An identification system, an identification method and a sorting machine

Abstract

The present disclosure relates to the field of material sorting, and particularly relates to a recognition system, a recognition method and a sorting machine. The recognition system comprises: a conveying device; a first recognition device configured to acquire a first image of a first range of the conveying device to obtain a first material image of a material passing through the first range; a second recognition device configured to acquire a second image of a second range of the conveying device to obtain a second material image of the material passing through the second range, the second side being the other side opposite to the first side, and the second range and the first range having an overlapping range in the width direction; a thickness measuring sensor arranged above the conveying device and configured to determine at least the pose information of the material passing through the overlapping range; and a controller. The recognition system does not need image splicing when performing image recognition, and is simple to calculate, low in consumption of computing resources, effectively guarantees image quality, and improves material classification accuracy. The real-time performance and processing throughput of the recognition system are high, and the reliability of the recognition result is improved.

Description

Technical Field

[0001] This disclosure relates to the field of material sorting, specifically to an identification system, identification method, and sorting machine. Background Technology

[0002] For dual-source X-ray identification technology used in mining, two X-ray emitters and detectors are used for separate detection. During material detection, an image stitching scheme is typically employed to combine the images acquired from the two detection areas into a complete wide-area image before identification. However, this method is computationally complex and resource-intensive, and errors and distortions are prone to occur during the stitching process, leading to low identification accuracy. Furthermore, since this method requires both images to be acquired and stitched before identification can begin, the significant delay results in poor real-time performance, and it fails to fully utilize auxiliary information provided by other detectors, leading to resource waste. Summary of the Invention

[0003] To overcome the problems existing in related technologies, an exemplary embodiment of this disclosure provides an identification system, the identification system comprising: a conveying device for conveying a material to be identified; a first identification device for acquiring a first image of a first range of the conveying device to obtain a first material image of the material passing through the first range, wherein the first range is the range from a first side to the middle in the width direction of the conveying device; a second identification device for acquiring a second image of a second range of the conveying device to obtain a second material image of the material passing through the second range, wherein the second range is the range from a second side to the middle in the width direction of the conveying device, the second side being the opposite side to the first side, and the second range and the first range having an overlapping range in the width direction; a thickness sensor disposed above the conveying device for at least determining the pose information of the material passing through the overlapping range; and a controller communicatively connected to the first identification device, the second identification device, and the thickness sensor for determining the category of the material based on the first material image, the second material image, and the pose information.

[0004] In some embodiments, the first identification device includes: a first emitting end disposed above the first range of the conveying device for emitting X-rays into the first range; and a first receiving end disposed below the first range of the conveying device for receiving the X-rays emitted by the first emitting end; the second identification device includes: a second emitting end disposed above the second range of the conveying device for emitting X-rays into the second range; and a second receiving end disposed below the second range of the conveying device for receiving the X-rays emitted by the second emitting end; wherein the first identification device and the second identification device are offset from each other in the conveying direction of the conveying device.

[0005] In some embodiments, the controller is configured to: determine the material category based on the first material image for materials that pass through the first range but not through the overlapping range; determine the material category based on the second material image for materials that pass through the second range but not through the overlapping range; and determine the material category based on the first category determined according to the first material image, the second category determined according to the second material image, and the pose information for materials that pass through the overlapping range.

[0006] Secondly, this disclosure also provides an identification method applied to the identification system described in the first aspect, the identification method comprising: acquiring a first image through the first identification device; acquiring a second image through the second identification device; determining the pose information of material passing through the overlapping range through the thickness sensor; and determining the category of the material based on the first image, the second image and the pose information.

[0007] In some embodiments, determining the category of the material based on the first image, the second image, and the pose information includes: if the material passes through the first range but not through the overlapping range, then determining the category of the material based on the first material image; if the material passes through the second range but not through the overlapping range, then determining the category of the material based on the second material image; if the material passes through the overlapping range, then determining the category of the material based on the first category determined according to the first material image, the second category determined according to the second material image, and the pose information.

[0008] In some embodiments, determining the material category based on a first category determined from the first material image, a second category determined from the second material image, and the pose information includes: determining the first material image, the second material image, and the pose information corresponding to the material; determining corresponding first category information and second category information based on the first material image and the second material image; if the material categories represented by the first category information and the second category information are consistent, then determining the material category; if the material categories represented by the first category information and the second category information are inconsistent, then: determining a weight parameter based on the pose information, wherein the weight parameter includes a first weight corresponding to the first category information and a second weight corresponding to the second category information; and determining the material category based on the weight parameter, the first category information, and the second category information.

[0009] In some embodiments, determining the first material image, the second material image, and the pose information corresponding to the material includes: determining first feature information based on the first material image, wherein the first feature information includes at least a first timestamp; determining second feature information based on the second material image, wherein the second feature information includes at least a second timestamp; and determining the first material image, the second material image, and the pose information corresponding to the material based at least on the first timestamp, the second timestamp, and the pose timestamp in the pose information.

[0010] In some embodiments, determining the weight parameters based on the pose information includes: determining the position information of the material based on the pose information, wherein the position information represents the position of the material relative to the width direction of the conveying device; and determining the weight parameters based on the position information, wherein if the position information represents the material being closer to the first side, then the first weight is greater than the second weight, and if the position information represents the material being closer to the second side, then the second weight is greater than the first weight.

[0011] In some embodiments, determining the weight parameters based on the pose information further includes: determining the height information of the material based on the pose information, wherein the height information represents the thickness of the material; and updating the weight parameters based on the height information, wherein if the height information represents the thickness of the material exceeding a threshold, the difference between the first weight and the second weight is increased.

[0012] Thirdly, this disclosure also provides a sorting machine, comprising: an identification system as described in the first aspect, for determining the category of the material; and a sorting device disposed downstream of the conveying device of the identification system, for sorting the material based on the category determined by the identification system.

[0013] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure.

[0014] The technical solutions provided by the embodiments of this disclosure can include the following beneficial effects: the recognition system does not require image stitching when performing image recognition, simplifying calculations and reducing computational resource consumption. During the recognition process, it can effectively ensure image quality, eliminate the interference of image errors on accuracy, and effectively improve the accuracy of material classification. In addition, the recognition system provided by this disclosure has high real-time performance and processing throughput, and can make full use of the data from the thickness sensor, further improving the reliability of the recognition results. Attached Figure Description

[0015] This disclosure can be better understood by describing exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, in which: Figure 1 This is a schematic diagram of an identification system according to exemplary embodiments of the present disclosure; Figure 2 This is a flowchart illustrating an identification method according to exemplary embodiments of the present disclosure; Figure 3 This is a flowchart illustrating an identification method according to exemplary embodiments of the present disclosure; Figure 4 This is a flowchart illustrating an identification method according to exemplary embodiments of the present disclosure. Detailed Implementation

[0016] The following describes specific embodiments of this disclosure. It should be noted that, in order to maintain brevity, this specification cannot provide a detailed description of all features of the actual embodiments. It should be understood that, in the actual implementation of any embodiment, just as in any engineering or design project, various specific decisions are often made to achieve the developer's specific goals and to meet system-related or business-related constraints, and this can change from one embodiment to another. Furthermore, it is understood that although the efforts made in this development process may be complex and lengthy, for those skilled in the art related to the content of this disclosure, changes in design, manufacturing, or production based on the technical content disclosed herein are merely conventional technical means and should not be construed as insufficient content of this disclosure.

[0017] Unless otherwise defined, the technical or scientific terms used in this disclosure shall have the ordinary meaning understood by one of ordinary skill in the art to which this disclosure pertains. The terms “first,” “second,” and similar terms used in this disclosure do not indicate any order, quantity, or importance, but are merely used to distinguish different components. The terms “a” or “one,” etc., do not indicate a quantity limitation, but rather indicate the presence of at least one. The terms “comprising,” “including,” etc., mean that the element or object preceding “comprising” or “including” encompasses the elements or objects listed following “comprising” or “including” and their equivalents, and do not exclude other elements or objects. The terms “connected,” “linked,” etc., are not limited to physical or mechanical connections, nor are they limited to direct or indirect connections.

[0018] Based on the aforementioned technical issues, such as Figure 1 As shown, this disclosure provides an identification system, which includes: a conveying device 110, a first identification device 120, a second identification device 130, a thickness sensor 140, and a controller 150.

[0019] The conveying device 110 is used to convey the material to be identified. The conveying device 110 can be a belt conveyor, chain conveyor, etc., and its upper surface can be used to place the material. By moving the conveying device 110, the material can be moved to achieve the conveying of the material.

[0020] The first identification device 120 is used to acquire a first image of a first range of the conveying device 110 to obtain a first material image of the material passing through the first range. The first range is the area from the first side to the middle of the conveying device 110 in the width direction. The first identification device 120 can be an image acquisition device, such as an industrial camera or an X-ray identification device. The range on the conveying device 110 that the first identification device 120 can acquire is the first range. The first identification device 120 can be used to acquire images of the first range on the conveying device 110 to determine the first image. When the first identification device 120 acquires images, images of the material located within the first range can be acquired. The first identification device 120 can be disposed above the conveying device 110, located on the first side in the width direction of the conveying device 110, so that the first range covers the first side to the middle in the width direction of the conveying device 110, thereby enabling the first identification device 120 to acquire the material from the first side to the middle in the width direction of the conveying device 110, thus obtaining a first material image. The first material image may include one or more materials.

[0021] The second identification device 130 is used to acquire a second image of a second range passing through the conveying device 110, thereby obtaining a second material image of the material passing through the second range. The second range is the area from the second side to the middle of the conveying device 110 in the width direction, the second side being the opposite side to the first side, and the second range and the first range overlap in the width direction. The second identification device 130 can be an image acquisition device, such as an industrial camera or an X-ray identification device. The range on the conveying device 110 that the second identification device 130 can acquire is the second range. The second identification device 130 can be used to acquire images of the second range on the conveying device 110, thereby determining the second image. When the second identification device 130 performs image acquisition, images of the material located within the second range can be acquired. The second identification device 130 can be positioned above the conveying device 110, located on the second side in the width direction of the conveying device 110, so that the second range covers the second side to the middle in the width direction of the conveying device 110, thereby enabling the second identification device 130 to acquire the material from the second side to the middle in the width direction of the conveying device 110, thus obtaining a second material image. The second material image may include one or more materials. The first identification device 120 and the second identification device 130 can be arranged side by side in the width direction of the conveying device 110, such that the first and second ranges respectively captured by the first identification device 120 and the second identification device 130 can cover the width of the conveying device 110, and the first and second ranges overlap in the middle of the conveying device 110. Therefore, the first identification device 120 and the second identification device 130 can independently perform image acquisition and recognition, and the acquired images can partially overlap. Based on the aforementioned overlapping range, after the first identification device 120 and the second identification device 130 each acquire the first material image and the second material image, they can determine whether the first identification device 120 and the second identification device 130 have acquired the same material based on temporal correlation and feature matching within the overlapping range, and then perform matching of the first material image and the second material image in the subsequent process.

[0022] A thickness sensor 140 is disposed above the conveying device 110 and is used at least to determine the pose information of the material passing through the overlapping area. The thickness sensor 140 may also be located above the conveying device 110, and may be disposed between the first identification device 120 and the second identification device 130. The thickness sensor 140 can detect the material within the overlapping area, thereby obtaining the pose information of the material within the overlapping area, including the material's position, orientation, height, and other information.

[0023] The controller 150 is communicatively connected to the first identification device 120, the second identification device 130, and the thickness sensor 140, and is used to determine the material category based on a first material image, a second material image, and pose information. The controller 150 can communicate with the first identification device 120, the second identification device 130, and the thickness sensor 140, thereby receiving the first and second material images acquired by the first identification device 120 and the second identification device 130 respectively. It can also receive the pose information of the material within the overlapping area detected by the thickness sensor 140. Furthermore, the controller 150 can determine the material category based on the first material image, the second material image, and the pose information, thereby accurately and quickly outputting the material category. Specifically, the controller 150 can identify and detect the first material image and the second material image respectively, and determine their corresponding identification results. Then, based on the features or acquisition time of the first and second material images, it can match them to determine if they belong to the same material. Thus, there is no need to stitch the images together; simply matching them confirms that they belong to the same material. Finally, based on the pose information detected by the thickness sensor 140, a material identification strategy can be determined. Specifically, based on pose information such as position, height, and orientation, the weights corresponding to the identification results of the first and second material images can be determined. Finally, the identification results of the first and second material images are fused based on these weights to determine the category to which the material belongs.

[0024] According to this embodiment, by setting a first identification device 120 and a second identification device 130 in the width direction of the conveying device 110, the first identification device 120 acquires a first material image within a first range, and the second identification device 130 acquires a second material image within a second range, with the first range and the second range overlapping. This allows the two identification devices to independently acquire and identify the material within their respective coverage areas, without needing to stitch the first material image with the second material image. Subsequent identification processing can be completed simply by determining whether the two images correspond to the same material based on temporal correlation and feature matching of the overlapping range. This avoids complex calculations such as registration, geometric correction, and brightness equalization during image stitching, reducing system computational complexity and computational resource consumption. It also avoids the impact of stitching errors on the identification results, which is beneficial to improving the accuracy and stability of material identification. Since the first recognition device 120 and the second recognition device 130 can independently complete image acquisition and recognition processing, the two recognition processes can be executed in parallel without waiting for all images on both sides to be acquired and stitched together before recognition. This effectively shortens the recognition delay, improves the system's real-time response capability and processing throughput per unit time, and meets the rapid recognition requirements of the conveying device 110 when continuously conveying materials. By setting a thickness sensor 140 to acquire the pose information of materials passing through the overlapping range, and the controller 150 combining the first material image, the second material image, and the pose information to jointly determine the material category, the position, height, and posture information acquired by the thickness sensor 140 directly participate in the recognition decision, which can further improve the reliability and accuracy of the recognition results. By matching the images on both sides based on the acquisition time and image features of the first and second material images, the controller 150 can accurately determine the image correspondence of the same material from multiple materials, providing a reliable basis for subsequent recognition result fusion. This helps reduce the probability of mismatch during continuous conveying of multiple materials and improves the system's recognition robustness under complex working conditions. Furthermore, the controller 150 determines the fusion weight of the corresponding recognition results of the first material image and the second material image based on the pose information, and determines the material category based on the weighted fusion result. This allows the final recognition result to make full use of the recognition information on both sides and the auxiliary information provided by the thickness sensor 140, avoiding the influence of factors such as imaging angle, occlusion or projection quality differences on a single recognition result, thereby further improving the accuracy of material classification and the reliability of recognition results.

[0025] In some embodiments, the first identification device 120 includes a first transmitter and a first receiver.

[0026] A first emitting end is disposed above a first range of the conveying device 110 and is used to emit X-rays into the first range. The first emitting end can be a radiation source, disposed above the first range of the conveying device 110, capable of emitting X-rays to the outside. The X-rays emitted by the first emitting end can cover the first range, so that after the first emitting end emits X-rays into the first range, the X-rays can pass through the material conveyed into the first range. After the X-rays penetrate the first range, they can be received by the first receiving end, thereby further determining a first image based on the X-rays received by the first receiving end.

[0027] A first receiving end is disposed below a first area of ​​the conveying device 110 and is used to receive X-rays emitted by a first transmitting end. The first receiving end may be disposed opposite to the first transmitting end, and the X-rays emitted by the first transmitting end can pass through the material and the conveying device 110 and be received by the first receiving end. The first receiving end can receive the X-rays emitted by the first transmitting end and determine the corresponding second image based on the received X-rays.

[0028] The second identification device 130 includes a second transmitter and a second receiver.

[0029] The second emitting end is disposed above the second range of the conveying device 110 and is used to emit X-rays into the second range. The second emitting end can be a radiation source, disposed above the second range of the conveying device 110, capable of emitting X-rays to the outside. The X-rays emitted by the second emitting end can cover the second range, so that after the second emitting end emits X-rays into the second range, the X-rays can pass through the material conveyed into the second range. After the X-rays penetrate the second range, they can be received by the second receiving end, thereby further determining a second image based on the X-rays received by the second receiving end.

[0030] The second receiving end is disposed below the second range of the conveying device 110 and is used to receive X-rays emitted by the second transmitting end. The second receiving end can be disposed opposite to the second transmitting end, and the X-rays emitted by the second transmitting end can pass through the material and the conveying device 110 and be received by the second receiving end. The second receiving end can receive the X-rays emitted by the second transmitting end and determine the corresponding second image based on the received X-rays.

[0031] Among them, such as Figure 1As shown, the first identification device 120 and the second identification device 130 are staggered in the conveying direction of the conveying device 110. The first range and the second range can cover the entire conveying area along the width direction of the conveying device 110, and form an overlapping detection area in the middle, so that the material passing through the overlapping detection area can be collected by the first identification device 120 and the second identification device 130 respectively, so as to obtain the corresponding first material image and second material image respectively. The first identification device 120 and the second identification device 130 can be staggered in the conveying direction of the conveying device 110. Specifically, since both the first identification device 120 and the second identification device 130 acquire images and identify materials using X-rays, if the coverage areas of the first identification device 120 and the second identification device 130 overlap, it may cause the X-rays emitted by the first and second transmitting ends to interfere with each other, and affect the signals received by the first and second receiving ends, resulting in abnormal acquisition of the first and second images. Therefore, the first identification device 120 and the second identification device 130 can be staggered in the conveying direction of the conveying device 110, and the first identification device 120 and the second identification device 130 can overlap in the width direction of the conveying device 110. This can effectively avoid mutual interference while ensuring the integrity and accuracy of image acquisition of the first identification device 120 and the second identification device 130, and further improve the accuracy of image acquisition.

[0032] According to this embodiment, the first identification device 120 includes a first transmitting end and a first receiving end arranged opposite to each other, and the second identification device 130 includes a second transmitting end and a second receiving end arranged opposite to each other. The first and second transmitting ends emit X-rays into their respective detection ranges, and the corresponding receiving ends receive the X-ray signals after penetrating the material, thereby obtaining a first material image and a second material image respectively. This enables independent X-ray imaging of the two detection areas, improving the stability and reliability of image acquisition. By having the first and second ranges jointly cover the entire conveying area in the width direction of the conveying device 110, and forming an overlapping detection area in the middle, materials located at any position on the conveying device 110 can be completely acquired by at least one identification device. Materials passing through the overlapping detection area can obtain a first material image and a second material image respectively, ensuring that image acquisition covers the width of the conveying device 110, thus improving the completeness and accuracy of material identification. By staggering the first identification device 120 and the second identification device 130 along the conveying direction of the conveying device 110, the irradiation areas corresponding to the two X-ray emitting ends are spatially offset, avoiding the crosstalk caused by two X-ray sources irradiating the same area simultaneously and the mutual interference of received signals. This ensures that the first and second receiving ends can stably receive their respective transmitted X-ray signals, improving the imaging quality of the first and second material images and reducing image noise and detection errors caused by X-ray interference. Therefore, the identification system can obtain high-quality bilateral detection images without image stitching, fully utilizing the advantages of dual-source detection while ensuring the independence of each detection result, which is beneficial for further improving the accuracy of material classification and the reliability of the identification system.

[0033] In some embodiments, the controller 150 is configured to: determine the material category based on a first material image for materials that pass through a first range but not through an overlapping range; determine the material category based on a second material image for materials that pass through a second range but not through an overlapping range; and determine the material category based on a first category determined according to the first material image, a second category determined according to the second material image, and pose information for materials that pass through an overlapping range. For materials that pass through the first range but not through an overlapping range, it can be determined that they only pass through the first range and not through the second range, thus the first material image can be determined as a complete image of the material. Therefore, the material category can be determined based solely on the first material image. Similarly, for materials that pass through the second range but not through an overlapping range, it can be determined that they only pass through the second range and not through the first range, thus the second material image can be determined as a complete image of the material. Therefore, the material category can be determined based solely on the second material image. For materials that overlap, it can be determined that the material is partially located within the first range and partially within the second range. Therefore, both the first identification device 120 and the second identification device 130 have acquired at least a portion of the image of the material. In this case, the material category can be finally determined based on the first category determined from the first material image and the second category determined from the second material image, combined with the first category, the second category, and the pose information.

[0034] According to this embodiment, materials are categorized into three cases based on whether they pass through an overlapping range: those passing through only a first range, those passing through only a second range, and those passing through an overlapping range. Different recognition strategies are employed for each case. For materials passing through only the first or second range, the material category is determined directly based on the corresponding side's material image without the need for fusion of two-sided recognition results. Only for materials passing through an overlapping range are the first category, the second category, and pose information combined to determine the final category. This avoids performing two-sided fusion calculations on all materials, thereby reducing system computational complexity and improving recognition processing efficiency and real-time performance. For materials completely captured by a single-sided recognition device, category recognition is directly completed using the complete single-sided material image, ensuring recognition accuracy while reducing data processing volume and improving recognition efficiency and system resource utilization. For materials passing through an overlapping range, the recognition results obtained by the first and second recognition devices 120 and the pose information obtained by the thickness sensor 140 are used simultaneously to determine the material category. This effectively avoids misjudgments caused by relying solely on a single-sided image, improving the accuracy of material classification and the reliability of recognition results under complex working conditions.

[0035] Based on the same inventive concept, this disclosure also provides an identification method, such as Figure 2 As shown, the identification method applied to any of the aforementioned embodiments includes the following steps.

[0036] Step S210: The first image is acquired by the first recognition device 120. Firstly, the first image of a first range on the conveying device 110 can be acquired by the first recognition device 120, i.e., as shown... Figure 1 The first image shown is located within the area covered from the left to the center of the image. The first image can capture at least a portion of the material within the first range, and a first material image corresponding to that material can be captured from the first image.

[0037] In step S220, a second image is acquired by the second recognition device 130. The second recognition device 130 can acquire a second image of a second range on the conveying device 110, i.e., as shown... Figure 1 The second image shown is located within the area covered from the right to the center of the image. The second image can capture at least a portion of the material within the second range, and a second material image corresponding to that material can be captured from the second image. Since the first identification device 120 and the second identification device 130 independently complete image acquisition, there is no need to stitch the first and second images together; image recognition can be performed separately, and each can determine its corresponding result. Because the first identification device 120 and the second identification device 130 can each capture material images, and the first and second ranges can cover the surface of the conveying device in the width direction, and there is an overlap between the first and second ranges, for any given material, it may be captured only by the first identification device 120, only by the second identification device 130, or by both the first and second identification devices 120 and 130. Therefore, especially for materials captured by both the first and second identification devices 120 and 130, it is necessary to match the images on both sides to determine that they belong to the same material.

[0038] In step S230, the thickness sensor 140 determines the pose information of the material passing through the overlapping area. The thickness sensor 140 is positioned above the overlapping area and can detect the material passing through it, thereby determining the material's position in the width direction, as well as its posture and height. Since both the first identification device 120 and the second identification device 130 can acquire images of the material within this overlapping area, the pose information can be combined to determine the correspondence between the images on both sides and the same material.

[0039] Step S240: Based on the first image, the second image, and the pose information, determine the material category. Based on the first image, a first material image corresponding to the material within the first image can be determined; based on the second image, a second material image corresponding to the material within the second image can be determined. Since non-uniform materials such as ores have different shapes and target element contents at different locations, performing identification and detection on the first and second material images separately may yield the same or different results. Therefore, to ensure the stability of the final material category determination, the material category is comprehensively determined based on the first and second images, combined with the pose information of the material within the overlapping area, to ensure the accuracy of material identification.

[0040] According to this embodiment, the first recognition device 120 and the second recognition device 130 independently complete image acquisition and recognition processing, enabling the two detection processes to be executed in parallel without waiting for the images on both sides to be stitched together before recognition begins. This shortens the overall processing latency from image acquisition to recognition result output, improves the real-time response capability and processing throughput per unit time of the recognition system, and is applicable to online rapid detection in continuous conveying scenarios such as ore sorting. The thickness sensor 140 acquires the pose information of the material passing through the overlapping range, and uses the pose information together with the first and second images as the basis for determining the material category. This fully utilizes the information provided by the thickness sensor 140, improving the reliability and accuracy of the recognition results. For the first and second material images corresponding to the same material, the final material category is determined by combining the recognition information from both sides and the material pose information. This allows the information acquired from different detection angles to complement each other and reduces the impact of inconsistent recognition results on the final classification result caused by uneven material composition, differences in local element distribution, or changes in unilateral imaging quality. This improves the accuracy and robustness of material classification under complex working conditions.

[0041] In some embodiments, such as Figure 3 As shown, step S240, based on the first image, the second image, and the pose information, determines the category of the material, including the following steps.

[0042] If the material passes through the first range but not the overlapping range, then step S241 is executed to determine the material category based on the first material image. When the material only passes through the first range and not the overlapping range, it can be assumed that only the first identification device 120 can capture the image of the material based on its location. Therefore, the first material image corresponding to this material in the first image captured by the first identification device 120 is the complete image of the material, and the material category can be directly determined based on the first material image, ensuring the accuracy of the identification result.

[0043] If the material passes through the second range but not the overlapping range, step S242 is executed to determine the material category based on the second material image. When the material only passes through the second range and not the overlapping range, it can be assumed that only the second identification device 130 can capture the image of the material based on its location. Therefore, the second material image corresponding to this material in the second image captured by the second identification device 130 is the complete image of the material, and the material category can be directly determined based on the second material image, ensuring the accuracy of the identification result.

[0044] If the material crosses the overlapping area, step S243 is executed, determining the material's category based on the first category determined from the first material image, the second category determined from the second material image, and the pose information. When the material crosses the overlapping area, it can be determined that the material has been captured by both the first recognition device 120 and the second recognition device 130. A first material image corresponding to the material exists in the first image, and a second material image corresponding to the material also exists in the second image. Since part of the material is captured by the first recognition device 120 and part by the second recognition device 130, to ensure the accuracy of recognition and detection, the categories of the first and second material images can be determined separately. The first category is determined based on the first material image, and the second category is determined based on the second material image. Subsequently, based on the first and second categories, combined with the pose information of the material within the overlapping area, the material's category is further determined. This achieves more accurate material detection.

[0045] According to this embodiment, for materials that only pass through the first or second range, since the corresponding identification device can acquire a complete image of the material, category identification is directly completed based on the complete material image, avoiding unnecessary data fusion processing. This effectively improves identification efficiency while ensuring accuracy. For materials that pass through overlapping ranges, the first and second categories are determined based on the first and second material images, respectively, and the final material category is determined by combining the pose information acquired by the thickness sensor 140. This effectively improves the accuracy and reliability of the identification results. Because materials such as ores have uneven composition and differences in element distribution and morphological characteristics in different parts, a single-sided image may only reflect local features of the material. This embodiment determines the final identification result by fusing the first category, the second category, and pose information. This reduces the impact of inconsistent identification results on the final classification result caused by differences in local imaging, detection angles, or material non-uniformity, improving the accuracy and robustness of material classification under complex working conditions.

[0046] In some embodiments, such as Figure 4As shown, step S243, which determines the category of the material based on the first category determined from the first material image, the second category determined from the second material image, and the pose information, includes the following steps.

[0047] Step S2431: Determine the first material image, the second material image, and the pose information corresponding to the material. First, based on the first image, determine the first material image corresponding to the material collected in the first image. Based on the second image, determine the second material image corresponding to the same material collected in the second image. According to the information collected by the thickness sensor 140 on the material in the overlapping area, the pose information of the same material in the overlapping area can be determined, including the height of the material, the location of the material, and the size of the material.

[0048] Step S2432: Based on the first material image and the second material image, determine the corresponding first category information and second category information respectively. For the first material image and the second material image belonging to the same material, they can be identified separately. Based on the first material image, the first category information corresponding to the first material image can be determined; based on the second material image, the second category information corresponding to the second material image can be determined. Although the first identification device 120 and the second identification device 130 overlap in the image acquisition range along the width direction of the conveying device 110, the areas of material that the first identification device 120 and the second identification device 130 can acquire are different. Therefore, the results of the first category information and the second category information determined based on the first material image and the second material image may not be the same. Especially for the identification and sorting of ore, the distribution of coal and ore elements within it is not uniform. Therefore, there may be situations where the first category information and the second category information represent different material categories, or the first category information and the second category information represent the same material category. Furthermore, this embodiment can make the material identification process more efficient and flexible, which is conducive to achieving efficient online identification of materials in continuous conveying scenarios such as ore sorting, and further improves the overall operating efficiency and resource utilization of the system.

[0049] If the material categories represented by the first category information and the second category information are consistent, then step S2433 is executed to determine the material category. Based on the first category information and the second category information, the corresponding material category can be determined separately. When the material categories represented by the first category information and the second category information are consistent, it can be directly determined that the material belongs to the category represented by both the first and second category information, without further determination.

[0050] If the material categories represented by the first category information and the second category information are inconsistent, then step S2434 is executed to determine the weight parameters based on the pose information, wherein the weight parameters include a first weight corresponding to the first category information and a second weight corresponding to the second category information; and step S2435, the material category is determined based on the weight parameters, the first category information, and the second category information. When the material categories represented by the first category information and the second category information are inconsistent, it may be due to uneven distribution of elements within the material, mutual occlusion, or other reasons. In this case, further determination of the material category is required. Therefore, when the material categories represented by the first category information and the second category information are inconsistent, the weight parameters corresponding to the first category information and the second category information can be determined based on the pose information, and further, based on the weight parameters and the first and second category information, the material category is determined through weighted calculation. Specifically, the weight parameters can be the weights of the first category information and the second category information in the material category determination process, wherein the weight parameters include a first weight and a second weight. The first weight can be the weight corresponding to the first category information in the material category determination process, and the second weight can be the weight corresponding to the second category information in the material category determination process. The material category can be determined by weighted fusion based on the first category information, the second category information, and the weight parameters, resulting in higher accuracy.

[0051] According to this embodiment, by acquiring the first material image, the second material image, and the pose information corresponding to the same material, and determining the corresponding first category information and second category information, the first identification device 120 and the second identification device 130 can independently complete the identification, improving the efficiency of material identification. When the material categories represented by the first category information and the second category information are consistent, the consistent identification result is directly used as the final material category without additional fusion calculation. This reduces computational overhead, shortens the identification processing time, and improves the system's real-time performance and processing efficiency while ensuring identification accuracy. When the material categories represented by the first category information and the second category information are inconsistent, the weight parameters corresponding to the first category information and the second category information are determined by combining the pose information acquired by the thickness sensor 140. Based on the weight parameters, the identification results on both sides are weighted and fused, which can more accurately determine the material category. Since the internal element distribution of materials such as ores is uneven, the X-ray images corresponding to different detection areas may reflect the characteristics of different parts of the material. Therefore, the identification results on one side may differ. This embodiment adaptively adjusts the weights of the recognition results on both sides based on pose information, so that the recognition result on the side with higher imaging quality and higher detection reliability can have a greater influence in the final judgment. This reduces the impact of local feature differences, occlusion or projection changes on the final classification result, and improves the accuracy and robustness of material classification under complex working conditions.

[0052] In some embodiments, step S2431, determining the first material image, the second material image, and the pose information corresponding to the material, includes the following steps.

[0053] Based on a first material image, first feature information is determined, wherein the first feature information includes at least a first timestamp. The first feature information can be determined based on the first material image and can characterize unique information about the material. Specifically, the first feature information can include at least a first timestamp. The first timestamp can represent the time when the first material image corresponding to the material was acquired. Based on the first timestamp, the time when the material arrived at the first identification device 120 can be determined.

[0054] Based on the second material image, second feature information is determined, wherein the second feature information includes at least a second timestamp. The second feature information can be determined based on the second material image and can characterize unique information about the material. Specifically, the second feature information can include at least a second timestamp. The second timestamp can represent the time when the second material image corresponding to the material was acquired. Based on the second timestamp, the time when the material arrived at the second identification device 130 can be determined.

[0055] Based at least on the first timestamp, the second timestamp, and the pose timestamp in the pose information, the first material image, the second material image, and the pose information corresponding to the material are determined. The pose timestamp can be the moment when the thickness sensor 140 detects the pose information of the material. According to the first timestamp, the second timestamp, and the pose timestamp, the first material image, the second material image, and the pose information corresponding to the same material can be determined. Specifically, for the recognition system, the conveying speed of its conveying device 110 is v, and the distance between the two ray sources in the material conveying direction is D. Based on the above data, the theoretical time difference between the arrival of the material at the first recognition device 120 and the second recognition device 130 can be determined, as shown in the following formula: Δt=D / v, where Δt is the theoretical time difference. If the first time stamp t1 and the second time stamp t2 satisfy the following formula: |t2-t1-Δt|<ε, where ε can be a preset time difference threshold, it can be determined that the first material image and the second material image corresponding to each other are images of the same material. Thus, the first material image and the second material image can be associated and bound by time sequence without merging the first material image and the second material image into a complete image. This can effectively save computing power, avoid errors caused by image fusion, and also effectively reduce the amount of computation and improve the efficiency of matching the first material image and the second material image.

[0056] Furthermore, in some embodiments, the first feature information may include a first grayscale feature, and the second feature information may include a second grayscale feature. The first and second grayscale features can be features that characterize the grayscale information of the materials in the first and second material images, such as grayscale histograms, grayscale mean values, grayscale variances, or texture features (e.g., LBP local binary mode features, GLCM grayscale co-occurrence matrix features, etc.). Based on the first and second grayscale features, their feature similarity can be calculated, and a preset similarity threshold can be set. When the feature similarity S_gray between the first and second grayscale features is greater than the preset similarity threshold Th_gray, matching of the first and second material images can be achieved using the first and second grayscale features.

[0057] In some embodiments, the first feature information may include a first shape feature, and the second feature information may include a second shape feature. The first and second shape features may include information describing the shape, such as the material's outline, length, area, aspect ratio, roundness, rectangularity, and Hu moment invariant. Specifically, by calculating the Euclidean distance or cosine similarity between the feature vectors of the first and second shape features, if the Euclidean distance is less than a preset threshold, or the cosine similarity is greater than a preset cosine similarity threshold, it can be determined that the first and second shape features match, and the corresponding first and second material images are images of the same material. Thus, matching of the first and second material images can be achieved using the first and second shape features.

[0058] In some embodiments, the first feature information may include a first height profile feature, and the second feature information may include a second height profile feature. The thickness sensor 140 can acquire the height profile corresponding to the material, and based on the first material image and the second material image, the corresponding first height profile feature and second height profile feature can be determined respectively, and matching can be performed based on the first height profile feature and the second height profile feature to determine whether the two correspond to the same material.

[0059] Furthermore, the matching relationship between the first material image and the second material image can be determined by comparing one or more of the aforementioned first and second feature information. Further, a comprehensive matching score can be determined based on the above information, calculated as follows: S_total = w1 × S_time + w2 × S_gray + w3 × S_shape + w4 × S_height. Where S_total is the comprehensive matching score, S_time is the time difference score, S_gray is the grayscale feature consistency score, S_shape is the shape feature consistency score, and S_height is the height contour consistency score. w1, w2, w3, and w4 are the weights corresponding to each parameter. For parameters with higher reliability, their corresponding weights can be increased. In the above formula, the time difference score has the highest reliability, and the grayscale feature consistency has relatively high reliability, so it can be assigned a larger weight. The shape feature consistency and height contour consistency have relatively lower reliability, so their weights can be assigned smaller values. Specifically, the following set of weights is provided as an example: w1=0.5, w2=0.3, w3=0.1, w4=0.1.

[0060] According to this embodiment, by extracting the first feature information and the second feature information corresponding to the first material image and the second material image respectively, and combining them with the pose information obtained by the thickness sensor 140, an association relationship for the same material is established. This enables the data obtained by the first identification device 120, the second identification device 130, and the thickness sensor 140 to accurately correspond to the same material, providing a reliable data foundation for the fusion of subsequent identification results and improving the accuracy and stability of the overall system identification process. By using the first timestamp, the second timestamp, and the pose timestamp to establish the temporal association relationship of the material, and combining the conveying speed and the relative position between the two identification devices to calculate the theoretical time difference, rapid association of the same material between different identification devices is achieved. This eliminates the need for image stitching of the first material image and the second material image, avoiding complex calculation processes such as image registration, geometric transformation, and image fusion, reducing the system's computational complexity and computational resource consumption. At the same time, it avoids the impact of image stitching errors on subsequent identification results, improving the system's real-time processing capability. Based on multi-dimensional feature information such as grayscale features, shape features, and height contour features, matching the first and second material images allows different types of features to complement and cross-validate each other, effectively improving the accuracy and robustness of matching images corresponding to the same material. This is particularly suitable for complex working conditions such as continuous conveying of multiple materials and small material spacing in ore sorting. It effectively improves the reliability and interpretability of matching the first and second material images.

[0061] In some embodiments, determining weight parameters based on pose information may include the following steps.

[0062] Based on the pose information, the position information of the material is determined, whereby the position information represents the position of the material relative to the width direction of the conveying device 110. First, based on the pose information, the position of the material in the width direction of the conveying device 110 can be determined, i.e., the position information. Specifically, the position of the center of the material, or the position of the highest point of the material, can be determined as the position information.

[0063] Based on location information, weight parameters are determined. If the location information indicates that the material is closer to the first side, then the first weight is greater than the second weight; conversely, if the location information indicates that the material is closer to the second side, then the second weight is greater than the first weight. The location information can be the center of the material or the location of its highest point. If the location information indicates that the material is closer to the first side... Specifically, the pixel range of the material in the data detected by the thickness sensor 140 can be determined, and converted into actual position coordinates according to calibration parameters. The left boundary x_left and right boundary x_right of the material detected by the thickness sensor 140 are calculated, thereby further determining its centroid position x_center. Where x_center = (x_left + x_right) / 2. Subsequently, the width W of the material can be determined as: W = x_right - x_left. The width of the conveying device 110 is D_belt, the center position of the conveying device 110 is D_belt / 2, and the position deviation threshold is δ_pos. Therefore, the material position represented by the material position information can be determined in the following way.

[0064] If x_center < D_belt / 2 - δ_pos, then the material is considered to be significantly to the left, i.e., closer to the first side. Since the detection by the first identification device 120 is more reliable, a larger value can be assigned to its corresponding first weight. Specifically, the first weight can be set to greater than or equal to 0.6 and less than or equal to 0.8, and the second weight can be set to greater than or equal to 0.2 and less than or equal to 0.4.

[0065] If x_center > D_belt / 2 + δ_pos, the material is significantly biased to the right (closer to the second identification device), making the second identification device's detection more reliable. Therefore, a larger value can be assigned to its corresponding second weight. The second weight can be set to greater than or equal to 0.6 and less than or equal to 0.8, and the first weight can be set to greater than or equal to 0.2 and less than or equal to 0.4.

[0066] If |x_center - D_belt / 2| ≤ δ_pos, then the material can be considered centered, and the detection conditions on both sides are similar. Therefore, the first weight and the second weight can be equal, i.e., w1_base = w2_base = 0.5. Here, w1_base is the first weight, and w2_base is the second weight. Thus, the final material category can be determined by averaging the information from the first category and the second category.

[0067] According to this embodiment, the position information of the material in the width direction of the conveying device 110 is determined by the pose information acquired by the thickness sensor 140. The weight parameters corresponding to the first and second category information are adaptively determined based on the position information. This allows the fusion process to dynamically adjust the influence of the two-sided recognition results based on the actual spatial position of the material, rather than using fixed weights or simple averaging, thus improving the accuracy of material category determination. When the position information indicates that the material is closer to the first side, the first weight corresponding to the first category information is increased; when the position information indicates that the material is closer to the second side, the second weight corresponding to the second category information is increased. This allows the recognition result of the side closer to the material and with better imaging conditions to have a greater weight in the final decision. Since the X-ray penetration path is more reasonable and the projection broadening and occlusion effects are relatively smaller when the material is biased to one side, the corresponding recognition result usually has higher reliability. By dynamically adjusting the fusion weights based on the actual position of the material, the differences in imaging quality between the two sides can be fully utilized, improving the accuracy of the final material classification and the reliability of the recognition results.

[0068] In some embodiments, determining the weight parameters based on pose information may further include the following steps.

[0069] Based on the pose information, the height information of the material is determined, where the height information represents the material's thickness. First, the material's thickness can be determined based on the pose information, and then characterized by the height information determined by the thickness sensor 140. Subsequently, the weighting parameters are adjusted based on the height information to avoid the influence of the material's thickness on material category detection.

[0070] Based on height information, the weight parameters are updated. If the height information indicates that the material thickness exceeds a threshold, the difference between the first weight and the second weight is increased. When the material is too thick, the portion within the overlapping area may project during image acquisition by the first recognition device 120 and the second recognition device 130, affecting the accuracy of the first and second images. The height of the material can affect the clarity and broadening of the X-ray projection. Therefore, the weight parameters can be updated based on height information. Specifically, height thresholds h_low and h_high can be preset, and the weight parameters can be updated according to the following method.

[0071] If h_max≤ h_low, then the material is considered to be low-profile, with a small projection width and similar clarity on both sides, and a height adjustment coefficient α_h=1.0 is assigned.

[0072] If h_low < h_max ≤ h_high, then the material can be considered to be of medium height, with a certain degree of projection widening. The source detection is clearer on the side closer to the material. The weight can be adjusted according to the position, and the height adjustment coefficient can be appropriately increased. For example, the height adjustment coefficient of the recognition device on the side closer to the material can be adjusted to α_h=1.1.

[0073] If h_max > h_high, the material is considered tall and thick with significant projection widening. The detection result on the side closer to the material must be selected first. The height adjustment coefficient α_h on the corresponding side can be adjusted to be greater than or equal to 1.2 and less than or equal to 1.5.

[0074] Finally, based on the original first and second weights, and combined with the height adjustment coefficient, the first and second weights are updated as shown in the following formula: w1 = w1_base × α_h1 / (w1_base×α_h1 + w2_base×α_h2) w2 = 1 - w1 Where w1 is the updated first weight and w2 is the updated second weight.

[0075] Finally, when the material categories represented by the first category information and the second category information are inconsistent, the material category can be determined based on the following method. Here, w1 is the first weight, w2 is the second weight, P1 is the confidence level of the first category information, and P2 is the confidence level of the second category information.

[0076] If w1×P1 >> w2×P2, then the first category information can be selected, and the material category corresponding to the first category information will be output.

[0077] If w2×P2 >> w1×P1, then the second category of information can be selected, and the material category corresponding to the second category of information will be output.

[0078] If the weighted confidence levels on both sides are similar, the importance of the category can be further considered based on the sorting environment and sorting requirements. For example, in the ore sorting process, the failure to detect gangue will lead to a reduction in the time and efficiency of ore sorting and a higher cost. In this case, a conservative strategy can be adopted, classifying the uncertain ore material as gangue, or directly outputting an "uncertain" alarm, and further determining the category of the ore material through manual review.

[0079] According to this embodiment, the height information of the material is determined based on the pose information acquired by the thickness sensor 140. The first and second weights are adaptively updated according to the material height, so that the fused weights not only reflect the position of the material in the width direction of the conveying device 110, but also dynamically adjust the bilateral recognition results based on the material's thickness. This achieves collaborative decision-making between position and height information, effectively improving adaptability to materials of different shapes and maintaining high detection accuracy for all materials. When the material thickness is small, the weights on both sides remain basically consistent due to the small difference in X-ray imaging quality on both sides. As the material thickness increases, the weight corresponding to the recognition result closer to the material is gradually increased, allowing the fused weights to reflect the influence of material height on X-ray projection broadening and imaging quality. This avoids the problem of fixed weights being unable to adapt to materials of different thicknesses, improving the reliability of recognition results under complex working conditions. For thicker materials, since the X-ray projection broadening is more pronounced, the side farther from the material is more susceptible to projection distortion, occlusion, and other factors. This embodiment increases the weight of the recognition result corresponding to the side closer to the material, giving the side with better imaging quality a greater influence in the final recognition decision. This reduces misjudgments caused by projection broadening and imaging differences, thereby improving the classification accuracy of materials with greater thickness. According to this embodiment, the accuracy of material classification, the reliability of recognition results, and the system's adaptability to complex working conditions can be effectively improved.

[0080] Based on the same inventive concept, this disclosure also provides a sorting machine, including: an identification system and a sorting device as described in any of the foregoing embodiments.

[0081] An identification system is used to determine the category of materials. The identification system detects materials on the conveying device 110 and determines their corresponding category. Specifically, the identification system can acquire a first material image collected by the first identification device 120, a second material image collected by the second identification device 130, and pose information acquired by the thickness sensor 140. Based on the position and area traversed by the material during conveying, it employs corresponding identification and classification strategies, and can determine the material category based on the first and second material images combined with pose information, achieving high accuracy.

[0082] The sorting device, located downstream of the conveying device 110 of the identification system, is used to sort materials based on the categories determined by the identification system. When materials are conveyed to the sorting position, the sorting device performs the corresponding sorting action according to the material category determined by the identification system, allowing different categories of materials to enter their respective collection areas, thus achieving automatic material sorting. The sorting device can employ an air-blowing sorting mechanism, a mechanical feeding mechanism, a baffle-type sorting mechanism, or other sorting mechanisms capable of performing sorting actions according to material category.

[0083] According to this embodiment, the reliability of material category determination is effectively improved, enabling the sorting device to perform corresponding sorting actions based on more reliable identification results, thereby further improving the sorting accuracy, operational reliability, and adaptability under complex working conditions of the entire sorting machine.

[0084] This disclosure uses specific terms to describe embodiments of the present disclosure. Terms such as "an embodiment," "one embodiment," and / or "some embodiments" refer to a particular feature, structure, or characteristic associated with at least one embodiment of the present disclosure. Therefore, it should be emphasized and noted that references to "an embodiment," "one embodiment," or "an alternative embodiment" in different locations throughout this specification do not necessarily refer to the same embodiment. Furthermore, certain features, structures, or characteristics in one or more embodiments of the present disclosure can be appropriately combined.

[0085] In the context of this disclosure, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" do not specifically refer to the singular and may also include the plural. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of expressly identified steps and elements, which do not constitute an exclusive list, and the method or apparatus may also include other steps or elements.

[0086] Similarly, it should be noted that, in order to simplify the description of this disclosure and thus aid in the understanding of one or more embodiments, the foregoing description of embodiments of this disclosure may sometimes combine multiple features into a single embodiment, drawing, or description thereof. However, this disclosure method does not imply that the subject matter of this disclosure requires more features than the features claimed. In fact, the embodiments contain fewer features than all the features of the single embodiments disclosed above.

[0087] The basic concepts have been described above. It is obvious that the above disclosure is merely illustrative and does not constitute a limitation of this disclosure. Although not explicitly stated herein, various modifications, improvements, and corrections may be made to this disclosure by those skilled in the art. Such modifications, improvements, and corrections are suggested in this disclosure and therefore remain within the spirit and scope of the embodiments of this disclosure.

Claims

1. A recognition method applied to a recognition system, wherein, The identification system includes: a conveying device for conveying materials to be identified; a first identification device for acquiring a first image of a first range of the conveying device to obtain a first material image of the material passing through the first range, wherein the first range is the range from a first side to the middle of the conveying device in the width direction; a second identification device for acquiring a second image of a second range of the conveying device to obtain a second material image of the material passing through the second range, wherein the second range is the range from a second side to the middle of the conveying device in the width direction, the second side being the opposite side to the first side, and the second range and the first range have an overlapping range in the width direction; a thickness sensor disposed above the conveying device, for at least determining the pose information of the material passing through the overlapping range; and a controller communicatively connected to the first identification device, the second identification device, and the thickness sensor, for determining the category of the material based on the first material image, the second material image, and the pose information, characterized in that the identification method includes: The first image is acquired by the first recognition device; The second image is acquired by the second recognition device; The thickness sensor is used to determine the position and orientation information of the material passing through the overlapping range. Based on the first image, the second image, and the pose information, the category of the material is determined; Determining the category of the material based on the first image, the second image, and the pose information includes: If the material passes through the first range but not the overlapping range, then the material category is determined based on the first material image; If the material passes through the second range but not the overlapping range, then the material category is determined based on the second material image; If the material passes through the overlapping area, then the material category is determined based on the first category determined according to the first material image, the second category determined according to the second material image, and the pose information; The process of determining the material category based on the first category determined from the first material image, the second category determined from the second material image, and the pose information includes: Determine the first material image, the second material image, and the pose information corresponding to the material; Based on the first material image and the second material image, the corresponding first category information and second category information are determined respectively; If the material categories represented by the first category information and the second category information are consistent, then the category of the material is determined; If the material categories represented by the first category information and the second category information are inconsistent, then: based on the pose information, a weight parameter is determined, wherein the weight parameter includes a first weight corresponding to the first category information and a second weight corresponding to the second category information; and the category of the material is determined based on the weight parameter, the first category information, and the second category information.

2. The identification method as described in claim 1, characterized in that, The step of determining the first material image, the second material image, and the pose information corresponding to the material includes: Based on the first material image, first feature information is determined, wherein the first feature information includes at least a first timestamp; Based on the second material image, second feature information is determined, wherein the second feature information includes at least a second timestamp; Based at least on the first timestamp, the second timestamp, and the pose timestamp in the pose information, the first material image, the second material image, and the pose information corresponding to the material are determined.

3. The identification method as described in claim 1, characterized in that, The step of determining the weight parameters based on the pose information includes: Based on the pose information, the position information of the material is determined, wherein the position information represents the position of the material relative to the width direction of the conveying device; Based on the location information, the weight parameters are determined, wherein if the location information indicates that the material is close to the first side, the first weight is greater than the second weight, and if the location information indicates that the material is close to the second side, the second weight is greater than the first weight.

4. The identification method as described in claim 3, characterized in that, The step of determining the weight parameters based on the pose information further includes: Based on the pose information, the height information of the material is determined, wherein the height information represents the thickness of the material; Based on the height information, the weight parameters are updated, wherein if the height information indicates that the thickness of the material exceeds a threshold, the difference between the first weight and the second weight is increased.

Images