Image registration method and system for high-precision measurement of missile engine main shaft size

Abstract

The application relates to the field of image processing, in particular to an image registration method and system for high-precision measurement of the size of a main shaft of a missile engine. The method comprises the following steps: acquiring multiple to-be-registered images, extracting multiple candidate structure bands in each to-be-registered image, and recording the structure band attributes corresponding to the multiple candidate structure bands; calculating the size-sensitive saliency of the ith candidate structure band corresponding to the current to-be-registered image, the axial sequential continuity between the ith candidate structure band in the current to-be-registered image and the jth candidate structure band in the next to-be-registered image, and the helical phase correspondence, determining the measurement fidelity registration degree between the ith candidate structure band in the current to-be-registered image and the jth candidate structure band in the next to-be-registered image; and according to the measurement fidelity registration degree, solving the registration transformation relationship between the current to-be-registered image and the next to-be-registered image, so as to realize the registration of the to-be-registered images. The application can improve the registration accuracy.

Description

Technical Field

[0001] This invention relates to the field of image processing, and specifically to an image registration method and system for high-precision measurement of the main shaft dimensions of a missile engine. Background Technology

[0002] The current method for measuring the dimensions of a missile engine spindle typically involves first acquiring local images of the spindle from different positions or perspectives, then registering multiple images, and finally extracting dimensional parameters such as step distance, diameter, slot width, thread start and end positions, and helical segment positions from the registration results.

[0003] Existing registration schemes mostly rely on texture features, grayscale similarity, or overall contour overlap. For ordinary parts, these methods can achieve image alignment, but the surfaces of spindle-type parts are mostly smooth metal surfaces with weak texture and strong reflection. Long cylindrical segments, repeating stage stages, and periodic spiral segments have a large number of similar structures in the image, which can easily lead to misregistration of adjacent steps, adjacent slots, adjacent spiral periods, or local highlight bands as the same structure.

[0004] Existing spindle image registration methods lack specific screening and acceptance judgment for dimension-sensitive structural zones. They cannot distinguish between repeating cylindrical contours that are "visible but not suitable for dimensional measurement" and stepped shoulders, relief grooves, thread start and stop edges, and helical junctions that "truly correspond to the key dimensions in the drawing." This causes registration errors to continue to be transmitted to the key dimension measurement results. Summary of the Invention

[0005] This invention provides an image registration method and system for high-precision measurement of the main shaft dimensions of a missile engine, in order to solve existing problems.

[0006] The image registration method for high-precision measurement of the main shaft dimension of a missile engine according to the present invention adopts the following technical solution: One embodiment of the present invention provides an image registration method for high-precision measurement of the main shaft dimensions of a missile engine, the method comprising the following steps: For the main axis of the missile engine, multiple images to be registered are acquired sequentially along the main axis direction. Multiple candidate structure bands are extracted from each image to be registered, and the structure band attributes corresponding to the multiple candidate structure bands are recorded. Based on the bilateral continuity ratio and key size overlap ratio of the i-th candidate structural band in the current image to be registered, determine the size sensitivity significance corresponding to the i-th candidate structural band in the current image to be registered. Based on the order consistency ratio and spacing deviation between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, determine the axial order continuity between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered. Based on whether the i-th candidate structure band in the current image to be registered is located in a helical segment, the helical direction information of the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, and the phase offset between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, the helical phase correspondence between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered is determined; Based on the size sensitivity significance of the i-th candidate structure band in the current image to be registered, the axial sequence continuity between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, and the spiral phase correspondence between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, the measurement fidelity registration degree between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered is determined. Based on the measurement fidelity registration between the i-th candidate structure band in the current image to be registered and the J candidate structure bands in the next image to be registered, the registration transformation relationship between the current image to be registered and the next image to be registered is solved to achieve the registration of the images to be registered.

[0007] This invention proposes an image registration system for high-precision measurement of the main shaft dimension of a missile engine, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the computer program is executed by the processor, it implements the steps of the image registration method for high-precision measurement of the main shaft dimension of a missile engine.

[0008] The beneficial effects of the technical solution of the present invention are: (1) It can remove long cylindrical segments, large-area reflective edges and local repetitive textures from the candidate registration objects, and prioritize the retention of structural bands that truly affect size extraction, thereby reducing the mismatch of repetitive structures; (2) It can perform axial sequence judgment on the shoulder of the step, the two sides of the groove and the edge of the relief groove in the multi-step spindle, and perform phase correspondence judgment on the spiral peak and valley positions in the left and right spiral axes, thereby avoiding mistaking adjacent structural periods as structures of the same size.

[0009] (3) The registration results are no longer evaluated solely based on the degree of image overlap. Instead, the relative spacing and boundary position of the key dimension structural bands after registration are further checked to ensure that the registration results directly serve high-precision dimension measurement. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 A flowchart illustrating an embodiment of the image registration method for high-precision measurement of missile engine spindle dimensions provided by the present invention; Figure 2 This is a structural diagram of an image registration system for high-precision measurement of the main shaft dimensions of a missile engine, provided in an embodiment of the present invention. Detailed Implementation

[0012] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of the image registration method for high-precision measurement of the main shaft dimensions of a missile engine proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0013] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0014] The following description, in conjunction with the accompanying drawings, details the specific scheme of the image registration method for high-precision measurement of the main shaft dimensions of missile engines provided by this invention.

[0015] This invention provides an image registration method and system for high-precision measurement of the main shaft dimensions of a missile engine. Please refer to [link to relevant documentation]. Figure 1 The diagram illustrates a flowchart of an image registration method for high-precision measurement of the main shaft dimension of a missile engine, provided by an embodiment of the present invention. The method includes the following steps: S101. For the main axis of the missile engine, acquire multiple images to be registered sequentially along the main axis direction, extract multiple candidate structure bands in each image to be registered, and record the structure band attributes corresponding to the multiple candidate structure bands.

[0016] For example, when acquiring the image to be registered, the missile engine spindle is placed in a rotary clamping or directional conveying station, and at least one set of industrial cameras and telecentric optical imaging components are used to acquire the outer contour image of the spindle (i.e., the image to be registered). For long-axis parts, multiple local images can be acquired in segments along the axial direction; for helical shafts, additional sampling images can be added at different rotational positions. Each image retains the calibrated pixel size relationship and the spindle axis direction.

[0017] Because the spindle surface has strong metallic reflection, directly using the original bright edge as the registration object can easily confuse the highlight edge with the real structural boundary. Therefore, after acquiring the image, edge enhancement and reflection suppression processing can be performed first to preserve stable contour edges; then, according to the spindle axis direction, the edges in each image are organized into candidate structural zones in the form of "paired boundaries + intermediate spacing". A structural zone refers to a local contour unit that can collectively represent a structure of the same size, such as the front and rear sides of a step shoulder, the left and right sides of a relief groove, the two sides of an annular groove opening, the edge combination of a thread start and end position, and the junction zone of a helical blade and a cylindrical section.

[0018] For example, the structural band attributes include the axial interval in which the structural band is located, the local spacing between the two sides, the continuous length of the two sides, the edge polarity, the local transition curvature, the front-to-back order with adjacent structural bands, and whether it is located in the helical segment region.

[0019] Thus, the present invention has obtained candidate structural bands and their corresponding axial intervals, bilateral contours, local spacing, curvature and adjacent order information in each principal axis measurement image (i.e. image to be registered).

[0020] S102. Based on the bilateral continuity ratio and key size overlap ratio of the i-th candidate structural band in the current image to be registered, determine the size sensitivity significance corresponding to the i-th candidate structural band in the current image to be registered.

[0021] Specifically, S102 may include: Extract the front and back boundaries of the i-th candidate structure band and perform boundary segmentation to obtain N corresponding front boundary segments and M back boundary segments. For the n-th front boundary segment, calculate the overlap length between the n-th front boundary segment and the M back boundary segments. Based on the overlap length, the axial coverage interval of the n-th front boundary segment, and the axial coverage interval of the back boundary segments, determine the back boundary segments that are effectively paired with the n-th front boundary segment. Calculate the ratio of the sum of the overlap lengths of all effectively paired front and back boundary segments to the theoretical interval length of the i-th candidate structure band to obtain the bilateral continuity ratio of the i-th candidate structure band. When the i-th candidate structural band is a non-critical structure, the critical dimension overlap ratio of the i-th candidate structural band is determined to be 0; when the i-th candidate structural band is a critical structure, the ratio of the overlap length between the axial interval of the i-th candidate structural band and the critical dimension interval of the drawing to the critical dimension interval of the drawing is calculated to obtain the critical dimension overlap ratio of the i-th candidate structural band. The size sensitivity significance of the i-th candidate structural band is determined based on the bilateral continuity ratio and the critical size overlap ratio.

[0022] It's understandable that structures truly worth using for dimensional measurement registration typically meet two conditions simultaneously. First, the structure itself corresponds to the location of key dimensions marked on the drawing, such as stepped shoulders, relief grooves, annular grooves, thread start and end edges, end face transition zones, and helical junctions. Second, this structure can appear in the image in a relatively stable bilateral form, with the local spacing between the two sides not changing drastically within a short interval. Conversely, while long straight cylindrical segments may have continuous edges, the corresponding structures are too repetitive to uniquely determine dimensional locations; and while local highlight edges may be bright, the bilateral relationship is unstable and unsuitable as measurement benchmarks.

[0023] For example, all edge points of the candidate structure band can be divided into two groups: Front boundary: A set of edge points with smaller axial coordinates, corresponding to the starting side of the structure; Back boundary: A set of edge points with larger axial coordinates, corresponding to the termination side of the structure.

[0024] For example, due to reasons such as metal reflection and edge blurring, the original boundary is usually not a continuous straight line. Therefore, it is necessary to divide each boundary into several continuous boundary segments with a length greater than a preset pixel threshold (such as 5 pixels, which can filter out scattered edge points caused by noise while retaining real structural boundary fragments) according to the axial coordinate, forming N front boundary segments and M back boundary segments respectively.

[0025] For example, a sequence of front boundary segments can be constructed based on multiple front boundary segments, and a sequence of back boundary segments can be constructed based on multiple back boundary segments: Front boundary segment sequence QB=[QB1,QB2,...,QB...] N The sequence of the back boundary segments is HB=[HB1,HB2,...,HB]. M ].

[0026] For example, taking the previous boundary segment QB1 as an example, we can calculate the relationship between QB1 and HB1, HB2, ..., HB M The overlap length.

[0027] It can be understood that the overlap length is the length of the overlap between the axial intervals of the front boundary segment and the axial intervals of the rear boundary segment.

[0028] For example, if the overlap length between the front boundary segment and the rear boundary segment is greater than 0.8 times the shorter axial interval length of the front boundary segment and the rear boundary segment, the front boundary segment and the rear boundary segment can be considered to be effectively paired. It should be noted that if a front boundary segment corresponds to multiple effectively paired rear boundary segments, the rear boundary segment with the largest overlap length is taken as the effectively paired rear boundary segment with the front boundary segment.

[0029] For example, the theoretical interval length of the i-th candidate structural zone is the length of the axial interval of the i-th candidate structural zone.

[0030] For example, taking the axial interval of the i-th candidate structure band as [100.0, 102.5], the front boundary segment sequence QB=[QB1, QB2], the rear boundary segment sequence HB=[HB1, HB2], the axial interval of the front boundary segment QB1 as [100.0, 100.8], the axial interval of the front boundary segment QB2 as [101.2, 102.5], the axial interval of the rear boundary segment HB1 as [100.1, 100.9], and the axial interval of the rear boundary segment HB2 as [101.3, 102.4] as an example: The overlap length between QB1 and HB1 can be calculated as 100.8-100.1=0.7, the overlap length between QB1 and HB2 is 0; the overlap length between QB2 and HB1 is 0, and the overlap length between QB2 and HB2 is 102.4-101.3=1.1. The axial interval lengths of QB1 and HB1 are both 100.8-100.0=100.9-100.1=0.8. The overlap length of QB1 and HB1 is 0.7>0.8×0.8. Therefore, QB1 and HB1 are effectively paired, that is, HB1 is the back boundary segment that is effectively paired with QB1. The axial interval length of QB2 is 102.5-101.2=1.3, and the axial interval length of HB2 is 102.4-101.3=1.1 (the smaller value). The overlap length of QB2 and HB2 is 1.1>0.8×1.1, so QB2 and HB2 are effectively paired, that is, HB2 is the back boundary segment that is effectively paired with QB2. The sum of the overlap lengths of all valid pairing front and back boundary segments is 0.7 + 1.1 = 1.8. The theoretical interval length of the i-th candidate structure band is 102.5 - 100.0 = 2.5. Therefore, the bilateral continuity ratio of the i-th candidate structure band is 1.8 / 2.5 = 0.72.

[0031] For example, non-critical structures may include long straight cylindrical segments (continuous cylindrical surfaces without abrupt diameter changes), localized high-gloss edges (false single edges formed by metallic reflection), machining burrs / scratches (scattered edges formed by surface defects), and tooling fixture edges (structures not of the spindle itself); critical structures may include stepped shoulders (both sides at abrupt diameter changes in cylindrical segments), relief grooves (left and right walls of relief grooves at the root of threads / splines), annular groove openings (upper and lower edges of annular grooves), thread start and stop edges (both sides at the junction of threaded segments and cylindrical segments), helical transition zones (transition areas between helical blades and cylindrical segments), and end face transition zones (junctions between the two end faces of the spindle and the outer cylindrical surface).

[0032] It should be noted that the structure category marker of the i-th candidate structure in the drawing can be read to determine whether it is a non-critical structure or a critical structure.

[0033] For example, all critical dimensions with specified tolerances are exported from the main spindle drawing, and a mapping relationship between the global coordinate system of the drawing and the image coordinate system is established, thereby determining the interval values ​​of the critical dimensions in the image coordinate system (i.e., the critical dimension interval of the drawing).

[0034] For example, the drawing information of the low-pressure shaft of a missile engine includes key dimensions: the third step shoulder, the axial range after conversion (i.e. the key dimension range of the drawing) is [250.0, 250.8], and the structural category of the i-th candidate structural band is step shoulder; The axial interval of the i-th candidate structural band is [250.1, 250.7]. The overlap length between the axial section of the i-th candidate structural zone and the critical dimension section of the drawing is calculated as 250.7-250.1=0.6, and the length of the critical dimension section of the drawing is 250.8-250.0=0.8. The ratio of the overlap length to the length of the critical dimension section of the drawing is 0.6 / 0.8=0.75. 0.75 is the critical dimension overlap ratio of the i-th candidate structural zone.

[0035] Specifically, based on the bilateral continuity ratio and the critical size overlap ratio, the size sensitivity significance corresponding to the i-th candidate structural band is determined, including: The significance of size sensitivity is determined based on the formula for calculating the significance of size sensitivity; the formula for calculating the significance of size sensitivity is:

[0036] In the formula, This represents the size-sensitive saliency of the i-th candidate structural band in the current image to be registered. This represents the bilateral continuity ratio of the i-th candidate structural band in the current image to be registered. This represents the critical size overlap ratio of the i-th candidate structural band in the current image to be registered.

[0037] In this formula, It can indicate whether the i-th candidate structure band in the image exists stably. This can be demonstrated by how much significance the key dimensional attributes of the drawing can supplement and enhance the image when stability is not yet fully achieved. This approach avoids directly elevating structures that are only significant on the drawing but have unstable image edges, and also avoids mistaking structural bands that, although continuous, belong to ordinary repetitive cylindrical segments as key registration references.

[0038] Thus, by examining the bilateral continuity and key dimension correlation of candidate structural bands, the saliency of the size-sensitive structural band for the i-th candidate structural band was obtained. It can be understood that the size-sensitive saliency reflects whether a candidate structural band is worth entering subsequent registration judgment. Its calculation first establishes a basic saliency value based on the bilateral continuity of the candidate structural band in the image, and then uses the degree of overlap between the structural band and the key dimension position on the drawing to correct the remaining parts. The result obtained in this way highlights structural bands that "both appear stably and truly correspond to the key dimensions."

[0039] S103. Based on the order consistency ratio and spacing deviation between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, determine the axial order continuity between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered.

[0040] Specifically, S103 may include: Based on the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, construct the first structure band order list and the second structure band sequence list; Based on the first structural band sequence list and the second structural band sequence list, determine whether the order of each pair of adjacent structural bands in the first structural band sequence list is consistent with the order of the corresponding adjacent structural bands in the second structural band sequence list. Calculate the ratio of the number of adjacent structural band pairs with consistent order to the total number of adjacent structural band pairs to obtain the order consistency ratio between the i-th candidate structural band in the current image to be registered and the j-th candidate structural band in the next image to be registered. Based on the first structural band sequence list, determine the spacing corresponding to each pair of adjacent structural bands in the first structural band sequence list, and construct a first spacing sequence; based on the second structural band sequence list, determine the spacing corresponding to each pair of adjacent structural bands in the second structural band sequence list, and construct a second spacing sequence; based on the first spacing sequence and the second spacing sequence, determine the spacing ratio difference sequence; based on the spacing ratio difference sequence, determine the spacing deviation between the i-th candidate structural band in the current image to be registered and the j-th candidate structural band in the next image to be registered; Based on the sequence consistency ratio and spacing deviation, the axial sequence continuity between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered is determined.

[0041] Understandably, for low-pressure spindles and multi-step spindles, the most common mismatch is not the complete absence of boundaries, but rather mistaking two visually similar adjacent step shoulders, adjacent notches, or adjacent relief grooves for the same structure. To avoid this mismatch, comparing only the local shape of a single structural band is insufficient; it is also necessary to consider whether the order of its preceding and following structural bands is consistent, whether the spacing ratio between bands is consistent, and whether the edge polarity is consistent. Only when these three factors are relatively stable simultaneously are two candidate structural bands in adjacent images more suitable as a correspondence of the same physical structure.

[0042] For example, a first structural band order list can be constructed based on the i-th candidate structural band of the current image to be registered. The first structural band order list may include the (i-2)-th candidate structural band of the current image to be registered, the (i-1)-th candidate structural band of the previous image, the i-th candidate structural band of the previous image, the (i+1)-th candidate structural band of the previous image, and the (i+2)-th candidate structural band of the previous image, QS=[QS i-2 QS i-1 QS i QS i+1 QS i+2 ]; Based on the j-th candidate structure band in the next image to be registered, a second structure band order list is constructed. This second structure band order list may include the (j-2)-th candidate structure band of the next image to be registered, the (j-1)-th candidate structure band of the next image, the j-th candidate structure band of the next image, the (j+1)-th candidate structure band of the next image, and the (j+2)-th candidate structure band of the next image. HS=[HS j-2 ,HS j-1 ,HS j ,HS j+1 ,HS j+2 ]; Next, the QS can be determined. i-2 QS i-1 The order of precedence, and HS j-2 HS j-1 Whether the order of events is consistent; if consistent, then QS i-2 QS i-1 and HS j-2 HS j-1The adjacent structural bands belong to the adjacent structural bands with consistent sequential relationship; similarly, it can be determined whether the subsequent adjacent structural bands belong to the adjacent structural bands with consistent sequential relationship; in this example, all adjacent structural bands belong to the adjacent structural bands with consistent sequential relationship, that is, the number of adjacent structural bands with consistent sequential relationship is 4. It is understandable that the number of all adjacent structural band pairs is 4, including QS. i-2 QS i-1 With HS j-2 HS j-1 (First pair), QS i-1 QS i With HS j-1 HS j (Second pair), QS i QS i+1 With HS j HS j+1 (Third pair), QS i+1 QS i+2 With HS j+1 HS j+2 (The fourth pair); The ratio of the number of adjacent structural band pairs with consistent order (i.e., 4) to the total number of all adjacent structural band pairs (i.e., 4) is calculated to obtain the order consistency ratio between the i-th candidate structural band in the current image to be registered and the j-th candidate structural band in the next image to be registered, which is 1.

[0043] For example, continuing with the previous example, the spacing between each pair of adjacent structural bands in the first structural band sequence list is 10 (QS). i-2 With QS i-1 ), 20 (QS i-1 With QS i ), 15 (QS) i With QS i+1 ), 15 (QS) i+1 With QS i+2 The spacing between each pair of adjacent structural zones in the second structural zone sequence list is 20 (HS). j-2 With HS j-1 ), 15 (HS) j-1 With HS j ), 15 (HS) j With HS j+1 ), 20 (HS) j+1 With HS j+2 If ), then the first spacing sequence is [10,20,15,15], and the second spacing sequence is [20,15,15,20]. For example, based on the first spacing sequence and the second spacing sequence, the ratio between the corresponding values ​​in the first spacing sequence and the second spacing sequence is first calculated, and then the absolute value of the difference between the ratio and 1 is calculated to obtain the spacing ratio difference value, and then the spacing ratio difference value sequence is obtained.

[0044] For example, continuing with the previous example, the absolute value of (10 / 20)-1 is calculated to be 0.5, the absolute value of (20 / 15)-1 is 0.333, the absolute value of (15 / 15)-1 is 0, and the absolute value of (15 / 20)-1 is 0.25. The resulting interval ratio difference sequence is [0.5, 0.333, 0, 0.25].

[0045] For example, after obtaining the spacing ratio difference sequence, the average value of all spacing ratio differences can be calculated and normalized according to the allowable spacing deviation of the drawing to obtain the spacing deviation between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered.

[0046] For example, continuing with the previous example, the average value of all spacing ratio differences is calculated as (0.5 + 0.333 + 0 + 0.25) / 4 = 0.27075. Taking the allowable spacing deviation of the drawing as an example, the ratio of the average value to the allowable spacing deviation of the drawing is calculated as 0.27075 / 0.005 = 54.15, which is the spacing deviation between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered.

[0047] Specifically, based on the sequence consistency ratio and spacing deviation, the axial sequence continuity between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered is determined, including: The axial sequential bearing capacity is determined based on the formula for calculating the axial sequential bearing capacity. The formula for calculating the axial sequential bearing capacity is as follows:

[0048] In the formula, This represents the axial sequential continuity between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered. This represents the order consistency ratio between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered. This represents the spacing deviation between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered.

[0049] It should be noted that, for the i-th candidate structure band of the current image to be registered and the J candidate structure bands of the next image to be registered, a total of J axial sequential continuity can be obtained.

[0050] It can be understood that the axial sequence continuity can reflect the accuracy of how well the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered can be connected in the same structure according to the principal axis sequence.

[0051] S104. Based on whether the i-th candidate structure band in the current image to be registered is located in a helical segment, the helical direction information of the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, and the phase offset between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, determine the helical phase correspondence between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered.

[0052] Specifically, S104 may include: When the i-th candidate structure band in the current image to be registered is located in the spiral segment, the spiral phase correspondence between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered is determined based on the spiral direction information of the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, and the phase offset between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered. If the i-th candidate structure band in the current image to be registered is not located in the spiral segment, the spiral phase correspondence between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered is determined to be 1.

[0053] It is understandable that, for left-handed and right-handed helical axes, simply relying on axial sequence may still align adjacent helical periods to the same structure. This is because the helical blades are repeatedly distributed along the axial direction, and adjacent periods appear very similar in local images, but their peak-valley phase positions, helical direction signs, and period start-end relationships are not the same. Only when the helical direction is consistent, the period positions are contiguous, and the peak-valley phase offset is small, do the helical structure bands in adjacent images truly correspond to the same physical location. For structure bands not located within the helical segment, the helical phase correspondence can be considered as full value, and registration is controlled solely by the results of the first two steps.

[0054] For example, the direction of rotation information can be determined by the "inclination direction of the helical edge relative to the main axis". Along the main axis from left to right, the helical edge tilting upward is right-handed and the helical edge tilting downward is left-handed.

[0055] For example, for the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, multiple parallel sampling lines can be uniformly extracted within the axial range of the candidate structure band along a direction perpendicular to the principal axis. For each sampling line: extract its intersection point with the spiral edge of the i-th candidate structure band / j-th candidate structure band; identify peak position, valley position and period boundary based on the intersection point, and construct the first phase position sequence and the second phase position sequence; The elements in the first phase position sequence and the second phase position sequence are paired one by one according to the same period number, and the normalized offset of the corresponding peak and valley positions within the period is calculated to form a phase offset sequence. The average of all elements in the phase offset sequence is used to obtain the phase offset between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered.

[0056] Specifically, based on the spiral orientation information of the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, and the phase offset between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, the spiral phase correspondence between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered is determined, which may include: The spiral phase correspondence is determined based on the formula for solving the spiral phase correspondence. The formula for solving the spiral phase correspondence is as follows:

[0057] In the formula, This represents the spiral phase correspondence between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered. This indicates the consistency of the rotation direction between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered. This represents the phase offset between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered.

[0058] For example, if the rotation direction information of the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered is consistent (e.g., both are left-handed or both are right-handed). It can take the value 1; if they are inconsistent, It can take the value 0.

[0059] It should be noted that, for the i-th candidate structure band of the current image to be registered and the J candidate structure bands of the next image to be registered, a total of J spiral phase correspondences can be obtained.

[0060] It can be understood that the spiral phase correspondence can reflect the degree of consistency between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered in terms of spiral direction and periodic phase.

[0061] S105. Based on the size sensitivity significance of the i-th candidate structure band in the current image to be registered, the axial sequence continuity between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, and the spiral phase correspondence between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, determine the measurement fidelity registration degree between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered.

[0062] Specifically, S105 may include: Based on the formula for calculating measurement fidelity registration, the measurement fidelity registration is determined; the formula for calculating measurement fidelity registration is:

[0063] In the formula, This represents the measurement fidelity registration between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered. This represents the size-sensitive saliency of the i-th candidate structural band in the current image to be registered. This represents the axial sequential continuity between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered. This represents the spiral phase correspondence between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered.

[0064] It should be noted that, for the i-th candidate structure band of the current image to be registered and the J candidate structure bands of the next image to be registered, a total of J measurement-fidelity registration scores can be obtained.

[0065] It is understandable that measuring the fidelity of registration can reflect the overall degree of registration between the i-th candidate structural band in the current image to be registered and the j-th candidate structural band in the next image to be registered.

[0066] S106. Based on the measurement fidelity registration degree between the i-th candidate structure band in the current image to be registered and the J candidate structure bands in the next image to be registered, solve the registration transformation relationship between the current image to be registered and the next image to be registered, so as to achieve the registration of the images to be registered.

[0067] For example, from the J candidate structure bands of the next image to be registered, the candidate structure band with the highest measurement fidelity registration is selected as the registration anchor structure band of the i-th candidate structure band in the current image to be registered, and the registration transformation relationship between the current image to be registered and the next image to be registered is solved accordingly.

[0068] It should be noted that solving the registration transformation relationship between the current image to be registered and the next image to be registered, and performing subsequent image registration, are conventional techniques in this field and will not be elaborated here.

[0069] It is understandable that for multi-step long axes, priority should be given to retaining structural band pairs with higher measurement fidelity registration in the shoulder of the step, the relief groove, and the two sides of the groove opening; for helical axes, priority should be given to retaining structural band pairs with higher measurement fidelity registration in the helical junction and the start and end positions of the cycle. If there are multiple segmented images of the same spindle, adjacent images are registered to a unified coordinate system in axial order, and it is checked whether the structural bands in the later added images still maintain the same axial order and helical period relationship as the preceding structural bands.

[0070] After obtaining a unified coordinate system, the system extracts the step distance, groove width, thread start and end positions, helical segment positions, and diameter dimensions within this coordinate system. For each critical dimension, the system prioritizes structural bands with high measurement fidelity and registration accuracy as measurement boundaries. If the same dimension can be repeatedly extracted from multiple structural bands, the system compares the fluctuations between the extraction results. If a dimension shows minimal fluctuation during repeated extractions, the current registration result is retained, and the dimension measurement value is output. If the fluctuation is significant, it indicates that although the current registration relationship can align at the contour level, it is still insufficient to stably support dimension extraction. In this case, structural band pairs with lower contributions are discarded, and local matching or supplementary image acquisition is re-executed.

[0071] For left-hand and right-hand helical axes, the system further checks whether the front and rear positions of adjacent period boundaries remain continuous in a unified coordinate system to avoid mistaking adjacent helical periods for the same measurement reference. For low-pressure axes and multi-step long axes, the system focuses on checking whether the relative spacing of the step shoulders, grooves, and relief grooves remains consistent. The registration results obtained after verification are directly used as the dimensional measurement reference, enabling the image registration results to stably serve the high-precision dimensional measurement of the missile engine spindle.

[0072] This invention is now complete.

[0073] In summary, in this embodiment of the invention, by determining the size sensitivity saliency corresponding to the i-th candidate structural band in the current image to be registered, the axial sequential continuity between the i-th candidate structural band in the current image to be registered and the j-th candidate structural band in the next image to be registered, and the helical phase correspondence between the i-th candidate structural band in the current image to be registered and the j-th candidate structural band in the next image to be registered, the measurement fidelity registration degree, which reflects the overall registerability of the i-th candidate structural band in the current image to be registered and the j-th candidate structural band in the next image to be registered, can be calculated. Then, based on the measurement fidelity registration degree, the current image to be registered and the next image to be registered are registered, thereby improving the registration accuracy.

[0074] This invention also proposes an image registration system for high-precision measurement of the main shaft dimensions of a missile engine. Please refer to [link to relevant documentation]. Figure 2 The diagram shows a structural diagram of an image registration system for high-precision measurement of the main shaft dimension of a missile engine provided in an embodiment of the present invention. The system includes: an acquisition module 201, a determination module 202, and a registration module 203.

[0075] The acquisition module 201 is used to sequentially acquire multiple images to be registered along the main axis of the missile engine, extract multiple candidate structure bands in each image to be registered, and record the structure band attributes corresponding to the multiple candidate structure bands. The determination module 202 is used to determine the size sensitivity saliency of the i-th candidate structure band in the current image to be registered based on the bilateral continuity ratio and key size coincidence ratio of the i-th candidate structure band in the current image to be registered; to determine the axial sequence continuity between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered based on the sequence consistency ratio and spacing deviation between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered; and to determine the spiral direction information of the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered based on whether the i-th candidate structure band in the current image to be registered is located in a spiral segment, and the current... The phase offset between the i-th candidate structure band in the image to be registered and the j-th candidate structure band in the next image to be registered is used to determine the spiral phase correspondence of the i-th candidate structure band in two adjacent images to be registered. Based on the size sensitivity significance of the i-th candidate structure band in the current image to be registered, the axial sequence continuity between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, and the spiral phase correspondence between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, the measurement fidelity registration between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered is determined. The registration module 203 is used to solve the registration transformation relationship between the current image to be registered and the next image to be registered based on the measurement fidelity registration degree between the i-th candidate structure band in the current image to be registered and the J candidate structure bands in the next image to be registered, so as to achieve the registration of the images to be registered.

[0076] The specific implementation methods and beneficial effects of this system embodiment can be found in the foregoing method embodiments, and will not be repeated here.

[0077] It should be noted that the system provided in the above embodiments is only an example of the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer device can be divided into different functional modules to complete all or part of the functions described above. In addition, the image registration system for high-precision measurement of the main shaft dimension of missile engines provided in the above embodiments and the image registration method for high-precision measurement of the main shaft dimension of missile engines belong to the same concept. The specific implementation process is detailed in the method embodiments and will not be repeated here.

[0078] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0079] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0080] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An image registration method for high-precision measurement of the main shaft dimensions of a missile engine, characterized in that, include: For the main axis of the missile engine, multiple images to be registered are acquired sequentially along the main axis direction. Multiple candidate structure bands are extracted from each image to be registered, and the structure band attributes corresponding to the multiple candidate structure bands are recorded. Based on the bilateral continuity ratio and key size overlap ratio of the i-th candidate structural band in the current image to be registered, determine the size sensitivity significance corresponding to the i-th candidate structural band in the current image to be registered. Based on the order consistency ratio and spacing deviation between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, determine the axial order continuity between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered. Based on whether the i-th candidate structure band in the current image to be registered is located in a helical segment, the helical direction information of the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, and the phase offset between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, the helical phase correspondence between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered is determined; Based on the size sensitivity significance of the i-th candidate structure band in the current image to be registered, the axial sequence continuity between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, and the spiral phase correspondence between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, the measurement fidelity registration degree between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered is determined. Based on the measurement fidelity registration between the i-th candidate structure band in the current image to be registered and the J candidate structure bands in the next image to be registered, the registration transformation relationship between the current image to be registered and the next image to be registered is solved to achieve the registration of the images to be registered.

2. The image registration method for high-precision measurement of the main shaft dimension of a missile engine according to claim 1, characterized in that, The step of determining the size sensitivity saliency corresponding to the i-th candidate structural band in the current image to be registered based on the bilateral continuity ratio and key size overlap ratio of the i-th candidate structural band in the current image to be registered includes: Extract the front and back boundaries of the i-th candidate structure band and perform boundary segmentation to obtain N corresponding front boundary segments and M back boundary segments. For the n-th front boundary segment, calculate the overlap length between the n-th front boundary segment and the M back boundary segments. Based on the overlap length, the axial coverage interval of the n-th front boundary segment, and the axial coverage interval of the back boundary segments, determine the back boundary segments that are effectively paired with the n-th front boundary segment. Calculate the ratio of the sum of the overlap lengths of all effectively paired front and back boundary segments to the theoretical interval length of the i-th candidate structure band to obtain the bilateral continuity ratio of the i-th candidate structure band. When the i-th candidate structural band is a non-critical structure, the critical dimension overlap ratio of the i-th candidate structural band is determined to be 0; when the i-th candidate structural band is a critical structure, the ratio of the overlap length between the axial interval of the i-th candidate structural band and the critical dimension interval of the drawing to the critical dimension interval of the drawing is calculated to obtain the critical dimension overlap ratio of the i-th candidate structural band. The size sensitivity significance of the i-th candidate structural band is determined based on the bilateral continuity ratio and the critical size overlap ratio.

3. The image registration method for high-precision measurement of the main shaft dimension of a missile engine according to claim 2, characterized in that, The determination of the size sensitivity significance corresponding to the i-th candidate structural band based on the bilateral continuity ratio and the critical size overlap ratio includes: The size sensitivity significance is determined based on the formula for calculating the size sensitivity significance; the formula for calculating the size sensitivity significance is as follows: In the formula, This represents the size-sensitive saliency of the i-th candidate structural band in the current image to be registered. This represents the bilateral continuity ratio of the i-th candidate structural band in the current image to be registered. This represents the critical size overlap ratio of the i-th candidate structural band in the current image to be registered.

4. The image registration method for high-precision measurement of the main shaft dimension of a missile engine according to claim 1, characterized in that, The step of determining the axial sequence continuity between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, based on the sequence consistency ratio and spacing deviation between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, includes: Based on the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, construct the first structure band order list and the second structure band sequence list; Based on the first structural band sequence list and the second structural band sequence list, determine whether the order of each pair of adjacent structural bands in the first structural band sequence list is consistent with the order of the corresponding adjacent structural bands in the second structural band sequence list. Calculate the ratio of the number of adjacent structural band pairs with consistent order to the total number of adjacent structural band pairs to obtain the order consistency ratio between the i-th candidate structural band in the current image to be registered and the j-th candidate structural band in the next image to be registered. Based on the first structural band sequence list, determine the spacing corresponding to each pair of adjacent structural bands in the first structural band sequence list, and construct a first spacing sequence; based on the second structural band sequence list, determine the spacing corresponding to each pair of adjacent structural bands in the second structural band sequence list, and construct a second spacing sequence; based on the first spacing sequence and the second spacing sequence, determine the spacing ratio difference sequence; based on the spacing ratio difference sequence, determine the spacing deviation between the i-th candidate structural band in the current image to be registered and the j-th candidate structural band in the next image to be registered; Based on the sequence consistency ratio and spacing deviation, the axial sequence continuity between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered is determined.

5. The image registration method for high-precision measurement of the main shaft dimension of a missile engine according to claim 4, characterized in that, The step of determining the axial sequence continuity between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, based on the sequence consistency ratio and spacing deviation, includes: The axial sequential bearing capacity is determined based on the formula for calculating the axial sequential bearing capacity; the formula for calculating the axial sequential bearing capacity is as follows: In the formula, This represents the axial sequential continuity between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered. This represents the order consistency ratio between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered. This represents the spacing deviation between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered.

6. The image registration method for high-precision measurement of the main shaft dimension of a missile engine according to claim 1, characterized in that, The step of determining the helical phase correspondence between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, based on whether the i-th candidate structure band in the current image to be registered is located in a helical segment, the helical direction information of the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, and the phase offset between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, includes: When the i-th candidate structure band in the current image to be registered is located in the spiral segment, the spiral phase correspondence between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered is determined based on the spiral direction information of the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, and the phase offset between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered. If the i-th candidate structure band in the current image to be registered is not located in the spiral segment, the spiral phase correspondence between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered is determined to be 1.

7. The image registration method for high-precision measurement of the main shaft dimension of a missile engine according to claim 6, characterized in that, The step of determining the spiral phase correspondence between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, based on the spiral orientation information of the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, and the phase offset between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, includes: The spiral phase correspondence is determined based on the formula for solving the spiral phase correspondence; the formula for solving the spiral phase correspondence is as follows: In the formula, This represents the spiral phase correspondence between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered. This indicates the consistency of the rotation direction between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered. This represents the phase offset between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered.

8. The image registration method for high-precision measurement of the main shaft dimension of a missile engine according to claim 1, characterized in that, The determination of the measurement fidelity registration degree between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, based on the size sensitivity salience corresponding to the i-th candidate structure band in the current image to be registered, the axial sequence continuity between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, and the spiral phase correspondence between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered, includes: Based on the formula for calculating measurement fidelity registration, the measurement fidelity registration is determined; the formula for calculating measurement fidelity registration is: In the formula, This represents the measurement fidelity registration between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered. This represents the size-sensitive saliency of the i-th candidate structural band in the current image to be registered. This represents the axial sequential continuity between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered. This represents the spiral phase correspondence between the i-th candidate structure band in the current image to be registered and the j-th candidate structure band in the next image to be registered.

9. The image registration method for high-precision measurement of the main shaft dimension of a missile engine according to claim 1, characterized in that, The structural band attributes include the axial interval in which the structural band is located, the local spacing between the two sides, the continuous length of the two sides, the edge polarity, the local transition curvature, the front-to-back order with adjacent structural bands, and whether it is located in the helical segment region.

10. An image registration system for high-precision measurement of the main shaft dimensions of a missile engine, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When a computer program is executed by a processor, it implements the steps of the image registration method for high-precision measurement of the main shaft dimension of a missile engine as described in any one of claims 1-9.

Images