An image processing method, apparatus and transmission scanning system

Abstract

The disclosure provides an image processing method, device and transmission scanning system, and relates to the technical field of radiation imaging. The image processing method comprises: performing sampling processing on a transmission image to obtain a to-be-processed image, the transmission image being obtained by detecting transmission intensity when an object is subjected to transmission scanning by using a plurality of detectors; determining a to-be-identified region from the to-be-processed image; performing reconstruction processing on the to-be-identified region by using a candidate depth to obtain a reconstruction result of the to-be-identified region; determining edge definition of the reconstruction result; and in the case that the edge definition is greater than a definition threshold, taking the candidate depth as a depth of the to-be-identified region. Through the above method, the depth information of the object can be obtained without adjusting the detector structure, which helps to improve the efficiency and identification effect of the object inspection.

Description

Technical Field

[0001] This disclosure relates to the field of radiation imaging technology, and in particular to an image processing method, apparatus and transmission scanning system. Background Technology

[0002] X-ray transmission inspection technology examines objects by measuring the attenuation of X-ray intensity after it passes through them. X-ray transmission inspection technology has advantages such as strong penetration, short measurement time, and high image resolution, and is widely used in cargo inspection at customs, airports, and other locations.

[0003] A transmission image of an object contains projection information of the object along one or more ray irradiation directions. Due to information overlap along the ray irradiation directions, it is difficult to directly extract the object's depth information from the transmission image, which causes inconvenience for inspection.

[0004] In related technologies, scanning equipment, including a sparse array detector, is used to perform transmission scanning on an object to obtain scanning data. Then, by reconstructing the scanning data and filtering out defocused pixel values, the object's depth information is identified. Summary of the Invention

[0005] This disclosure provides an image processing method, apparatus, and transmission scanning system.

[0006] According to a first aspect of this disclosure, an image processing method is proposed, comprising: sampling a transmission image to obtain an image to be processed, wherein the transmission image is obtained by detecting the transmission intensity when an object is scanned using a multi-array detector; determining a region to be identified from the image to be processed; reconstructing the region to be identified using a candidate depth to obtain a reconstruction result of the region to be identified; determining the edge sharpness of the reconstruction result; and, if the edge sharpness is greater than a sharpness threshold, using the candidate depth as the depth of the region to be identified.

[0007] In some embodiments, the transmitted image includes images acquired by the multi-column detector under each of multiple pulses. The sampling processing of the transmitted image to obtain the image to be processed includes: extracting K pixels from each row of pixels in the image acquired under each pulse to obtain a pixel sequence corresponding to the image acquired under each pulse, where K is an integer greater than or equal to 1 and less than N, and N is the number of columns of the crystal unit of the multi-column detector; and splicing the pixel sequence corresponding to the image acquired under each of the multiple pulses to obtain the image to be processed.

[0008] In some embodiments, K is 1, and extracting K pixels from each row of pixels in the image acquired under each pulse includes: extracting the q-th pixel from the i-th row of the image acquired under each pulse according to the element value q of the i-th element in the sampling sequence, where i is any integer from 1 to M, M is the number of rows of the crystal unit of the multi-column detector, and q is an integer greater than or equal to 1 and less than or equal to N.

[0009] In some embodiments, the sampling sequence includes a first sampling sequence in which element values ​​are determined based on a random function.

[0010] In some embodiments, the sampling sequence includes a second sampling sequence, wherein the arrangement of pixels extracted by the second sampling sequence conforms to a specified pattern.

[0011] In some embodiments, determining the edge sharpness of the reconstruction result includes: extracting edge pixels of the reconstruction result using an edge extraction algorithm; and determining the edge sharpness of the reconstruction result based on the edge pixels.

[0012] In some embodiments, the edge sharpness of the reconstruction result is determined based on the edge pixels: the number of edge pixels contained in a specified neighborhood of each pixel in the reconstruction result is calculated; the number of pixels whose number of edge pixels is less than a quantity threshold is calculated; and the edge sharpness of the reconstruction result is determined based on the number of pixels whose number of edge pixels is less than the quantity threshold and the total number of edge pixels in the reconstruction result.

[0013] In some embodiments, determining the edge sharpness of the reconstruction result based on the number of pixels whose number of edge pixels is less than a quantity threshold and the total number of edge pixels in the reconstruction result includes: using the ratio of the number of pixels whose number of edge pixels is less than a quantity threshold to the total number of edge pixels in the reconstruction result as the edge sharpness of the reconstruction result.

[0014] In some embodiments, the reconstruction process of the region to be identified using candidate depth includes: determining a focusing ratio based on the distance from the multi-column detectors to the X-ray source and the candidate depth; and focusing the coordinate components of the pixels in the region to be identified in a specified direction based on the focusing ratio, wherein the specified direction is parallel to the scanning direction of the object.

[0015] In some embodiments, determining the focusing ratio based on the distance from the multi-row detectors to the radiation source and the candidate depth includes: calculating the difference between the distance from the multi-row detectors to the radiation source and the candidate depth; and using the ratio of the difference to the distance from the multi-row detectors to the radiation source as the focusing ratio.

[0016] In some embodiments, the image processing method further includes: when the edge sharpness is less than or equal to the sharpness threshold, iteratively performing a candidate depth update, region reconstruction based on the updated candidate depth, and an edge sharpness determination step, until the edge sharpness is greater than the sharpness threshold.

[0017] In some embodiments, the image processing method further includes: reconstructing the region to be identified based on the depth of the region to be identified, so as to obtain a reconstructed image.

[0018] In some embodiments, the image processing method further includes: displaying the reconstructed image and the depth information of the region to be identified.

[0019] According to a second aspect of this disclosure, an image processing apparatus is provided, comprising: a memory; and a processor coupled to the memory, the processor being configured to perform the image processing method as described above based on instructions stored in the memory.

[0020] According to a third aspect of this disclosure, a transmission scanning system is proposed, comprising: an image processing device as described above; a multi-row detector; and a radiation source.

[0021] According to a fourth aspect of this disclosure, a computer-readable storage medium is provided that stores computer program instructions thereon, which, when executed by a processor, implement the image processing method as described above.

[0022] According to a fifth aspect of this disclosure, a computer program product is provided having computer program instructions stored thereon, which, when executed by a processor, implement the image processing method as described above.

[0023] Other features and advantages of this disclosure will become clear from the following detailed description of exemplary embodiments with reference to the accompanying drawings. Attached Figure Description

[0024] The accompanying drawings, which form part of this specification, illustrate embodiments of this disclosure and, together with the specification, serve to explain the principles of this disclosure.

[0025] This disclosure can be more clearly understood with reference to the accompanying drawings and the following detailed description.

[0026] Figure 1A schematic diagram illustrating the principle of including depth information in a transmission image.

[0027] Figure 2 This is a schematic flowchart of an image processing method according to some embodiments of the present disclosure.

[0028] Figure 3 This is a schematic diagram illustrating the principle of image sampling according to some embodiments of the present disclosure.

[0029] Figure 4 This is a schematic diagram illustrating the principle of image sampling according to other embodiments of this disclosure.

[0030] Figure 5 This is a schematic flowchart of a region reconstruction process according to some embodiments of the present disclosure.

[0031] Figure 6 This is a schematic diagram of the imaging principle based on a multi-array detector according to some embodiments of the present disclosure.

[0032] Figure 7 This is a schematic diagram of a process for determining edge sharpness according to some embodiments of the present disclosure.

[0033] Figure 8a This is a schematic diagram of a neighborhood template according to some embodiments of the present disclosure.

[0034] Figure 8b This is a schematic diagram of a sharp-edge image according to some embodiments of the present disclosure.

[0035] Figure 8c This is a statistical histogram of the number of edge pixels in the neighborhood corresponding to a clear edge image according to some embodiments of the present disclosure.

[0036] Figure 8d This is a schematic diagram of a non-sharp edge image according to some embodiments of the present disclosure.

[0037] Figure 8e This is a statistical histogram of the number of edge pixels in the neighborhood corresponding to a non-sharp edge image according to some embodiments of the present disclosure.

[0038] Figure 9 This is a schematic flowchart of an image processing method according to other embodiments of the present disclosure.

[0039] Figure 10a This is a schematic diagram of a transmission image according to some embodiments of the present disclosure.

[0040] Figure 10b This is a schematic diagram of an image after sampling processing according to some embodiments of the present disclosure.

[0041] Figure 10cThis is a schematic diagram of an image processed according to a depth reconstruction according to some embodiments of the present disclosure.

[0042] Figure 10d This is a schematic diagram of an image processed according to another depth reconstruction according to some embodiments of the present disclosure.

[0043] Figure 10e This is a schematic diagram of an image with depth information added according to some embodiments of the present disclosure.

[0044] Figure 11 This is a schematic diagram of the structure of an image processing apparatus according to some embodiments of the present disclosure.

[0045] Figure 12 This is a schematic diagram of the structure of a transmission scanning system according to some embodiments of the present disclosure.

[0046] Figure 13 This is a schematic diagram of the structure of a computer system according to some embodiments of the present disclosure. Detailed Implementation

[0047] Various exemplary embodiments of the present disclosure will now be described in detail with reference to the accompanying drawings. It should be noted that, unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps set forth in these embodiments do not limit the scope of the present disclosure.

[0048] At the same time, it should be understood that, for ease of description, the dimensions of the various parts shown in the accompanying drawings are not drawn according to actual scale.

[0049] The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit this disclosure or its application or use.

[0050] Techniques, methods, and equipment known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and equipment should be considered part of the specification.

[0051] In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values.

[0052] It should be noted that similar labels and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be discussed further in subsequent figures.

[0053] To make the objectives, technical solutions, and advantages of this disclosure clearer, the following detailed description is provided in conjunction with specific embodiments and the accompanying drawings.

[0054] The inventors of this disclosure have discovered the following shortcomings in the related technology: First, the related technology uses a sparse array detector, which requires significant modifications to the detector structure in the scanning system, and the sparse array detector is difficult to manufacture, install, and fix; Second, the related technology has not conducted further research on how to accurately extract the depth information of the object based on image processing algorithms.

[0055] In view of this, this disclosure proposes an image processing method that can obtain the depth information of an object without adjusting the conventional detector structure, which helps to improve the efficiency and recognition effect of object inspection.

[0056] Figure 1 This is a schematic diagram illustrating the principle of including depth information in a transmission image. (Example) Figure 1 As shown, a transmission image of an object is obtained by scanning using multiple rows of detectors. These multiple rows of detectors, also known as detector matrices or multi-row detectors, contain at least two rows of detector crystals (or two rows of crystal units). The transmission image can be, for example, a DR (Digital Radiography) image.

[0057] like Figure 1 As shown, assume multiple detectors are placed vertically, with their column directions parallel to the detector centerline 102 and their row directions parallel to the line indicated by the arrow. The central ray emitted by the ray source 101 passes through points A0 and B0 at different depths on the object and is received by the same crystal D0 on the detector centerline 102. In this case, the projection information of the object detected by the detector in different depth directions overlaps, making it difficult to detect the object's depth information from this overlapping information.

[0058] Another ray emitted by X-ray source 101 forms an angle α with the central ray. This ray passes through point A1 on the object and is detected by crystal D on the detector. A Received. Simultaneously, another ray from ray source 101 passes through point B1 on the object and is detected by crystal D on the detector. C Reception. This shows that, due to the different angles of the rays emitted by the X-ray source, points A1 and B1 at different depths on the object are received by crystals at different positions on the detector (i.e., crystals with different horizontal displacements from the detector's centerline 102), and not by the same crystal D. B The received data will be transmitted through the multi-array detectors (specifically, detectors with horizontal displacement) to contain depth information of the object. The following will explain how to extract the depth information of the object from the transmitted image, in conjunction with an embodiment.

[0059] Figure 2 This is a schematic flowchart illustrating an image processing method according to some embodiments of the present disclosure. Figure 2 As shown, the image processing method includes steps S201 to S205.

[0060] In step S201, the transmission image is sampled to obtain the image to be processed.

[0061] In some embodiments, the image processing method is performed by the image processing apparatus described later.

[0062] The transmission image is obtained by detecting the transmission intensity when an object is scanned using multiple detectors. For example, a DR (Digital Radiography) scanning device, which includes an X-ray source and multiple detectors, is used to scan an object to obtain a transmission image of the object.

[0063] In some embodiments, the transmission image includes images acquired by the multi-column detector under each of multiple pulses. In these embodiments, the transmission image is sampled as follows: K pixels are extracted from each row of pixels in the image acquired under each pulse to obtain a pixel sequence corresponding to the image acquired under each pulse; the pixel sequences corresponding to each pulse in the multiple pulses are concatenated to obtain the image to be processed. Here, K is an integer greater than or equal to 1 and less than N, and N is the number of columns of the crystal units of the multi-column detector.

[0064] In some examples, K is 1. In these examples, a pixel is extracted from each row of pixels in the image as follows: based on the element value q of the i-th element in the sampling sequence, the q-th pixel is extracted from the i-th row of the image acquired under each pulse. Here, i is any integer from 1 to M, M is the number of rows of the crystal unit in the multi-column detector, and q is an integer greater than or equal to 1 and less than or equal to N. Then, the extracted pixels are concatenated sequentially according to the extraction order to obtain the pixel sequence corresponding to the image acquired under each pulse. Next, the pixel sequences corresponding to multiple pulses are concatenated to obtain the image to be processed.

[0065] For example, in a multi-column detector array with 10 columns and 5 rows, and a sampling sequence of (1, 3, 5, 7, 10), the first pixel is extracted from the first row of pixels in the image acquired under the first pulse, based on the element value of the first element (1). The third pixel is extracted from the second row of pixels in the image acquired under the first pulse, based on the element value of the second element (3). The fifth pixel is extracted from the third row of pixels in the image acquired under the first pulse, based on the element value of the third element (5). The seventh pixel is extracted from the fourth row of pixels in the image acquired under the first pulse, based on the element value of the fourth element (7). The tenth pixel is extracted from the fifth row of pixels in the image acquired under the first pulse, based on the element value of the fifth element (10). These five extracted pixels are then concatenated to obtain the pixel sequence corresponding to the image acquired under the first pulse.

[0066] In some examples, the sampling sequence includes a first sampling sequence, in which the element values ​​are determined based on a random function. In these examples, the first sampling sequence is used for sampling images acquired under different pulses.

[0067] In some examples, such as Figure 3 As shown, assuming the multi-column detector is an M-row, N-column crystal array, a random function generates an integer sequence S (i.e., the first sampling sequence) with values ​​ranging from 1 to N. This integer sequence S includes M elements, each of which is a random integer between 1 and N. For example, assuming the multi-column detector is a 4-row, 5-column crystal array, the integer sequence S includes 4 elements, whose values ​​can be 1, 2, 5, 3, and 5 respectively. Using the integer sequence S, the image 301 acquired under pulse 1 is sampled to obtain pixel sequence 1; using the integer sequence S, the image 302 acquired under pulse 2 is sampled to obtain pixel sequence 2. This process continues, sampling is performed on images acquired under multiple pulses to obtain multiple pixel sequences. Then, multiple pixel sequences, including pixel sequence 1 and pixel sequence 2, are concatenated to obtain the image to be processed. It should be noted that pixel sequence 1 and pixel sequence 2 consist of pixels containing horizontal displacement relative to the detector centerline 102, thus carrying horizontal displacement information.

[0068] In some examples, the sampling sequence includes a second sampling sequence, the second sampling sequence extracting pixels arranged according to a specified pattern. In these examples, the second sampling sequence is used for sampling images acquired under different pulses.

[0069] In some examples, the second sampling sequence is a sine sequence, an m-sequence, or other sequences. Among them, the sine sequence is... Figure 4The extracted pixels are arranged in a sinusoidal sampling sequence. The m-sequence refers to a sampling sequence where the extracted pixels are arranged in an m-shape. In practice, after obtaining the original transmission image, different sampling sequences can be used to process the transmission images of objects with different contours. For example, for contours resembling straight lines or diagonal lines, random sequences or m-sequences are more effective.

[0070] For example, such as Figure 4 As shown, assuming the multi-column detector is a crystal array with M rows and N columns, using coincidence... Figure 4 The second sampling sequence (containing M elements) of the specified pattern is used to sample image 301 acquired under pulse 1 to obtain pixel sequence 1; the second sampling sequence is then used to sample image 302 acquired under pulse 2 to obtain pixel sequence 2. This process is repeated for images acquired under multiple pulses to obtain multiple pixel sequences. These multiple pixel sequences, including pixel sequence 1 and pixel sequence 2, are then concatenated to obtain the image to be processed.

[0071] In some examples, the sampling sequence includes a first sampling sequence and a second sampling sequence. In these examples, the first sampling sequence is used to sample images acquired under a portion of multiple pulses; the second sampling sequence is used to sample images acquired under another portion of multiple pulses.

[0072] In this embodiment of the disclosure, considering that the projection information of different rays passing through different depths of an object overlaps, making it difficult to extract depth information, the transmission image is sampled. In this way, the information collected by the crystal unit at different positions in the detector can be highlighted, which is convenient for subsequent depth recognition.

[0073] In step S202, the region to be identified is determined from the image to be processed.

[0074] In some embodiments, the image to be processed by step S201 includes multiple objects. In these embodiments, the regions where each object is located are segmented from the image to be processed to serve as multiple regions to be identified. For example, the regions where different objects are located are segmented based on features such as image grayscale, geometric shape, and spatial texture.

[0075] In some embodiments, the image to be processed includes multiple objects. In these embodiments, the region containing the object of interest is segmented from the image to be processed as the region to be identified.

[0076] In some embodiments, the same object in the image to be processed is segmented into multiple regions to be identified. For example, for a vehicle in the image, the vehicle is segmented into two regions to be identified: the front and the rear.

[0077] In step S203, the candidate depth is used to reconstruct the region to be identified in order to obtain the reconstruction result of the region to be identified.

[0078] Among them, the process of reconstructing the region to be identified using candidate depth can also be referred to as focusing the region to be identified using candidate depth.

[0079] In some embodiments, there are multiple regions to be identified, and the same range of candidate depth values ​​is set for different regions to be identified.

[0080] For example, the range of candidate depth values ​​can be determined based on the distance between the X-ray source and the detector in the scanning system. Assuming the distance from the X-ray source to the detector is 7 meters, the range of candidate depth values ​​for each region to be identified can be set to 1–6 meters. For instance, the initial candidate depth value could be 1 meter, the increment step could be 0.1 meters, and the maximum candidate depth value could be 6 meters.

[0081] For example, the range of candidate depth values ​​can be determined based on the object's size estimation information. Assuming the object is a container, the range of candidate depth values ​​for each region to be identified can be set to 1 to 3 meters. For instance, the initial value of the candidate depth can be 1 meter, the increment step can be 0.2 meters, and the maximum value of the candidate depth can be 3 meters.

[0082] In some embodiments, there are multiple regions to be identified, and different candidate depths are set for different regions to be identified.

[0083] In some embodiments, in step S203, according to Figure 5 The process shown reconstructs the region to be identified.

[0084] In step S204, the edge sharpness of the reconstruction result is determined.

[0085] Edge sharpness is an indicator used to characterize whether the edges of the reconstructed image region are sharp.

[0086] In some embodiments, according to Figure 7 The process shown determines the edge sharpness of the reconstructed image. Specifically, it determines edge sharpness by statistically analyzing the features of edge pixels in the reconstructed image.

[0087] In step S205, if the edge sharpness is greater than the sharpness threshold, the candidate depth is used as the depth of the region to be identified.

[0088] After determining the edge sharpness of the reconstruction result in step S204, the edge sharpness is compared with a sharpness threshold. If the edge sharpness is greater than the sharpness threshold, the candidate depth is taken as the depth of the region to be identified.

[0089] In some embodiments, the image processing method further includes: when the edge sharpness is less than or equal to a sharpness threshold, iteratively performing a candidate depth update, region reconstruction based on the updated candidate depth, and an edge sharpness determination step, until the edge sharpness is greater than the sharpness threshold.

[0090] For example, when the candidate depth is 1 meter, steps S203 and S204 are executed to obtain the edge sharpness. If the edge sharpness obtained based on the candidate depth of 1 meter is less than or equal to the sharpness threshold, the candidate depth is updated. For example, the candidate depth is increased from 1 meter to 1.1 meters, with a step size of 0.1 meters. Then, for the candidate depth of 1.1 meters, steps S203 and S204 are executed to obtain the new edge sharpness. If the edge sharpness obtained based on the candidate depth of 1.1 meters is greater than the sharpness threshold, then 1.1 meters is taken as the depth of the region to be identified.

[0091] In this embodiment of the disclosure, by sampling the transmission image, reconstructing the region to be identified based on the candidate depth, and determining the depth of the region to be identified based on the edge sharpness of the reconstructed region, the depth information of the object can be extracted accurately and efficiently without adjusting the detector structure, thereby helping to improve the efficiency and recognition effect of object inspection.

[0092] Figure 5 This is a schematic diagram illustrating the process of region reconstruction according to some embodiments of this disclosure. Figure 5 As shown, the region reconstruction process includes steps S501 and S502.

[0093] In step S501, the focusing ratio is determined based on the distance from the multiple detectors to the X-ray source and the candidate depth.

[0094] In some embodiments, there is one region to be identified. The focusing ratio is calculated based on the candidate depth corresponding to the region to be identified.

[0095] In some embodiments, there are multiple regions to be identified, and each region has a corresponding candidate depth. The candidate depths for different regions to be identified can be the same or different. In these embodiments, for each region to be identified, the corresponding focus ratio is calculated using its corresponding candidate depth.

[0096] In some embodiments, the focusing ratio is determined by calculating the difference between the distance from the multi-array detectors to the X-ray source and the candidate depth; the ratio of this difference to the distance from the multi-array detectors to the X-ray source is used as the focusing ratio.

[0097] In step S502, the coordinate components of the pixels in the region to be identified in the specified direction are focused according to the focusing ratio.

[0098] The specified direction is parallel to the scanning direction of the object. For example, such as... Figure 1 As shown, the scanning direction of the object is parallel to D0D. A The direction of the line in question. (DOD) A The line and the detector centerline 102 together form the detector plane 104.

[0099] In some embodiments, focusing is performed as follows: the coordinate components of the pixels in the region to be identified in a specified direction are multiplied by the focusing ratio to obtain an offset; the coordinate components of the pixels in the region to be identified in the specified direction are translated according to the offset to obtain the focused coordinate components.

[0100] In this embodiment of the disclosure, the above steps achieve the reconstruction of the region to be identified based on candidate depth. This facilitates the subsequent determination of the depth of the region to be identified based on the edge sharpness of the reconstruction result.

[0101] The following combination Figure 6 The reconstruction process will be further explained. For example... Figure 6 As shown, it is assumed that the scanning direction of object 103 is parallel to the X-axis, for example, object 103 moves along a direction parallel to the X-axis, or the detector and the X-ray source move along a direction parallel to the X-axis. The multi-row detector includes multiple rows of crystal units, and the multiple rows of crystal units are distributed along the X-axis direction on the detector plane 104. Here, it is assumed that the detector plane 104 is set to be perpendicular to the horizontal plane.

[0102] Suppose that a certain column of crystal cells has a certain horizontal displacement ΔD relative to the detector centerline 102 (if the detector centerline is taken as the vertical axis, it is equivalent to the horizontal coordinate of the pixel point corresponding to that column of crystal cells in the image). For a crystal cell with a horizontal displacement ΔD, point A on the plane at a distance Z from the detector plane will appear in the image lags behind by a distance Δx compared to when the crystal cell is located on the centerline 102. Therefore, shifting the scan line obtained from this crystal cell forward by Δx will align it with the centerline. For another crystal cell, point A2 on the plane at a distance Z from the detector plane will appear in the image a certain distance earlier than when the crystal cell is located on the centerline 102. Therefore, shifting the scan line obtained from this crystal cell backward by a certain distance will align it with the centerline. Furthermore, during image focusing, each pixel is shifted backward by a distance Δx, thereby obtaining a focused image with depth information.

[0103] Furthermore, assuming the distance from point A on object 103 to the detector is Z, and the distance from the radiation source 101 to the detector is L, based on the similarity relationship of triangles, the following relationship can be obtained:

[0104]

[0105] As can be seen from the above formula, when the horizontal displacement ΔD of the crystal unit from the detector centerline is fixed, the offset Δx of the imaging position is different for object points at different depths on the same projection ray emitted by the X-ray source.

[0106] During the focusing process, the translation distance Δx is calculated based on a specific depth position Z, and therefore is only effective for objects near that depth Z. That is, for a given candidate depth, according to the above formula, the pixel coordinates in the sampled image to be processed are aligned with the center line according to the focusing ratio (LZ) / L, based on the detector center position, thereby obtaining the focused image at the given candidate depth position, i.e., the reconstruction result.

[0107] Figure 7 This is a schematic diagram illustrating a process for determining edge sharpness according to some embodiments of the present disclosure. Figure 7 As shown, the process for determining edge sharpness includes steps S701 and S702.

[0108] In step S701, the edge pixels of the reconstruction result are extracted using an edge extraction algorithm.

[0109] In some embodiments, gradient calculation operators such as the Roberts operator, Prewitt operator, or Sobel operator are used to extract edge pixels from the reconstructed results.

[0110] In some embodiments, the Laplace operator, the Canny operator, or a sub-pixel edge detection algorithm based on the Facet model are used to extract the edge pixels of the reconstructed result.

[0111] In step S702, the edge sharpness of the reconstruction result is determined based on the edge pixels.

[0112] In some embodiments, edge sharpness is determined as follows: the number of edge pixels contained in a specified neighborhood for each pixel in the reconstruction result is calculated; the number of pixels whose number of edge pixels is less than a quantity threshold is calculated; and the edge sharpness of the reconstruction result is determined based on the number of pixels whose number of edge pixels is less than the quantity threshold and the total number of edge pixels in the reconstruction result. The quantity threshold can be 2, 3, or other values.

[0113] In some examples, the neighborhood is specified as an eight-point neighborhood, a twelve-point neighborhood, or other styles of neighborhood. For example, using... Figure 8a The neighborhood template shown counts the number of edge pixels in the neighborhood of each pixel.

[0114] In some examples, after counting the number of edge pixels less than the threshold and the total number of edge pixels in the reconstruction result, the ratio of the number of edge pixels less than the threshold to the total number of edge pixels in the reconstruction result is used as the edge sharpness of the reconstruction result.

[0115] For example, assuming the number threshold is 2, there are 50 edge pixels in the reconstructed area, and there are 20 pixels whose neighborhoods contain fewer than 2 edge pixels, then the edge sharpness is 0.4.

[0116] For example, such as Figures 8b to 8e As shown, when the image edges are sharp, the histogram of the number of edge pixels in its neighborhood is concentrated in the region with fewer edge pixels; while when the image edges are not sharp, the histogram of the number of edge pixels in its neighborhood does not show obvious clustering. Therefore, the edge sharpness determined in the above manner can accurately determine whether the image edges are sharp.

[0117] In this embodiment of the disclosure, by using factors such as the number of pixels with fewer than a certain threshold in the neighborhood edge pixels and the total number of edge pixels in the reconstruction result, the edge sharpness of the reconstructed region can be better characterized, thereby enabling the accurate determination of the depth of the region to be identified.

[0118] Figure 9 This is a schematic flowchart illustrating an image processing method according to other embodiments of the present disclosure. Figure 9As shown, the image processing method includes steps S901 to S909.

[0119] In step S901, the transmission image is sampled to obtain the image to be processed.

[0120] In some embodiments, in DR scanning mode, a scanning device including multiple rows of detectors is used to scan the object at an appropriate speed to obtain a transmission image of the object. For example, the obtained transmission image is as follows: Figure 10a As shown.

[0121] In some embodiments, the transmission image includes images acquired by a multi-column detector under each pulse of a plurality of pulses. In these embodiments, one pixel is extracted from each row of pixels in the image acquired under each pulse to obtain a pixel sequence corresponding to the image acquired under each pulse; the pixel sequences corresponding to the images acquired under each pulse of the plurality of pulses are then concatenated to obtain the image to be processed. For example, the image to be processed obtained through sampling is as follows: Figure 10b As shown.

[0122] In step S902, the region to be identified is determined from the image to be processed.

[0123] For example, Figure 10b The various objects that appear in the story (e.g.) Figure 10b The data (such as a bundle of wires, long sticks, etc.) is segmented out, and the area where each object is located is taken as a region to be identified.

[0124] In step S903, the candidate depth is used to reconstruct the region to be identified in order to obtain the reconstruction result of the region to be identified.

[0125] In some embodiments, the region to be identified is reconstructed as follows: a focusing ratio is determined based on the distances from the multiple detectors to the X-ray source and the candidate depth; based on the focusing ratio, the coordinate components of the pixels in the region to be identified in a specified direction parallel to the scanning direction of the object are focused. In step S904, the edge sharpness of the reconstruction result is determined.

[0126] In step S905, it is determined whether the edge sharpness is greater than the sharpness threshold.

[0127] If the edge sharpness is greater than the sharpness threshold, proceed to steps S906, S908, and S909; if the edge sharpness is less than or equal to the edge sharpness threshold, proceed to step S907.

[0128] In step S906, the candidate depth is used as the depth of the region to be identified. In step S907, the candidate depth is updated.

[0129] In some examples, the candidate depth is incremented within a certain range according to a set step size. For example, if the first value of the candidate depth is 0.1 meters, and the set step size is 0.1 meters, then the second value is 0.2 meters, and the third value is 0.3 meters.

[0130] After step S907, steps S903 to S905 are executed again, and then subsequent processing steps are selected based on the judgment result of step S905.

[0131] In step S908, the region to be identified is reconstructed according to its depth to obtain the reconstructed image.

[0132] In this step, when there are multiple regions to be identified, each region is reconstructed separately according to its depth.

[0133] For example, if the depth of region 1 to be identified is 3 meters, then region 1 will be reconstructed based on a depth of 3 meters. If the depth of region 2 to be identified is 3.6 meters, then region 2 will be reconstructed based on a depth of 3.6 meters. If region 1 and region 2 to be identified are reconstructed based on a depth of 3 meters, then region 1 will be clear while region 2 will be blurry.

[0134] For example, with a depth of 3800mm, for Figure 10b The image to be processed is reconstructed to obtain, as shown below. Figure 10c The reconstruction result is shown. Figure 10c As shown, in the reconstructed image, the region where the object has a depth of 3800mm (e.g., Figure 10c The bundle of wires on the left side of the middle section is relatively clear, while the areas containing objects at other depths (such as...) are clearer. Figure 10c The two long sticks on the right side of the middle are rather blurry.

[0135] For example, with a depth of 4600mm, for Figure 10b The image to be processed is reconstructed to obtain, as shown below. Figure 10d The reconstruction result is shown. Figure 10d As shown, in the reconstructed image, the image of an object with a depth of 4600mm (e.g.) Figure 10d The two long sticks on the right side of the image are clear, while images of objects at other depths (such as...) are sharp. Figure 10d The bundle of wires on the left side of the middle section is rather blurry.

[0136] In some examples, step S908 includes: determining a focusing ratio based on the distance from the multi-column detectors to the X-ray source and the depth of the area to be identified; and focusing the coordinate components of the pixels in the area to be identified in a specified direction, which is parallel to the scanning direction of the object, according to the focusing ratio.

[0137] In some examples, the focusing ratio is determined by: calculating the difference between the distance from the multi-array detectors to the X-ray source and the depth; and using the ratio of this difference to the distance from the multi-array detectors to the X-ray source as the focusing ratio.

[0138] An exemplary method for focusing the coordinate components according to the focusing ratio includes: multiplying the coordinate components of the pixels in the region to be identified in a specified direction by the focusing ratio to obtain an offset; and then, according to this offset, performing a translation operation on the coordinate components of the pixels in the region to be identified in the specified direction to obtain the focused coordinate components. Following this process, the focused image (i.e., the reconstructed image) can be obtained. In step S909, the reconstructed image and the depth information of the region to be identified are displayed.

[0139] In some examples, depth information is associated with the reconstructed image, and the associated image is then displayed. For instance, depth information can be overlaid on the reconstructed image using color-coded labels. Alternatively, depth information can be overlaid on the reconstructed image using tooltips, marker bars, or other similar methods. For example... Figure 10e As shown, Z1 represents a depth of 3.8 meters, Z2 represents a depth of 3 meters, and Z3 represents a depth of 4.6 meters. This makes it easier for inspectors to identify objects in the image, greatly reducing the difficulty of object identification and improving the efficiency of object inspection.

[0140] In this embodiment of the disclosure, through the above process, based on the main structure of a conventional X-ray transmission inspection system, multiple detectors can be used to scan at an appropriate scanning speed to obtain transmission images. After a series of image processing steps, depth information is extracted, enabling depth identification of objects such as containers and cargo. Based on the method of this embodiment, depth identification can be achieved based on transmission images without significant changes to the system hardware configuration. This makes it easy to embed the depth identification function into existing DR transmission scanning processes, reducing the difficulty of DR image interpretation and unpacking inspections, and facilitating object identification by inspection personnel.

[0141] Figure 11 This is a schematic diagram of the structure of an image processing apparatus according to some embodiments of the present disclosure. Figure 11 As shown, the image processing apparatus 110 includes a memory 111 and a processor 112 coupled to the memory 111. The memory 111 is used to store instructions for performing an embodiment of the image processing method. The processor 112 is configured to perform the image processing method in any of the embodiments of this disclosure based on the instructions stored in the memory 111.

[0142] Figure 12 This is a schematic diagram of the structure of a transmission scanning system according to some embodiments of the present disclosure. Figure 12 As shown, the transmission scanning system includes an image processing device 110, a radiation source 101, and a detector 121.

[0143] A radiation source 101 is used to emit radiation so as to perform a transmission scan of an object 103 based on the radiation.

[0144] Detector 121 is used to detect the transmission intensity when the object 103 is scanned by transmission in order to obtain a transmission image.

[0145] In some embodiments, the scanning speed of the object is limited to a set range to ensure that the geometry of the object in the image corresponding to the single-column detector is substantially normal, or that the geometry of the object in the image to be processed after sampling is substantially normal. This helps to improve the subsequent image processing effect, facilitating further improvement in the efficiency of object inspection and the object recognition effect.

[0146] Image processing apparatus 110 is used to perform the image processing method as described above.

[0147] In the embodiments of this disclosure, the above-described transmission scanning system can obtain the depth information of an object without adjusting the detector structure, which helps to improve the efficiency and recognition effect of object inspection.

[0148] Figure 13 This is a schematic diagram of the structure of a computer system according to some embodiments of the present disclosure.

[0149] like Figure 13 As shown, the computer system 130 can be represented in the form of a general computing device. The computer system 130 includes a memory 131, a processor 132, and a bus 133 connecting different system components.

[0150] The memory 131 may include, for example, system memory, non-volatile storage media, etc. The system memory may store, for example, an operating system, application programs, a boot loader, and other programs. The system memory may include volatile storage media, such as random access memory (RAM) and / or cache memory. The non-volatile storage media may store, for example, instructions for a corresponding embodiment of at least one image processing method being executed. Non-volatile storage media include, but are not limited to, disk storage, optical storage, flash memory, etc.

[0151] The processor 132 can be implemented using a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete hardware components such as discrete gates or transistors. Accordingly, each module, such as the sampling module, the determination module, and the reconstruction module, can be implemented by executing instructions in the central processing unit (CPU) running memory to perform the corresponding steps, or by implementing dedicated circuitry to perform the corresponding steps.

[0152] Bus 133 can use any of the various bus architectures. For example, bus architectures include, but are not limited to, the Industry Standard Architecture (ISA) bus, the Micro Channel Architecture (MCA) bus, and the Peripheral Component Interconnect (PCI) bus.

[0153] The computer system 130 can be connected via a bus 133 to interfaces 134, 135, and 136, as well as to memory 131 and processor 132. Input / output interface 134 provides a connection interface for input / output devices such as monitors, mice, and keyboards. Network interface 135 provides a connection interface for various networked devices. Storage interface 136 provides a connection interface for external storage devices such as floppy disks, USB flash drives, and SD cards.

[0154] Various aspects of this disclosure are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatus, and computer program products according to embodiments of this disclosure. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations thereof, can be implemented by computer-readable program instructions.

[0155] These computer-readable program instructions are provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable device to produce a machine, such that execution of the instructions by the processor produces means for implementing the functions specified in one or more boxes of the flowchart and / or block diagram.

[0156] These computer-readable program instructions may also be stored in a computer-readable storage medium. These instructions cause a computer to work in a particular manner to produce an article of manufacture, including instructions that implement the functions specified in one or more boxes in a flowchart and / or block diagram.

[0157] This disclosure may take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects.

[0158] The image processing method, apparatus, and transmission scanning system described in the above embodiments can obtain the depth information of an object without adjusting the detector structure, which helps to improve the efficiency and recognition effect of object inspection.

[0159] The image processing method, apparatus, and transmission scanning system according to this disclosure have now been described in detail. To avoid obscuring the concept of this disclosure, some details known in the art have not been described. Those skilled in the art can fully understand how to implement the technical solutions disclosed herein based on the above description.

Claims

1. An image processing method, comprising: The transmission image is sampled to obtain the image to be processed. The transmission image is obtained by detecting the transmission intensity when the object is scanned using a multi-column detector. The sampling process is used to highlight the information collected by the crystal unit at different positions in the multi-column detector. Determine the region to be identified from the image to be processed; The candidate depth is used to reconstruct the region to be identified in order to obtain the reconstruction result of the region to be identified. Determine the edge sharpness of the reconstruction result; If the edge sharpness is greater than the sharpness threshold, the candidate depth is used as the depth of the region to be identified. If the edge sharpness is less than or equal to the sharpness threshold, the steps of candidate depth update, region reconstruction based on the updated candidate depth, and edge sharpness determination are performed iteratively until the edge sharpness is greater than the sharpness threshold.

2. The image processing method according to claim 1, wherein, The transmitted image includes images acquired by the multi-column detectors under each pulse of multiple pulses, and the sampling processing of the transmitted image to obtain the image to be processed includes: From each row of pixels in the image acquired under each pulse, K pixels are extracted to obtain a pixel sequence corresponding to the image acquired under each pulse, where K is an integer greater than or equal to 1 and less than N, and N is the number of columns of the crystal unit of the multi-column detector. The pixel sequence corresponding to the image acquired under each of the multiple pulses is stitched together to obtain the image to be processed.

3. The image processing method according to claim 2, wherein, Where K is 1, extracting K pixels from each row of pixels in the image acquired under each pulse includes: Based on the element value q of the i-th element in the sampling sequence, the q-th pixel is extracted from the i-th row of the image acquired under each pulse. The value of i is any integer from 1 to M, where M is the number of rows of the crystal unit of the multi-column detector, and q is an integer greater than or equal to 1 and less than or equal to N.

4. The image processing method according to claim 3, wherein, The sampling sequence includes a first sampling sequence, in which the element values ​​are determined based on a random function.

5. The image processing method according to claim 3, wherein, The sampling sequence includes a second sampling sequence, wherein the arrangement of pixels extracted by the second sampling sequence conforms to a specified pattern.

6. The image processing method according to claim 1, wherein, Determining the edge sharpness of the reconstruction result includes: The edge pixels of the reconstructed result are extracted using an edge extraction algorithm; The edge sharpness of the reconstruction result is determined based on the edge pixels.

7. The image processing method according to claim 6, wherein, Based on the edge pixels, determine the edge sharpness of the reconstructed result: Calculate the number of edge pixels contained in a specified neighborhood for each pixel in the reconstruction result; Calculate the number of edge pixels whose number is less than the quantity threshold; The edge sharpness of the reconstruction result is determined based on the number of edge pixels that are less than the number threshold and the total number of edge pixels in the reconstruction result.

8. The image processing method according to claim 7, wherein, The step of determining the edge sharpness of the reconstruction result based on the number of edge pixels less than a threshold and the total number of edge pixels in the reconstruction result includes: The ratio of the number of edge pixels less than a certain threshold to the total number of edge pixels in the reconstruction result is used as the edge sharpness of the reconstruction result.

9. The image processing method according to claim 1, wherein, The process of reconstructing the region to be identified using candidate depth includes: The focusing ratio is determined based on the distance from the multi-row detectors to the X-ray source and the candidate depth; According to the focusing ratio, the coordinate components of the pixels in the region to be identified in a specified direction are focused, and the specified direction is parallel to the scanning direction of the object.

10. The image processing method according to claim 9, wherein, The focusing ratio is determined based on the distance from the multi-column detectors to the X-ray source and the candidate depth, including: Calculate the distance from the multi-column detectors to the X-ray source and the difference between this distance and the candidate depth; The ratio of the difference to the distance from the multi-column detector to the X-ray source is used as the focusing ratio.

11. The image processing method according to any one of claims 1 to 9, further comprising: Based on the depth of the region to be identified, the region to be identified is reconstructed to obtain a reconstructed image.

12. The image processing method according to claim 11, further comprising: The reconstructed image and the depth information of the region to be identified are displayed.

13. An image processing apparatus, comprising: Memory; as well as A processor coupled to the memory, the processor being configured to execute the image processing method as described in any one of claims 1 to 12 based on instructions stored in the memory.

14. A transmission scanning system, comprising: The image processing apparatus as described in claim 13; Multi-row detectors; as well as X-ray source.

15. A computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the image processing method as described in any one of claims 1 to 12.

16. A computer program product having stored computer program instructions thereon, which, when executed by a processor, implement the image processing method as described in any one of claims 1 to 12.

Images