Thermal-state casting blank image splicing method

Abstract

The invention discloses a thermal-state casting blank image splicing method, which comprises the following steps that: a middle camera and an edge camera are triggered at the same time according to a preset time interval, and a middle image and at least one edge image of a casting blank at the moment are respectively collected; determining a transformation matrix from the image coordinates of all the side cameras to the image coordinates of the middle camera, and converting all the side images collected at the same moment to a middle image coordinate system at the corresponding moment based on the transformation matrix; the method comprises the following steps: determining edge images having an overlapping region with a middle image, searching feature points in the overlapping region to complete image splicing, determining edge images having an overlapping region with a spliced image in the remaining edge images, and splicing all the edge images collected at the same moment to the middle image at the corresponding moment to complete image splicing. Obtaining a horizontal spliced image at the moment; and vertically splicing the horizontal splicing images at adjacent moments to obtain a final casting blank splicing image. According to the method, rapid and accurate splicing of the hot-state casting blank images is realized.

Description

Technical Field

[0001] The present invention relates to the field of automated detection in iron and steel metallurgy, and in particular to a method for splicing images of hot cast billets. Background Art

[0002] Continuous casting is an intermediate step in steel production, preceding steelmaking and following hot rolling. During the continuous casting process, various surface defects (common ones include cracks, scratches, and holes) can appear on the ingots due to factors such as raw materials, equipment, and production processes. Failure to promptly detect and address these defects can lead to widespread quality defects during the hot rolling process, resulting in unnecessary economic losses. Therefore, inspecting the surface quality of the ingots and implementing appropriate measures to address them can effectively ensure the surface quality of the hot-delivered ingots, thereby improving the quality of the end product.

[0003] Ingot image acquisition is a prerequisite for ingot surface defect detection. The field of view of a single industrial camera is limited. To ensure the quality of ingot surface imaging, multiple cameras are arranged side by side along the width of the ingot to ensure ingot image resolution. Due to certain errors in camera installation, it cannot be guaranteed that each camera is horizontal and parallel, and there are certain angles and displacements. The effect of simple image coordinate stitching is not good, and defect identification does not require a complete stitched image. Therefore, the existing ingot surface inspection system does not provide a stitched image of the entire ingot surface and cannot directly display the entire surface image of the ingot. As a result, on-site technicians cannot fully evaluate the surface condition of the entire ingot, and it is not convenient for defect tracing and analysis.

[0004] However, the image acquisition of hot ingots is easily affected by parameters such as light, environment, and camera installation angle, which can easily lead to poor feature point matching accuracy and long time consumption. The stitched images may have problems such as distortion, seams, and brightness differences, which affect the final stitching effect. Summary of the Invention

[0005] In order to solve the problems of poor splicing effect of simple image coordinates, poor feature point matching accuracy and long time consumption, the present invention provides a hot cast billet image splicing method to achieve automatic and rapid splicing of hot cast billet images in horizontal and vertical directions.

[0006] In a first aspect of the present invention, a method for splicing images of hot cast billets is provided, the method comprising: A central camera and at least one edge camera are set along the width direction of the billet, and any two adjacent cameras have overlapping fields of view; When the billet passes through the camera, the middle camera and the side camera are triggered simultaneously at a preset time interval to respectively capture a middle image and at least one side image of the billet at that moment; Determine the transformation matrix from the coordinates of all edge camera images to the coordinates of the central camera image, and based on the transformation matrix, transform all edge images collected at the same time into the central image coordinate system at the corresponding time; Determine the edge images that overlap with the central image, search for feature points in the overlapping areas to complete image stitching, and then determine the edge images that overlap with the stitched image in the remaining edge images. In this way, all edge images collected at the same time are stitched onto the central image at the corresponding time to obtain the horizontal stitching image at that time. The horizontal stitching images at adjacent moments are vertically stitched together to obtain the final stitching image of the ingot.

[0007] In the above scheme, a middle camera and at least one edge camera are arranged above the billet conveyor roller along the width direction of the billet, or a middle camera and at least one edge camera are arranged below the billet conveyor roller along the width direction of the billet, or a middle camera and at least one edge camera are arranged above and below the billet conveyor roller along the width direction of the billet.

[0008] In the above solution, when the billet passes through the camera, the middle camera and the side camera are triggered simultaneously at a preset time interval to respectively capture a middle image and at least one side image of the billet at that moment, including: When it is detected that the billet passes the camera, the middle camera and the side camera are triggered at the same time to respectively capture the middle image and at least one side image of the billet at that moment, and according to the billet number and camera position, the billet upper surface image is saved to the billet number + upper surface + camera position storage path, and the lower surface image is saved to the billet number + lower surface + camera position storage path, and the images captured by each camera are numbered according to the triggering order; The preset time interval for camera triggering is dynamically determined based on the speed of the conveyor roller to ensure that images captured by the same camera at adjacent moments have overlapping areas; When it is detected that the billet leaves the camera imaging area, the billet imaging is stopped.

[0009] In the above scheme, the transformation matrix from the image coordinates of all edge cameras to the image coordinates of the central camera is determined by the calibration plate, and recalibration is performed regularly or when the horizontal stitching effect of the image is poor.

[0010] In the above scheme, determining the edge images that overlap with the central image and searching for feature points in the overlapping area to complete image stitching includes: Determine all the side images that have overlapping areas with the central image, and use the image registration algorithm to find key points in the overlapping areas. Then use the feature matching algorithm to find strongly correlated key points as the feature points of the two images. Generate a transformation matrix based on the feature points, and perform perspective transformation to obtain a stitched image, so that all the side images that have overlapping areas with the central image are stitched onto the central image.

[0011] In the above solution, the obtained horizontally stitched images are subjected to image fusion processing, and invalid areas are cropped.

[0012] In the above scheme, an image fusion algorithm is used to smooth the horizontally stitched image, and a rectangle is used to cut off the black edges of the stitched image, retaining the image in the center area.

[0013] In the above solution, the horizontal stitching images at adjacent moments are vertically stitched to obtain the final stitching image of the ingot, including: If the number of horizontal image stitching is equal to 2, then extract the feature points of the first horizontal stitching image and the second horizontal stitching image, stitch the first horizontal stitching image and the second horizontal stitching image to obtain a stitching image in the vertical direction, and perform image fusion processing on the image; If the number of horizontal image stitching is greater than 2, extract the feature points of the last vertical stitching image and the current horizontal stitching image, stitch the last vertical stitching image and the current horizontal stitching image to obtain a new vertical stitching image, and perform image fusion processing on the image; The invalid area of ​​the final vertical stitching image is cut off to obtain the final ingot stitching image.

[0014] According to a second aspect of the present invention, a computer device is provided, comprising: a processor and a memory, wherein the memory stores programs or instructions that can be run on the processor, and when the programs or instructions are executed by the processor, the steps of the hot cast slab image stitching method described in any one of the first aspects are implemented.

[0015] According to a third aspect of the present invention, a computer-readable storage medium is provided, on which a program or instruction is stored. When the program or instruction is executed by a processor, the steps of the hot cast billet image stitching method described in any one of the first aspects are implemented.

[0016] In general, the above technical solutions conceived by the present invention can achieve the following beneficial effects compared with the prior art: The present invention provides a method for stitching images of hot ingots. Through feature point registration, horizontal stitching is first performed to obtain a panoramic image of the ingot width in the same frame. The horizontally stitched images are then vertically stitched together to obtain a complete image of the ingot surface, thereby achieving intelligent stitching of images of the upper and lower surfaces of the hot ingot. Furthermore, by transforming the images of the left and right edges of the ingot to the coordinates of the intermediate image and then performing feature point registration in the overlapping area, the present invention can reduce feature registration time and increase stitching speed. The stitched panoramic image of the ingot surface can be used by on-site technicians to comprehensively assess the surface condition of the entire ingot, providing strong support for tracing and analyzing surface quality defects of the ingot. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A schematic flow chart of a hot-cast slab image stitching method provided in an embodiment of the present application; Figure 2 A schematic diagram of synchronous collection of hot castings provided in an embodiment of the present application; Figure 3 A schematic diagram of image coordinate transformation provided in an embodiment of the present application; Figure 4 A schematic diagram of horizontal splicing of the left and middle images provided in an embodiment of the present application; Figure 5 A schematic diagram of horizontal splicing of a left-center spliced ​​image and a right image provided in an embodiment of the present application; Figure 6 A schematic diagram of cutting out an invalid area provided in an embodiment of the present application; Figure 7 A schematic diagram of the final horizontal splicing after fusion and cutting provided in an embodiment of the present application; Figure 8 A schematic diagram of an image of a whole hot cast billet after vertical splicing of a horizontal splicing image of a hot cast billet provided in an embodiment of the present application; Figure 9 A schematic diagram of the hardware structure of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. Based on the embodiments provided in this application, all other embodiments obtained by those of ordinary skill in the art without making creative work are within the scope of protection of the present invention.

[0019] Obviously, the drawings described below are merely examples or embodiments of the present application. Those skilled in the art can, without inventive effort, apply the present application to other similar scenarios based on these drawings. Furthermore, it is also understood that, although the effort involved in such a development process may be complex and lengthy, for those skilled in the art related to the content disclosed in this application, changes in design, manufacturing, or production based on the technical content disclosed in this application are merely conventional technical means and should not be construed as an insufficiency of the content disclosed in this application.

[0020] References to "embodiments" in this application mean that a particular feature, structure, or characteristic described in connection with the embodiment may be included in at least one embodiment of the application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it refer to independent or alternative embodiments that are mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described in this application may be combined with other embodiments unless there is a conflict.

[0021] Unless otherwise defined, the technical or scientific terms used in this application should have the ordinary meaning understood by a person of ordinary skill in the technical field to which this application belongs. The words "one", "a", "the" and the like used in this application do not indicate a limit on quantity and may indicate the singular or plural. The terms "include", "comprise", "have" and any variations thereof used in this application are intended to cover non-exclusive inclusions; for example, a process, method, system, product or device that includes a series of steps or modules (units) is not limited to the listed steps or units, but may also include steps or units that are not listed, or may also include other steps or units that are inherent to these processes, methods, products or devices. The words "connect", "connected", "coupled" and the like used in this application are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The word "multiple" used in this application refers to two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships can exist. For example, "A and / or B" can mean: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the objects before and after are in an "or" relationship. The terms "first", "second", "third", etc. involved in this application are only used to distinguish similar objects and do not represent a specific order for the objects.

[0022] This application provides a method for stitching images of hot ingots. This method performs feature point registration in overlapping areas, first stitching horizontally to obtain a panoramic image of the ingot width in the same frame, and then stitching the horizontally stitched images vertically to obtain a complete image of the ingot surface, thus achieving intelligent stitching of images of the upper and lower surfaces of the hot ingot.

[0023] like Figure 1 As shown, the hot-cast slab image splicing method of the present application comprises the following steps: Upon receiving the signal to start image acquisition of the slab, multiple cameras collect and store the surface images of the hot slab in real time; If the ingot image acquisition end signal is not received: read the same frame image, convert the edge camera image coordinates to the middle camera image coordinates; extract the feature points of the left and middle images, and splice the left and middle images; extract the feature points of the left middle image and the right image, splice the left middle image and the right image to obtain a spliced ​​image in the horizontal direction, perform image fusion processing on the image, and cut out the invalid area; if the number of horizontal splicing images is equal to 2, extract the feature points of the first horizontal splicing image and the second horizontal splicing image, splice the first horizontal splicing image and the second horizontal splicing image to obtain a spliced ​​image in the vertical direction, and perform image fusion processing on the image; if the number of horizontal splicing images is greater than 2, extract the feature points of the last vertical splicing image and the current horizontal splicing image, splice the last vertical splicing image and the current horizontal splicing image to obtain a new splicing image in the vertical direction, and perform image fusion processing on the image; If a signal for ending the slab image acquisition is received: there is a vertical stitching image, the invalid area of ​​the vertical stitching image is cut off to obtain the final slab stitching image.

[0024] In some embodiments, multiple cameras collect and store hot slab surface images in real time, including: Three industrial cameras are installed above and below the billet conveyor roller along the width of the billet, and any two adjacent cameras have overlapping fields of view; When it is detected that the ingot passes the camera, the six cameras on the upper and lower surfaces are triggered at the same time, and images are dynamically captured based on the roller speed. According to the ingot number and camera position, the images of the three cameras on the upper surface are saved to the storage path of ingot number + upper surface + camera position (left, middle, right), and the images of the three cameras on the lower surface are saved to the storage path of ingot number + lower surface + camera position (left, middle, right), and the images captured by each camera are numbered according to the triggering order (images captured at the same time have the same number).

[0025] If there are multiple side cameras, they are numbered as left 1, left 2, right 1, right 2, etc. according to the camera position. Of course, cameras can also be set only above or below the conveying roller as needed.

[0026] In some embodiments, the slab imaging end signal comes from the slab image acquisition system and is used to detect that the slab has left the camera imaging area.

[0027] In some embodiments, reading the same frame of image and converting the edge camera image coordinates to the center camera image coordinates includes: If the read billet number exists, read the left, middle and right numbers of the upper surface at the same time according to the order of drawing. i Secondary image acquisition Imgul i 、 Imgum i 、 Imgur i and the left, middle and right sides of the lower surface i Secondary image acquisition Imgdl i 、 Imgdm i 、 Imgdr i ; After the upper surface camera is installed, the transformation matrix from the image coordinates of the left camera on the upper surface to the middle camera is obtained through the calibration plate. M ulm And the transformation matrix of the camera image coordinates on the right side of the upper surface to the middle camera M urm ; After the lower surface camera is installed, the transformation matrix of the image coordinates of the left camera on the lower surface to the middle camera is obtained through the calibration plate. M dlm And the transformation matrix of the camera image coordinates on the right side of the lower surface to the middle camera M drm ; Through the transformation matrix M ulm and M urm , respectively obtain the left and right i Secondary image acquisition Imgul i 、 Imgur i The corresponding image on the image captured by the camera in the middle of the upper surface Imgult i 、 Imgurt i ; Through the transformation matrix M dlm and M drm , respectively obtain the left and right i Secondary image acquisition Imgdl i 、 Imgdr iThe corresponding image on the image collected by the camera in the middle of the lower surface Imgdlt i 、 Imgdrt i ; Upper surface transformation matrix M ulm 、 M urm and the lower surface transformation matrix M dlm 、 M drm ,When the regular or subsequent splicing effect is not good, it can be re-obtained through the calibration board.

[0028] In some embodiments, extracting feature points of the left and center images and stitching the left and center images includes: Find the upper surface left image through image registration algorithm Imgult i With China Library Imgum i The key points between the two images are selected using a feature matching algorithm to find the key points with strong correlation as the feature points of the two images. The transformation matrix is ​​generated based on the feature points, and the perspective transformation is performed to obtain the spliced ​​image of the left and middle images of the upper surface. Imgulm i ; Find the lower surface left image through image registration algorithm Imgdlt i With China Library Imgdm i The key points between the two images are selected using a feature matching algorithm to find the key points with strong correlation as the feature points of the two images. The transformation matrix is ​​generated based on the feature points, and the perspective transformation is performed to obtain the spliced ​​image of the left and middle images of the lower surface. Imgdlm i .

[0029] It should be noted that if there are multiple edge cameras, the edge images that have overlapping areas with the central image are first determined, and feature points are searched in the overlapping areas to complete image stitching. Then, the edge images that have overlapping areas with the stitched image are determined in the remaining edge images. In this way, all edge images collected at the same moment are stitched onto the central image at the corresponding moment to obtain the horizontal stitched image at that moment.

[0030] For example, there are two cameras on the left side of the middle camera, namely left 1 and left 2. Left 1 has an overlapping field of view with the middle camera, and left 2 has an overlapping field of view with left 1. The stitching of left 1 and the middle image is completed first, and then the stitching of left 2 and the stitched image of left 1 and the middle image is completed.

[0031] In some embodiments, feature points of the left-center image and the right image are extracted, the left-center image and the right image are spliced ​​together to obtain a spliced ​​image in the horizontal direction, image fusion processing is performed on the image, and invalid areas are cropped, including: Find the upper left middle image through image registration algorithm Imgulm i With the right picture Imgurt i The key points between the two images are found by feature matching algorithm, and the key points with strong correlation are used as the feature points of the two images. The transformation matrix is ​​generated based on the feature points, and the perspective transformation is performed to obtain the left middle image of the upper surface. Imgulm i With the right picture Imgurt i Stitched image Imguv i ; Find the lower left middle image through image registration algorithm Imgdlm i With the right picture Imgdrt i The key points between the two images are found by feature matching algorithm, and the key points with strong correlation are used as the feature points of the two images. The transformation matrix is ​​generated based on the feature points, and the perspective transformation is performed to obtain the left middle image of the lower surface. Imgdlm i With the right picture Imgdrt i Stitched image Imgdv i ; Image fusion algorithm is used to complete the horizontal stitching of left, middle and right images of the upper and lower surfaces Imguv i 、 Imgdv i Smoothing processing to improve the look and feel of stitched images; Due to the installation error of multiple cameras, the stitched image Imguv i 、 Imgdv i There are black edges. Through image processing technology, the black edges of the stitched image are rectangularly cropped to retain the image in the center area.

[0032] In some embodiments, extracting feature points of the first horizontal stitching image and the second horizontal stitching image, stitching the first horizontal stitching image and the second horizontal stitching image to obtain a stitching image in the vertical direction, and performing image fusion processing on the image includes: Find the first horizontal stitching image of the upper surface through the image registration algorithm Imguv 1 and the second horizontal stitching image ImguvThe key points between the two images are found by feature matching algorithm. The key points with strong correlation are used as the feature points of the two images. The transformation matrix is ​​generated based on the feature points. The perspective transformation is performed to obtain the stitching image of the first horizontal stitching image and the second horizontal stitching image of the upper surface, that is, the first vertical stitching image of the upper surface. Imguh 1; Find the first horizontal stitching image of the lower surface through the image registration algorithm Imgdv 1 and the second horizontal stitching image Imgdv The key points between the two images are found by feature matching algorithm. The key points with strong correlation are used as the feature points of the two images. The transformation matrix is ​​generated based on the feature points. The perspective transformation is performed to obtain the stitching image of the first horizontal stitching image and the second horizontal stitching image of the lower surface, that is, the first vertical stitching image of the lower surface. Imgdh 1; Use image fusion algorithm to complete the first vertical splicing of upper and lower surfaces Imguh 1. Imgdh 1 smoothing process to improve the look and feel of the stitched image.

[0033] In some embodiments, extracting feature points of the last vertical stitched image and the current horizontal stitched image, stitching the last vertical stitched image and the current horizontal stitched image to obtain a new vertical stitched image, and performing image fusion processing on the image includes: Find the last vertical stitching on the upper surface through image registration algorithm Imguh i-1 Stitch the image with the current level Imguv i The key points between the two images are found by feature matching algorithm, and the key points with strong correlation are used as the feature points of the two images. The transformation matrix is ​​generated based on the feature points, and the perspective transformation is performed to obtain the vertical splicing of the upper surface. Imguh i-1 Stitch the image with the current level Imguv i The stitched image of the upper surface i Second vertical stitching image Imguh i ; Find the next vertical stitching of the lower surface through image registration algorithm Imgdh i-1 Stitch the image with the current level Imgdv i The key points between the two images are selected by feature matching algorithm to find the key points with strong correlation as the feature points of the two images. The transformation matrix is ​​generated based on the feature points, and the perspective transformation is performed to obtain the vertical splicing of the lower surface. Imgdh i-1 Stitch the image with the current level Imgdv i The stitched image of the lower surfacei Second vertical stitching image Imgdh i ; The image fusion algorithm is used to complete the upper and lower surfaces i Second vertical stitching image Imguh i 、 Imgdh i Smoothing processing improves the appearance of stitched images.

[0034] In some embodiments, there is a vertical stitching image, and cutting out the invalid area of ​​the vertical stitching image to obtain the final stitching image of the casting strand includes: After the image is collected, the upper and lower surfaces are cut out and the image is spliced ​​together using image processing technology. Imguh i 、 Imgdh i Black border, keep the center area image.

[0035] In summary, after the casting camera is installed, the present application obtains the transformation matrix from the left camera image coordinates to the middle camera and the transformation matrix from the right camera image coordinates to the middle camera through the calibration plate, and then obtains the corresponding images of the left and right captured images under the capture coordinates of the middle camera through the transformation matrix, thereby compensating for the coordinate deviation between the side camera and the center camera, and improving the efficiency and quality of feature point extraction in the overlapping area between the subsequent side camera and the center camera. If the splicing effect is not good regularly or subsequently, the side camera transformation matrix is ​​re-obtained through the calibration plate to ensure the relative coordinate accuracy of the side camera image and the center camera image.

[0036] In addition, multiple intermediate stitching methods are used for horizontal stitching. The middle camera image is always used as the stitching target image. The edge transformation image is stitched with the intermediate image. The feature points of the overlapping area of ​​the edge transformation image and the intermediate image are found, and the edge image is stitched into the coordinates of the intermediate image. The stitched image is smoothed by image fusion, and the black edges generated by stitching can be automatically removed to improve the stitching quality and effect.

[0037] Specifically, the hot casting image splicing method of the embodiment of the present application is as follows Figure 1 As shown, the main steps are as follows: At step 101, if a signal for starting the slab drawing is received, the program flow goes to step 102, otherwise the signal for starting the slab drawing is waited for.

[0038] At step 102 , multiple cameras collect and store hot-state cast billet surface images in real time.

[0039] At step 103 , if a signal to stop the casting pattern is received, the program flow goes to step 201 , otherwise the program flow goes to step 104 .

[0040] At step 104 , the same frame of image is read, and the side camera image coordinates are converted to the middle camera image coordinates.

[0041] At step 105 , feature points of the left and center images are extracted, and the left and center images are spliced.

[0042] At step 106 , feature points of the left center image and the right image are extracted, and the left center image and the right image are spliced ​​together to obtain a spliced ​​image in the horizontal direction, and image fusion processing is performed on the image.

[0043] At step 107 , if the number of horizontally stitched images is greater than 1, the program flow goes to step 108 ; otherwise, the program flow goes to step 103 .

[0044] At step 108 , if the number of times of horizontally stitching images is equal to 2, the program flow goes to step 110 ; otherwise, the program flow goes to step 109 .

[0045] At step 109 , feature points of the last vertical stitched image and the current horizontal stitched image are extracted, the last vertical stitched image and the current horizontal stitched image are stitched together to obtain a new vertical stitched image, image fusion processing is performed on the image, and the program flow goes to step 103 .

[0046] At step 110 , feature points of the first horizontal stitching image and the second horizontal stitching image are extracted, the first horizontal stitching image and the second horizontal stitching image are stitched together to obtain a stitching image in the vertical direction, image fusion processing is performed on the image, and the program flow goes to step 103 .

[0047] At step 201 , if a vertically stitched image exists, the program flow goes to step 202 ; otherwise, the program flow goes to step 103 .

[0048] At step 202 , the invalid area of ​​the vertical stitching image is cut off to obtain the final stitching image of the ingot, and the program flow goes to step 103 .

[0049] Reference Figures 2 to 9 Taking the above surface image as an example, a hot casting slab image splicing method provided in an embodiment of the present application is described: Figure 2 Schematic diagram of synchronous acquisition of hot cast billets in the same frame on the upper surface provided in the embodiment of this application Imgul i 、 Imgum i 、 Imgur i , any two adjacent images have overlapping areas, such as Figure 2 In the middle dotted area, there are four reference points, which are used to indicate the differences in image coordinates between adjacent images.

[0050] After the upper surface camera is installed, the calibration board is used to obtain the transformation matrix from the upper surface left camera image coordinates to the middle camera using halcon create_hom_mat2d_proj or OpenCVgetPerspectiveTransform M ulm And the transformation matrix of the camera image coordinates on the right side of the upper surface to the middle camera M urm ; Through the transformation matrix M ulm and M urm , use halconaffine_trans_image or OpenCV warpPerspective function to transform the left and right sides of the upper surface i Secondary image acquisition Imgul i 、 Imgur i Transformed image converted to the coordinate system of the camera image captured in the middle of the upper surface Imgult i 、 Imgurt i , its schematic diagram is as follows Figure 3 shown.

[0051] Find the upper surface left image through OpenCV SIFT, SURF or ORB function Imgult i With China Library Imgum i The key points between the two images are found by using the BFMatcher or FlannBasedMatcher feature matching algorithm to find the strongly correlated key points as the feature points of the two images. The findHomography function is used to generate the transformation matrix based on the feature points. Finally, the warpPerspective function is used to perform perspective transformation to obtain the spliced ​​image of the left and middle images on the upper surface. Imgulm i ; You can also use the halconpoints_foerstner function to find the upper surface left image Imgult i With China Library Imgum i The proj_match_points_ransac feature matching algorithm is used to find the key points with strong correlation and obtain the transformation matrix. Finally, the gen_projective_mosaic function is used to obtain the spliced ​​image of the left and middle images of the upper surface. Imgulm i . Figure 4 It is a schematic diagram of the horizontal splicing of the left picture and the middle picture.

[0052] As above, extract the feature points of the left middle image and the right image, and splice the left middle image and the right image. The schematic diagram of the spliced ​​image is as follows: Figure 5 . Weighted, pyramid, gradient domain and other fusion algorithms can be used to complete the horizontal stitching of the left, middle and right images on the upper surface. Imguv i Smoothing processing improves the appearance of stitched images.

[0053] Due to the deviation of the field of view of multiple cameras, the stitched image Imguv i If there are black edges, we can use threshold segmentation to get the upper, lower, left, and right edge coordinates of the black edges, and then cut out the black edges of the spliced ​​image with a rectangle, such as Figure 6 As shown, the center area image is retained and the final horizontal stitching image is Imguv i like Figure 7 shown.

[0054] Reference horizontal stitching, Figure 8 This is a schematic diagram of vertical stitching. Use the image registration algorithm to find the last vertical stitching on the upper surface. Imguh i-1 Stitch the image with the current level Imguh i The key points between the two images are selected by feature matching algorithm to find the key points with strong correlation as the feature points of the two images. The transformation matrix is ​​generated based on the feature points, and the perspective transformation is performed to obtain the vertical splicing image of the upper surface. Imguh i-1 Stitch the image with the current level Imguv i The stitched image of the upper surface i Second vertical stitching image Imguh i ; Use image fusion algorithm to complete the upper surface i Second vertical stitching image Imguh i Smoothing is performed to improve the appearance of the stitched image; finally, the invalid area of ​​the vertical stitched image is cut off to obtain the final ingot stitched image.

[0055] It should be noted that the steps shown in the above process or the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0056] In addition, combined Figure 1 The hot-cast slab image stitching method described in the embodiment of the present application can be implemented by a computer device. Figure 9 Schematic diagram of the hardware structure of the computer device of the embodiment of the present application. Figure 9As shown, the device may include a processor 301 and a memory 302 storing computer program instructions.

[0057] Specifically, the processor 301 may include a central processing unit (CPU), or an application-specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiments of the present application.

[0058] Memory 302 may include a large-capacity memory for data or instructions. By way of example, and not limitation, memory 302 may include a hard disk drive (HDD), a floppy disk drive, a solid-state drive (SSD), flash memory, an optical disk, a magneto-optical disk, magnetic tape, or a Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, memory 302 may include removable or non-removable (or fixed) media. Where appropriate, memory 302 may be internal or external to the data processing device. In certain embodiments, memory 302 is non-volatile memory. In certain embodiments, memory 302 includes read-only memory (ROM) and random access memory (RAM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), an electrically alterable ROM (EAROM) or a flash memory (FLASH), or a combination of two or more of these. Under appropriate circumstances, the RAM can be a static random access memory (SRAM) or a dynamic random access memory (DRAM), where the DRAM can be a fast page mode dynamic random access memory (FPMDRAM), an extended data out dynamic random access memory (EDODRAM), a synchronous dynamic random access memory (SDRAM), etc.

[0059] The memory 302 may be used to store or cache various data files that need to be processed and / or used for communication, as well as possible computer program instructions executed by the processor 301 .

[0060] The processor 301 reads and executes computer program instructions stored in the memory 302 to implement any one of the hot cast slab image stitching methods in the above embodiments.

[0061] In some embodiments, the computer device for implementing the hot casting image stitching method described in any one of the first aspects may further include a communication interface 303 and a bus 300. Figure 9 As shown, the processor 301 , the memory 302 , and the communication interface 303 are connected via a bus 300 and communicate with each other.

[0062] The communication interface 303 is used to implement communication between the various modules, devices, units, and / or devices in the embodiments of the present application. The communication interface 303 can also implement data communication with other components such as: external devices, image / data acquisition equipment, databases, external storage, and image / data processing workstations.

[0063] Bus 300 includes hardware, software, or both, and couples components of a computer device to each other. Bus 300 includes, but is not limited to, at least one of the following: a data bus, an address bus, a control bus, an expansion bus, and a local bus. By way of example, and not limitation, bus 300 may include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a Hyper Transport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an InfiniBand interconnect, a Low Pin Count (LPC) bus, a memory bus, a Micro Channel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local Bus (VLB) bus, or other suitable buses, or a combination of two or more of these. Bus 300 may include one or more buses, where appropriate. Although embodiments herein describe and illustrate a particular bus, this application contemplates any suitable bus or interconnect.

[0064] The computer device can execute the hot casting billet image splicing method in the embodiment of the present application, thereby realizing the combination Figure 1 A hot-cast slab image stitching method is described.

[0065] In addition, in conjunction with the hot slab image stitching method in the above-mentioned embodiment, the present application embodiment may provide a computer-readable storage medium for implementation. The computer-readable storage medium stores computer program instructions; when the computer program instructions are executed by a processor, any of the hot slab image stitching methods in the above-mentioned embodiment is implemented.

[0066] It should be noted that the various technical features of the above-described embodiments can be combined in any manner. To simplify the description, not all possible combinations of the various technical features in the above-described embodiments are described. However, as long as there are no contradictions in the combination of these technical features, they should be considered to be within the scope of this specification. In addition, according to the needs of implementation, the various steps / components described in this application can be split into more steps / components, and two or more steps / components or partial operations of steps / components can be combined into new steps / components to achieve the purpose of the present invention.

[0067] Those skilled in the art will readily understand that the above-described embodiments merely represent several implementation methods of the present application, and their descriptions are relatively specific and detailed, but should not be construed as limiting the scope of the invention. It should be noted that a person of ordinary skill in the art may make several variations and improvements without departing from the concept of the present application, and these variations and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present patent application shall be based on the appended claims.

Claims

1. A method for splicing images of hot cast billets, characterized in that: The method includes: A central camera and at least one edge camera are set along the width direction of the billet, and any two adjacent cameras have overlapping fields of view; When the billet passes through the camera, the middle camera and the side camera are triggered simultaneously at a preset time interval to respectively capture a middle image and at least one side image of the billet at that moment; Determine the transformation matrix from the coordinates of all edge camera images to the coordinates of the central camera image, and based on the transformation matrix, transform all edge images collected at the same time into the central image coordinate system at the corresponding time; Determine the edge images that overlap with the central image, search for feature points in the overlapping areas to complete image stitching, and then determine the edge images that overlap with the stitched image in the remaining edge images. In this way, all edge images collected at the same time are stitched onto the central image at the corresponding time to obtain the horizontal stitching image at that time. The horizontal stitching images at adjacent moments are vertically stitched together to obtain the final stitching image of the ingot.

2. The hot casting image splicing method according to claim 1, characterized in that: A middle camera and at least one edge camera are arranged above the billet conveyor roller along the width direction of the billet, or a middle camera and at least one edge camera are arranged below the billet conveyor roller along the width direction of the billet, or a middle camera and at least one edge camera are arranged above and below the billet conveyor roller along the width direction of the billet.

3. The hot casting image splicing method according to claim 2, characterized in that: When the billet passes through the camera, the middle camera and the side camera are triggered simultaneously at a preset time interval to respectively capture a middle image and at least one side image of the billet at that moment, including: When it is detected that the billet passes the camera, the middle camera and the side camera are triggered at the same time to respectively capture the middle image and at least one side image of the billet at that moment, and according to the billet number and camera position, the billet upper surface image is saved to the billet number + upper surface + camera position storage path, and the lower surface image is saved to the billet number + lower surface + camera position storage path, and the images captured by each camera are numbered according to the triggering order; The preset time interval for camera triggering is dynamically determined based on the speed of the conveyor roller to ensure that images captured by the same camera at adjacent moments have overlapping areas; When it is detected that the billet leaves the camera imaging area, the billet imaging is stopped.

4. The hot casting slab image stitching method according to claim 1, characterized in that: The transformation matrix from the image coordinates of all edge cameras to the image coordinates of the central camera is determined by the calibration board, and recalibrated regularly or when the horizontal stitching effect is poor.

5. The hot casting slab image splicing method according to claim 1, characterized in that: Determine the edge images that overlap with the central image and search for feature points in the overlapping area to complete image stitching, including: Determine all the side images that have overlapping areas with the central image, and use the image registration algorithm to find key points in the overlapping areas. Then use the feature matching algorithm to find strongly correlated key points as the feature points of the two images. Generate a transformation matrix based on the feature points, and perform perspective transformation to obtain a stitched image, so that all the side images that have overlapping areas with the central image are stitched onto the central image.

6. The hot casting image splicing method according to claim 1, characterized in that: Perform image fusion processing on the obtained horizontal stitching images and crop out the invalid areas.

7. The hot casting slab image stitching method according to claim 6, characterized in that: The image fusion algorithm is used to smooth the horizontal stitching images, and the black edges of the stitching images are cut off using a rectangle to retain the image in the center area.

8. The hot casting slab image stitching method according to claim 1, characterized in that: The horizontal stitching images at adjacent moments are vertically stitched together to obtain the final stitching image of the ingot, including: If the number of horizontal image stitching is equal to 2, then extract the feature points of the first horizontal stitching image and the second horizontal stitching image, stitch the first horizontal stitching image and the second horizontal stitching image to obtain a stitching image in the vertical direction, and perform image fusion processing on the image; If the number of horizontal image stitching is greater than 2, extract the feature points of the last vertical stitching image and the current horizontal stitching image, stitch the last vertical stitching image and the current horizontal stitching image to obtain a new vertical stitching image, and perform image fusion processing on the image; The invalid area of ​​the final vertical stitching image is cut off to obtain the final ingot stitching image.

9. A computer device, characterized in that: include: A processor and a memory, wherein the memory stores programs or instructions that can be run on the processor, and when the programs or instructions are executed by the processor, the steps of the hot cast slab image stitching method described in any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium, characterized in that Programs or instructions are stored thereon, and when the programs or instructions are executed by the processor, the steps of the hot cast slab image splicing method described in any one of claims 1 to 8 are implemented.

Images