Method and device for positioning ore in ore image and storage medium

Abstract

The invention discloses a method and equipment for positioning ore in an ore image and a storage medium. The method comprises the following steps: acquiring an original ore image, and cutting the original ore image into a plurality of image blocks; based on the plurality of image blocks, performing ore detection by using a target recognition model to obtain a prediction recognition frame of the ore on each image block; mapping the prediction identification frame to the original ore image, and extracting the contour information of the ore in the original ore image; and according to the contour information, combining the prediction identification frames at the cutting positions of the image blocks in the original ore image to obtain a final identification frame of the ore so as to position the ore in the ore image. By means of the scheme, the splicing precision of the ore recognition frame can be improved, and the ore positioning accuracy is improved.

Description

Technical Field

[0001] The present application generally relates to the technical field of ore image segmentation. More specifically, the present application relates to a method, device and computer-readable storage medium for locating ore in an ore image. Background Art

[0002] In the field of ore sorting, the ore is often transported by a conveyor belt system, and the ore transported on the conveyor belt is scanned by combining X-ray transmission technology. The X-ray information is converted into image information to evaluate the quality of the ore and sort it. In actual application scenarios, the ore sorting process faces many challenges, especially the diversity of ore shapes, different sizes and close arrangement. These characteristics require extremely high positioning and segmentation accuracy to ensure the accuracy of sorting, and ore images often have adhesion and overlap, which directly affects the subsequent positioning and segmentation processing. Therefore, how to achieve accurate positioning of the ore has become the key to improving sorting accuracy.

[0003] Among the existing ore positioning methods, there are two main categories: traditional algorithms and neural networks. Traditional algorithms include using the color, texture, shape, etc. of the image for target detection, but it is difficult to design efficient and accurate algorithms when faced with complex scenes where the ore has variable shapes, different sizes, and is closely arranged. Although the neural network method has advantages in processing complex scenes, due to the relatively small size and low resolution of the ore particles, the overall image collected is large, and the neural network is prone to losing key image information during the downsampling process of the processing, resulting in positioning errors or omissions. At present, the problem of positioning small targets in the aforementioned large images has been solved by image cropping and splicing, but in the splicing process, only the recognition box is often considered, and more contextual information is ignored, resulting in limited accuracy.

[0004] In view of this, there is an urgent need to provide a solution for locating ore in ore images, so as to improve the splicing accuracy of the ore identification frame and enhance the accuracy of ore positioning. Summary of the invention

[0005] In order to at least solve one or more of the above-mentioned technical problems, the present application proposes a solution for locating ore in a ore image in multiple aspects.

[0006] In a first aspect, the present application provides a method for locating ore in an ore image, comprising: acquiring an original ore image, and cropping the original ore image into a plurality of image blocks; performing ore detection based on the plurality of image blocks using a target recognition model to obtain a predicted identification frame of the ore on each of the image blocks; mapping the predicted identification frame to the original ore image, and extracting contour information of the ore in the original ore image; and merging the predicted identification frames at the cropping locations of each of the image blocks in the original ore image according to the contour information to obtain a final identification frame of the ore, so as to locate the ore in the ore image.

[0007] In some embodiments, the original ore image is cropped into a plurality of image blocks by the following operations: obtaining an original height of the original ore image; and using the original height as a cropping width, cropping the original ore image into the plurality of image blocks in a horizontal direction.

[0008] In some other embodiments, the process of cropping the original ore image into the plurality of image blocks in the horizontal direction further includes: determining pixels at the end of a previous image block that are advanced by a target number of pixels forward as a starting point for cropping the next image block.

[0009] In some other embodiments, the process of cropping the original ore image into the plurality of image blocks in the horizontal direction further includes: in response to a cropping width of the image block being smaller than the original height, splicing the corresponding image block with its adjacent previous image block into one image block.

[0010] In some further embodiments, the contour information of the ore in the original ore image is extracted by the following operations: binarizing the original ore image; and extracting the contour information of the ore in the original ore image according to the binarization result.

[0011] In some other embodiments, the predicted identification frames at the cropping locations of each image block in the original ore image are merged according to the contour information to obtain the final identification frame of the ore, including: determining, according to the contour information, target identification frames that can be merged at the cropping locations of adjacent image blocks in the original ore image; and merging the target identification frames that can be merged to obtain the final identification frame of the ore.

[0012] In some other embodiments, determining a target recognition frame that can be merged at the cropping location of adjacent image blocks in the original ore image based on the contour information includes: obtaining a group of predicted recognition frames that intersect at the cropping location of adjacent image blocks in the original ore image; judging whether each predicted recognition frame in the group of predicted recognition frames intersects with the same contour based on the contour information; judging whether the center of the intersection area is located inside the contour in response to the intersection of each predicted recognition frame and the same contour; and determining the predicted recognition frame in the group of predicted recognition frames as the target recognition frame that can be merged in response to the center of the intersection area being located inside the contour.

[0013] In some other embodiments, merging the mergeable target identification frames to obtain the final identification frame of the ore includes: determining the maximum identification frame height among the target identification frames as the height of the final identification frame; determining the maximum horizontal distance between the target identification frames as the width of the final identification frame; and determining a new identification frame based on the height of the final identification frame and the width of the final identification frame to merge the mergeable target identification frames to obtain the final identification frame of the ore.

[0014] In a second aspect, the present application provides a device for locating ore in an ore image, comprising: a processor; and a memory, in which program instructions for locating ore in an ore image are stored, and when the program instructions are executed by the processor, the device implements one or more embodiments of the aforementioned first aspect.

[0015] In a third aspect, the present application provides a computer-readable storage medium having stored thereon computer-readable instructions for locating ore in an ore image, wherein when the computer-readable instructions are executed by one or more processors, one or more embodiments of the aforementioned first aspect are implemented.

[0016] Through the solution for locating ore in ore images provided above, the embodiment of the present application cuts the original ore image into multiple image blocks, first performs ore detection through the target recognition model to obtain a predicted recognition frame, then maps the predicted recognition frame to the original ore image, and merges the predicted recognition frames at the cutouts of the image blocks in combination with the contour information of the ore to obtain the final recognition frame of the ore. Based on this, the shape, size, position of the ore particles and the overall structural information of the image can be comprehensively considered, achieving a more accurate merging result, and further improving the accuracy of ore positioning. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] By reading the detailed description below with reference to the accompanying drawings, the above and other purposes, features and advantages of the exemplary embodiments of the present application will become easy to understand. In the accompanying drawings, several embodiments of the present application are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein:

[0018] Figure 1 is an exemplary flowchart showing a method for locating ore in an ore image according to an embodiment of the present application;

[0019] Figure 2 is an exemplary schematic diagram showing an original ore image and an ore contour according to an embodiment of the present application;

[0020] Figure 3 is an exemplary schematic diagram showing image cropping according to an embodiment of the present application;

[0021] Figure 4 is an exemplary schematic diagram showing a predicted identification box of an ore on each image block according to an embodiment of the present application;

[0022] Figure 5 is an exemplary schematic diagram showing the mapping of the predicted identification box of the ore on each image block to the original ore image according to an embodiment of the present application;

[0023] Figure 6 is an exemplary schematic diagram showing a merged identification frame according to an embodiment of the present application;

[0024] Figure 7 is an exemplary schematic diagram showing the final identification frame of all ores according to an embodiment of the present application;

[0025] Figure 8 It is an exemplary structural block diagram showing a device for locating ore in an ore image according to an embodiment of the present application. DETAILED DESCRIPTION

[0026] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0027] It should be understood that the terms "include" and "comprising" used in the specification and claims of the present application indicate the presence of described features, wholes, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or collections thereof.

[0028] It should also be understood that the terms used in this application specification are only for the purpose of describing specific embodiments and are not intended to limit the present application. As used in this application specification and claims, unless the context clearly indicates otherwise, the singular forms of "a", "an" and "the" are intended to include plural forms. It should also be further understood that the term "and / or" used in this application specification and claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes these combinations.

[0029] As used in this specification and claims, the term "if" may be interpreted as "when" or "upon" or "in response to determining" or "in response to detecting," depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is detected" may be interpreted as meaning "upon determination" or "in response to determining" or "upon detection of [described condition or event]" or "in response to detecting [described condition or event]," depending on the context.

[0030] It should be understood that in actual ore sorting, in order to optimize the utilization efficiency of the ore and the recognition ability of the equipment, the ore is usually crushed to control the particle size within the range of 10 to 20 mm. At the same time, in order to meet the needs of high-speed sorting, the conveyor belt width is often set to about 1500 mm. This makes the ore particles relatively small and low in resolution, and the overall image collected is large (that is, large image and small target), which further increases the complexity and challenge of sorting. As known from the above background technology, when neural networks deal with the problem of positioning small targets in large images, they are prone to lose key image information in the downsampling link. For example, after the ore image is downsampled multiple times by the neural network, the ore area in the image may become a point, resulting in positioning errors or omissions. In addition, the width of the ore image is wide, while the width of the input data of the neural network is narrow, so the input of the ore image into the neural network is bound to have the problem of image compression, which will also cause the ore area to become smaller, resulting in positioning errors or omissions.

[0031] At present, although the above problems can be solved by image cropping and splicing, by cropping into small images, the proportion of the ore area in the image is increased, so that more image information can be retained during the downsampling process of the neural network. However, in the splicing process, existing methods often only consider the intersection and union of the recognition boxes, but do not pay attention to the contour information of the target, or cannot determine the contour of the target, thereby ignoring more contextual information, resulting in limited accuracy of the final splicing.

[0032] Based on this, the present application provides a solution for locating ore in ore images. By cropping the original ore image into multiple image blocks and combining the inference results of ore detection and the contour information of the ore with the target recognition model, the global and local features of the ore image are comprehensively considered, and accurate image cropping, efficient reasoning and precise stitching are achieved, thereby improving the accuracy of ore positioning and segmentation, and meeting the high standards for ore sorting.

[0033] The specific implementation of the present application is described in detail below with reference to the accompanying drawings.

[0034] Figure 1 FIG. 1 is an exemplary flowchart of a method 100 for locating ore in an ore image according to an embodiment of the present application. Figure 1 As shown in FIG. 1 , at step S101, an original ore image is obtained and cut into a plurality of image blocks. In some embodiments, X-ray information can be obtained by scanning the ore conveyed on the conveyor belt, for example, using X-ray transmission technology, and then the X-ray information is converted into image information to obtain the original ore image (e.g. Figure 2 Based on the original ore image, it is cropped into multiple image blocks to increase the proportion of the ore area in the image and retain more key image information.

[0035] In some embodiments, the original height of the original ore image is obtained, and the original height is used as the cropping width, and the original ore image is cropped into multiple image blocks along the horizontal direction. Specifically, in some implementation scenarios, with the upper left corner of the original ore image as a reference, multiple squares with the same width and length as the original height are cropped along the horizontal direction to obtain multiple image blocks (for example Figure 3 In other embodiments, when the original ore image is cut into multiple image blocks along the horizontal direction, the pixels at the end of the previous image block that are pushed forward by the target number are determined as the starting point for cutting the next image block. Preferably, the end of the previous image block is pushed forward by two pixels as the starting point for cutting the next image block. That is, among the multiple image blocks obtained by cutting, except for the leftmost image block and the rightmost image block, each image block has an intersection of two pixels with the image blocks before and after it in the horizontal direction.

[0036] In some other embodiments, when the original ore image is cropped into multiple image blocks along the horizontal direction, in response to the cropped width of the image block being smaller than the original height, the corresponding image block and its adjacent previous image block are spliced ​​into one image block. In other words, when the width of the cropped image block is smaller than the height of the original ore image, it indicates that it is at the rightmost end (i.e., the end) of the original ore image. In this scenario, the image block and its adjacent previous image block are combined into one image block.

[0037] Based on the multiple image blocks obtained by the upper cropping, at step S102, based on the multiple image blocks, the target recognition model is used to perform ore detection to obtain the predicted recognition box of the ore on each image block. In some implementation scenarios, the aforementioned target recognition model may include but is not limited to a convolutional neural network model, a single-stage target detection model, or a feature pyramid network model, etc., and the present application does not impose any restrictions in this regard. By inputting multiple image blocks into the trained target recognition model, the predicted recognition box (i.e., the inference result) of the ore on each image block can be directly obtained.

[0038] Then, at step S103, the predicted identification frame is mapped to the original ore image, and the contour information of the ore in the original ore image is extracted. In some implementation scenarios, the predicted identification frame can be mapped to the original ore image based on the coordinate position of the predicted identification frame. In some embodiments, the original ore image can be binarized to extract the contour information of the ore in the original ore image according to the binarization result. As an example, by setting a threshold, the pixels above the threshold are set to 1, and the pixels below the threshold are set to 0 to achieve the aforementioned binarization operation, thereby extracting the contour information of the ore in the original ore image. In addition, the contour information of the ore in the original ore image can also be extracted by, for example, edge-based or region-based contour detection methods, which are not limited in this application. It can be understood that the aforementioned contour information of the ore includes the edge contour of the ore.

[0039] Further, at step S104, the predicted recognition frames at the cutouts of each image block in the original ore image are merged according to the contour information to obtain the final recognition frame of the ore, so as to locate the ore in the ore image. In some embodiments, the target recognition frames that can be merged at the cutouts of adjacent image blocks in the original ore image are determined according to the contour information, and the final recognition frame of the ore is obtained by merging the mergeable target recognition frames. More specifically, in some embodiments, a group of predicted recognition frames that intersect at the cutouts of adjacent image blocks in the original ore image is obtained, and it is determined according to the contour information whether each predicted recognition frame in the predicted recognition frame group has an intersection with the same contour. In response to the intersection of each predicted recognition frame with the same contour, it is determined whether the center of the intersection area is located inside the contour, and in response to the center of the intersection area being located inside the contour, the predicted recognition frame in the predicted recognition frame group is determined as a target recognition frame that can be merged.

[0040] That is, firstly, the prediction recognition frames where adjacent image blocks intersect are obtained to form a prediction recognition frame group, and then it is determined whether each prediction recognition frame in the prediction recognition frame group intersects with the outline of the same ore and whether the center of the intersection area is located inside the outline. When each prediction recognition frame intersects with the same outline and the center of the intersection area is located inside the outline, the prediction recognition frame of the prediction recognition frame group is determined as a target recognition frame that can be merged. It can be understood that the aforementioned intersection judgment or the judgment that the center of the intersection area is located inside the outline is based on the pixel coordinate position. If there are pixels with the same coordinate position, it is determined that there is an intersection or the center of the intersection area is located inside the outline.

[0041] In some embodiments, the maximum recognition frame height in the target recognition frame is determined as the height of the final recognition frame, the maximum horizontal distance between the target recognition frames is determined as the width of the final recognition frame, and then a new recognition frame is determined based on the height of the final recognition frame and the width of the final recognition frame, so as to merge the mergeable target recognition frames and obtain the final recognition frame of the ore. In some implementation scenarios, the aforementioned maximum horizontal distance is the horizontal distance from the leftmost side of one target recognition frame to the rightmost side of another target recognition frame. Thus, by drawing a new recognition frame with the maximum width from the leftmost side of one target recognition frame to the rightmost side of another target recognition frame and with the maximum recognition frame height, the recognition frame merging can be completed to obtain the final recognition frame. This will be combined later. Figure 6 The merging process of the aforementioned recognition frames is described in detail.

[0042] Combined with the above description, it can be seen that the embodiment of the present application obtains the final identification frame by cutting the original ore image into multiple image blocks, and combining the inference results of ore detection by the target recognition model and the contour information of the ore to merge the identification frame. Based on this, the local information such as the shape, size, and position of the ore particles and the overall structural information of the image are comprehensively considered to achieve accurate merging of the identification frame, thereby further improving the accuracy of ore positioning.

[0043] Figure 2 is an exemplary schematic diagram showing an original ore image and an ore contour according to an embodiment of the present application. Figure 2 Figure (a) shows the original ore image. According to the foregoing, X-ray information can be obtained by scanning the ore transmitted on the conveyor belt through X-ray transmission technology, and the original ore image is obtained based on the conversion of X-ray information. Multiple ores 201 are shown in the original ore image. As can be seen from the figure, the ore particles are relatively small and the resolution is low, which will cause the loss of key image information in the neural network downsampling. Therefore, the present application increases the proportion of the ore area in the image by cropping the original ore image into multiple image blocks to retain more key information. Figure 2Figure (b) shows the extracted ore contour. As mentioned above, the original ore image can be binarized to extract the contour information of the ore in the original ore image.

[0044] Figure 3 FIG. 1 is an exemplary schematic diagram showing image cropping according to an embodiment of the present application. Figure 3 As exemplarily shown in the figure, the original ore image is cropped into three image blocks. In some implementation scenarios, the original height H is used as the cropping width, and the upper left corner P of the original ore image is used as the reference, and multiple squares with the same width and length as the original height are cropped horizontally. Among them, the end of the previous image block is moved forward by two pixels as the starting point for cropping the next image block. As an example, the pixel point p1 in the previous image block is used as the starting point for cropping the next image block. In addition, for image blocks whose cropping width is less than the original ore image height, they are merged with the previous image block into one image block. For example, the third image block in the figure merges an image block whose cropping width w is less than the original height h.

[0045] Figure 4 is an exemplary schematic diagram showing the predicted identification box of the ore on each image block according to an embodiment of the present application. Figure 4 As shown in , the rectangular box at the ore on each image block is the predicted identification box of the ore on each image block. In some implementation scenarios, based on the above-mentioned multiple cropped image blocks, a trained target recognition model such as a convolutional neural network model, a single-stage target detection model, or a feature pyramid network model is used to perform ore detection, and a predicted identification box of the ore on each image block can be obtained.

[0046] Figure 5 FIG. 1 is an exemplary schematic diagram showing the mapping of the predicted identification box of the ore on each image block to the original ore image according to an embodiment of the present application. Figure 5 As shown in , the predicted identification frames of the ore on each image block are mapped to the original ore image, and there will be intersecting identification frames at the cutouts, such as those shown by arrows A, B, and C. In addition, the figure also shows the intersecting identification frames shown by arrow D. In the embodiment of the present application, by combining the outline of the ore to merge the intersecting identification frame groups, the shape, size, position of the ore particles and the overall structural information of the image can be comprehensively considered to obtain an accurate merging result.

[0047] Figure 6 is an exemplary schematic diagram showing a merged recognition frame according to an embodiment of the present application. In some embodiments, firstly, a predicted recognition frame having an intersection of adjacent image blocks is obtained to form a predicted recognition frame group, such as Figure 6 FIG. (a) shows an example of two sets of intersecting prediction recognition boxes (i.e., the above Figure 5Take the prediction identification frame group indicated by arrow A and arrow B in the figure as an example. Next, determine whether each prediction identification frame in the prediction identification frame group intersects with the outline of the same ore. For example, each prediction identification frame of the prediction identification frame group indicated by arrow A intersects with the outline S1 of the ore, and each prediction identification frame of the prediction identification frame group indicated by arrow B intersects with the outline S2 of the ore. Further, it is determined that the center of the intersection area is located inside the outline. For example, the figure shows that the centers o1 and o2 of the intersection areas in the two groups of prediction identification frame groups are respectively located inside the outlines S1 and S2, so that the prediction identification frames in the two groups of prediction identification frame groups are target identification frames that can be merged.

[0048] Figure 6 (b) shows the final recognition frame after merging. Specifically, taking the predicted recognition frame group shown by arrow A as an example, a new recognition frame is drawn by setting the maximum recognition frame height in the predicted recognition frame group as the height of the final recognition frame, and setting the maximum horizontal distance between the leftmost side of one target recognition frame and the rightmost side of another target recognition frame as the width of the final recognition frame to obtain the final recognition frame M. For example, the height h1 of the predicted recognition frame on the left side of the predicted recognition frame group is set as the height of the final recognition frame M, and the maximum horizontal distance w1 between the leftmost side of one target recognition frame and the rightmost side of another target recognition frame is set as the width of the final recognition frame M. Similarly, the intersecting recognition frames at arrows A, arrow B and arrow C can be merged. For the intersecting recognition frames at arrow D, since they do not intersect with the same ore contour, the group of recognition frames is not merged.

[0049] Figure 7 is an exemplary schematic diagram showing the final identification frame of all ores according to an embodiment of the present application. Figure 7 It can be seen that based on the embodiment of the present application, an accurate merging result can be obtained, which improves the accuracy of ore positioning. Subsequently, based on the ore positioning result, ore segmentation can be accurately achieved to improve the accuracy of ore sorting.

[0050] Figure 8 FIG. 8 is an exemplary structural block diagram showing a device 800 for locating ore in an ore image according to an embodiment of the present application. Figure 8As shown in , the device 800 of the present application may include a processor 801 and a memory 802, wherein the processor 801 and the memory 802 communicate with each other via a bus. The memory 802 stores program instructions for locating ore in an ore image. When the program instructions are executed by the processor 801, the method steps described in the above text in combination with the accompanying drawings are implemented: obtaining an original ore image, and cutting the original ore image into multiple image blocks; based on the multiple image blocks, using the target recognition model to perform ore detection to obtain a predicted recognition frame of the ore on each of the image blocks; mapping the predicted recognition frame to the original ore image, and extracting the contour information of the ore in the original ore image; and merging the predicted recognition frames at the cutouts of each of the image blocks in the original ore image according to the contour information to obtain the final recognition frame of the ore, so as to locate the ore in the ore image.

[0051] According to the above description in combination with the accompanying drawings, those skilled in the art can also understand that the embodiments of the present application can also be implemented by software programs. Therefore, the present application also provides a computer-readable storage medium. The computer-readable storage medium stores computer-readable instructions for locating ore in an ore image. When the computer-readable instructions are executed by one or more processors, the present application in combination with the accompanying drawings can be implemented. Figure 1 A method for locating a mineral in a mineral image is described.

[0052] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0053] It should be noted that although the operations of the method of the present application are described in a specific order in the accompanying drawings, this does not require or imply that the operations must be performed in this specific order, or that all the operations shown must be performed to achieve the desired results. On the contrary, the steps depicted in the flow chart can be performed in a different order. Additionally or alternatively, some steps can be omitted, multiple steps can be combined into one step, and / or one step can be decomposed into multiple steps.

[0054] It should be understood that when the terms "first", "second", "third" and "fourth" are used in the claims, the specification and the drawings of the present application, they are only used to distinguish different objects, rather than to describe a specific order. The terms "include" and "comprise" used in the specification and claims of the present application indicate the presence of the described features, wholes, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or their collections.

[0055] It should also be understood that the terms used in this application specification are only for the purpose of describing specific embodiments and are not intended to limit the present application. As used in this application specification and claims, unless the context clearly indicates otherwise, the singular forms of "a", "an" and "the" are intended to include plural forms. It should also be further understood that the term "and / or" used in this application specification and claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes these combinations.

[0056] Although the implementation methods of the present application are as above, the contents described are only examples adopted to facilitate the understanding of the present application, and are not intended to limit the scope and application scenarios of the present application. Any technician in the technical field described in the present application can make any modifications and changes in the form and details of the implementation without departing from the spirit and scope disclosed in the present application, but the scope of patent protection of the present application shall still be subject to the scope defined in the attached claims.

[0057] In addition, the collection and acquisition of various data in this application complies with relevant laws and regulations and is authorized by the data provider. Any organization or individual that needs to obtain external data must obtain authorization in accordance with the law and ensure data security. It is not allowed to illegally collect, use, process, or transmit unauthorized or unprotected data, or to illegally buy, sell, provide, or disclose unauthorized or unprotected data.

Claims

1. A method for locating an ore in an ore image, comprising: Acquire an original ore image, and cut the original ore image into multiple image blocks; Based on the multiple image blocks, use the target recognition model to perform ore detection to obtain a predicted recognition frame of the ore on each of the image blocks; Mapping the predicted identification frame to the original ore image, and extracting the contour information of the ore in the original ore image; as well as The predicted identification frames at the cutout positions of the image blocks are merged in the original ore image according to the contour information to obtain a final identification frame of the ore, so as to locate the ore in the ore image.

2. The method according to claim 1, wherein the original ore image is cropped into a plurality of image blocks by the following operations: Acquiring the original height of the original ore image; and The original ore image is cropped into the plurality of image blocks along a horizontal direction with the original height being used as the cropping width.

3. The method according to claim 2, wherein in the process of cutting the original ore image into the plurality of image blocks in the horizontal direction, further comprising: The pixels at the end of the previous image block that are advanced by the target number are determined as the starting point for cropping the next image block.

4. The method according to claim 2 or 3, wherein in the process of cutting the original ore image into the plurality of image blocks in the horizontal direction, the method further comprises: In response to the cropped width of the image block being smaller than the original height, the corresponding image block and its adjacent previous image block are spliced ​​into one image block.

5. The method according to claim 1, wherein the contour information of the ore in the original ore image is extracted by the following operations: Binarizing the original ore image; and The contour information of the ore in the original ore image is extracted according to the binarization result.

6. The method according to claim 1 or 5, wherein merging the predicted identification frames at the cutout positions of each image block in the original ore image according to the contour information to obtain the final identification frame of the ore comprises: Determine, according to the contour information, a target recognition frame that can be merged at the cutout locations of adjacent image blocks in the original ore image; as well as The target identification frames that can be merged are merged to obtain the final identification frame of the ore.

7. The method according to claim 6, wherein determining the target recognition frame that can be merged at the cutout position of adjacent image blocks in the original ore image according to the contour information comprises: Obtaining a predicted recognition frame group where there is an intersection at the adjacent image block cutouts in the original ore image; Determining, according to the contour information, whether each predicted identification frame in the predicted identification frame group has an intersection with the same contour; In response to the intersection of each predicted recognition frame and the same contour, determining whether the center of the intersection area is located inside the contour; and In response to the center of the intersection area being located inside the contour, the predicted recognition box in the predicted recognition box group is determined as the target recognition box that can be merged.

8. The method according to claim 7, wherein merging the mergeable target identification frames to obtain the final identification frame of the ore comprises: Determine the maximum recognition frame height in the target recognition frame as the height of the final recognition frame; Determining the maximum horizontal distance between the target recognition frames as the width of the final recognition frame; as well as A new identification frame is determined based on the height of the final identification frame and the width of the final identification frame, so as to merge the mergeable target identification frames and obtain the final identification frame of the ore.

9. A device for locating ore in an ore image, comprising: processor; as well as A memory, wherein program instructions for locating ore in an ore image are stored, and when the program instructions are executed by the processor, the device implements the method according to any one of claims 1-8.

10. A computer-readable storage medium having stored thereon computer-readable instructions for locating ore in an ore image, wherein when the computer-readable instructions are executed by one or more processors, the method according to any one of claims 1 to 8 is implemented.

Images