckets is:

[0053] a=T+b.

[0054] As an embodiment of the present invention, the present invention also provides a method for quickly calculating the increase or decrease of the number of buckets;

[0055] Since the buckets placed directly on the ground are placed from the inside to the outside according to convenience and common sense, when putting in the buckets, they are placed on the innermost side first, and then gradually outward. When taking out the buckets, the outermost buckets are taken first. Based on this, the outermost side image contour is first obtained, and the gradient operator is used to realize edge detection. Specifically, the set convolution template is moved in the image, and each pixel in the image is convolved with the template to obtain the response R of each pixel. R is used to represent the gray value change rate of each pixel, that is, the gray gradient value, so that the gray image can be converted into a gradient image after convolution with the template. The formula is as follows:

[0056]

[0057] Where Z represents the gray value of the pixel, is the template coefficient. Since the bucket image is in a single form, a 3*3 convolution template is used in this embodiment.

[0058] By setting the threshold, if the convolution result R is greater than the threshold, then the pixel is an edge point and the output is white; if R is less than the threshold, then the pixel is not an edge point and the output is black; finally, a black and white gradient image can be output to achieve edge detection.

[0059] And determine whether the images overlap based on the image contours. If they overlap, predict the occluded part. The specific method is as follows:

[0060] Establish a single bucket dataset: Use the camera to obtain single bucket images of buckets at all positions within the camera's field of view, and store them in different bucket datasets according to the camera angle;

[0061] All adjacent edge line combinations of the image to be detected are compared and identified with the images in the single bucket data set to identify similar pictures. Since the bucket image closest to the camera occludes the bucket image close to it, the remaining buckets are occluded in turn. When the bucket image closest to the camera is identified, the image is traversed to identify the vertical lines of the remaining blocked bucket images. The number of buckets in this direction is determined by the number of identified vertical lines, and the number of outermost buckets can be known. Since only the number of outermost buckets needs to be identified, the system's computing intensity is greatly reduced and the system's computing efficiency can be improved, which is especially suitable for operations of taking a small number of buckets multiple times.

[0062] Then the complete bucket image is compensated in proportion. Since the edge of one side of the bucket is completely identified, the complete bucket image can be compensated according to the proportion of each side of the bucket and the shooting angle of the camera. The compensated complete bucket image is compared with the bucket image at the same position in the single bucket data set to determine the accuracy of the compensated image, which is convenient for the calibration of the bucket recognition quantity.

[0063] As an embodiment of the present invention, the time when each bucket enters and leaves the warehouse is recorded, and a warehouse storage time alarm is set to alarm for buckets that are close to storage time to avoid excessive storage, which causes the water in the bucket to expire, making it easier for staff to deal with the problem.

[0064] like Figure 2 As shown, as an embodiment of the present invention, the present invention also provides a water station information data processing system, including:

[0065] Area recognition module: used to obtain bucket stacking images, convert coordinate calculation points in the camera coordinate system into plane coordinates to calculate the spacing between buckets. The area with equal spacing between buckets is a bucket stacking area.

[0066] The upper bucket quantity recognition module is used to remove the pixels outside the ends to obtain the unconnected bucket end images, and identify the number of buckets on the uppermost layer of the tray by counting the unconnected areas;

[0067] Pallet quantity identification module: used to identify the number of pallets;

[0068] Bucket total quantity calculation module: used to calculate the total quantity of buckets based on the number of bucket placement areas, the number of pallets, the number of buckets on the top of the pallets, and the number of buckets placed directly on the ground;

[0069] Bucket increase and decrease detection module: used to identify the outermost bucket on the ground through image edge detection technology to quickly identify the increase and decrease of buckets;

[0070] Three-dimensional model building module: used to build a three-dimensional model of water station storage management according to the total amount of water buckets and the placement of water buckets.

[0071] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present invention. A person skilled in the art may change, modify, replace and vary the above embodiments within the scope of the present invention.

Claims

1. A water station information data processing method, characterized in that: The steps include: 1) Get the bucket stacking image, convert the coordinate calculation points in the camera coordinate system into plane coordinates to calculate the spacing between buckets. The area with equal spacing between buckets is a bucket stacking area; 2) Remove pixels other than the ends to obtain an image of the ends of unconnected buckets, and identify the number of buckets on the top layer of the tray by counting the unconnected areas; 3) Identify the number of pallets; 4) Calculate the total number of buckets based on the number of bucket placement areas, the number of pallets, the number of buckets on the top of the pallets, and the number of buckets placed directly on the ground; 5) Use image edge detection technology to identify the outermost bucket on the ground to quickly identify the increase or decrease of the bucket; the specific method is to obtain the outermost side image contour, and determine whether the image overlaps based on the image contour. If it overlaps, predict the blocked part; The prediction method of the blocked part is as follows: compare and identify all adjacent edge line combinations of the image to be detected with the images in the single bucket data set, identify the bucket image closest to the camera, traverse the image, identify the vertical lines, and determine the number of buckets in this direction based on the number of identified vertical lines, so as to know the number of buckets on the outermost side; Based on the known vertical lines of the blocked bucket, the complete bucket image is compensated in combination with the ratio of each side of the bucket and the shooting angle of the camera. The compensated complete bucket image is compared with the bucket image at the same position in the single bucket dataset to determine the accuracy of the compensated image and calibrate the number of bucket recognitions. 6) Establish a three-dimensional model of water station storage management based on the total amount of water buckets and their placement.

2. A water station information data processing method according to claim 1, characterized in that: The specific method of step 1) is to obtain a binary image through image processing, and convert the image pixel points from the coordinate calculation points in the camera coordinate system into plane coordinates (X, Y), as follows: ； ； Where f is the focal length of the camera, , , is the point coordinate in the camera coordinate system; After converting the image from the camera's perspective into plane coordinates, the bucket spacing is calculated, and the areas with the same spacing between the bucket ends are identified as bucket placement areas, and the number of bucket placement areas is obtained.

3. A water station information data processing method according to claim 2, characterized in that: The specific method of step 2) is to remove the pixels other than the ends to obtain an unconnected bucket end image, and use the opencv algorithm to count the pixel values ​​of each non-adjacent white area after the coordinates are converted, so as to obtain the number of buckets on the top layer of the tray.

4. A water station information data processing method according to claim 1, characterized in that: The specific method of step 4) is to calculate the total number of buckets T based on the placement of pallets through the number of pallet placement areas n, the number of pallets in each area m, and the number of top buckets identified in each area k, and the formula is as follows: ； Among them, j is the number of buckets filled on the tray, is the number of pallets in the nth area, is the number of top buckets in the nth region; By using the camera to recognize and count the buckets placed directly on the ground, the total number of buckets can be calculated.

5. A water station information data processing method according to claim 1, characterized in that: The buckets that are increased or decreased in real time in step 6) are synchronized in real time through the established three-dimensional model.

6. A water station information data processing system, characterized in that: include: Area recognition module: used to obtain bucket stacking images, convert coordinate calculation points in the camera coordinate system into plane coordinates to calculate the spacing between buckets. The area with equal spacing between buckets is a bucket stacking area. The upper bucket quantity recognition module is used to remove the pixels outside the ends to obtain the unconnected bucket end images, and identify the number of buckets on the uppermost layer of the tray by counting the unconnected areas; Pallet quantity identification module: used to identify the number of pallets; Bucket total quantity calculation module: used to calculate the total quantity of buckets based on the number of bucket placement areas, the number of pallets, the number of buckets on the top of the pallets, and the number of buckets placed directly on the ground; Bucket increase and decrease detection module: used to identify the outermost bucket on the ground through image edge detection technology to quickly identify the increase and decrease of buckets; the specific method is to obtain the outermost side image contour, and determine whether the image overlaps based on the image contour, and if it overlaps, predict the blocked part; The prediction method of the blocked part is as follows: compare and identify all adjacent edge line combinations of the image to be detected with the images in the single bucket data set, identify the bucket image closest to the camera, traverse the image, identify the vertical lines, and determine the number of buckets in this direction based on the number of identified vertical lines, so as to know the number of buckets on the outermost side; Based on the known vertical lines of the blocked bucket, the complete bucket image is compensated in combination with the ratio of each side of the bucket and the shooting angle of the camera. The compensated complete bucket image is compared with the bucket image at the same position in the single bucket dataset to determine the accuracy of the compensated image and calibrate the number of bucket recognitions. Three-dimensional model building module: used to build a three-dimensional model of water station storage management according to the total amount of water buckets and the placement of water buckets.

Images