A method and system for monitoring changes in lumbar spine during sitting based on lumbar pelvic flexion distribution.
By using the lumbar-pelvic flexion distribution transition method, the contributions of posterior pelvic tilt and lumbar flattening are separated, the flexion distribution difference sequence is calculated, the lumbar-pelvic distribution transition position is determined, and the position of lumbar muscle activity reduction is converted. This solves the problem of difficulty in distinguishing the timing of lumbar muscle activity changes in existing technologies and generates an effective lumbar change monitoring sequence.
Patent Information
- Application Number
- CN202611577329.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-10-13
- Publication Date
- 2026-11-10
AI Technical Summary
Existing technologies struggle to effectively differentiate the timing of changes in lumbar muscle activity and the relative changes in the pelvis and lumbar spine during sustained sitting, making it difficult to identify change segments with value for monitoring lower back pain.
By acquiring lumbar electromyography signals, pelvic tilt sequence, and lumbar curvature sequence, the contribution of posterior pelvic tilt and lumbar flattening is separated using the lumbar flexion distribution transition method. The flexion distribution difference sequence is calculated to determine the lumbar flexion distribution transition position. The position of reduced lumbar muscle activity is converted into the relative flexion process offset to generate a lumbar flexion change monitoring sequence.
It was able to distinguish the relative changes in the pelvis and lumbar spine and the timing of muscle activity during the same flexion process, and generated monitoring results of lumbar changes over time, while retaining the cases where lumbar muscle activity did not decrease accordingly as monitoring records.
Smart Images

Figure CN123058673A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of lumbar motion measurement technology, and more specifically, this application relates to a method and system for monitoring lumbar changes in sitting posture based on lumbar pelvic flexion distribution transition. Background Technology
[0002] Prolonged sitting has become a common posture in office, study, and home settings, and lower back discomfort and pain are common during continuous sitting. These changes in lower back condition are usually slow and intermittent, and subjective discomfort reports may be delayed, while the overall sitting posture may not change significantly at the same time. Therefore, it is necessary to objectively and continuously monitor changes in the lower back during continuous sitting, and to screen out changes with monitoring value from frequent daily posture adjustments to provide a basis for the analysis of lower back pain.
[0003] Current continuous monitoring typically extracts features from surface electromyography (EMG) of the lumbar region or posture signals of the trunk and pelvis. For example, it reflects lumbar muscle activity based on changes in EMG amplitude, or records sitting posture adjustments based on posture angles and their changes. During lumbar pelvic flexion, current treatments focus more on whether muscle activity decreases or the degree of overall posture change, and pay insufficient attention to the relative changes in the pelvis and lumbar spine during the same flexion, as well as the specific stages at which muscle activity changes during flexion.
[0004] However, real continuous sitting posture includes various changes such as comfort adjustments, changes in external support, and habitual activities. Different changes may show similar electromyographic amplitudes or degrees of lumbar pelvic flexion. Even if the overall flexion degree is similar, the relative changes of the pelvis and lumbar spine and the timing of changes in lumbar muscle activity may be different. It is difficult to screen out the change segments with low back pain monitoring value from the messy sitting posture changes based solely on electromyographic changes or overall posture changes, and to form continuous monitoring results organized over time. Therefore, a sitting posture low back change monitoring method and system based on lumbar pelvic flexion distribution transition is proposed to solve this problem. Summary of the Invention
[0005] To address the aforementioned technical problems, a method and system for monitoring lumbar changes in sitting posture based on lumbar flexion distribution and turning is provided. This technical solution solves the problems mentioned in the background technology.
[0006] To achieve the above objectives, the technical solution of the present invention is as follows:
[0007] In a first aspect, this application provides a method for monitoring changes in lumbar spine during sitting based on lumbar-pelvic flexion distribution transition, the method comprising:
[0008] Acquire lumbar electromyography (EMG) signals, pelvic tilt sequence, and lumbar curvature sequence of subjects in a continuous sitting posture, and align them by time.
[0009] The posterior tilt change of the pelvic tilt sequence and the lordosis reduction of the lumbar curvature sequence are extracted and normalized into posterior tilt component and flattening component, respectively. The two components are added together at each time step to obtain the total change sequence. The interval where the total change sequence is continuously positive and the cumulative value reaches the preset flexion threshold is determined as the candidate flexion segment. The lumbar pelvic flexion amount is obtained by accumulating the total change sequence from the starting point of the segment.
[0010] Within the candidate buckling segment, the difference in the proportion of the backward tilt component and the flattening component is calculated according to a preset window to form a buckling distribution difference sequence. When the buckling distribution difference sequence changes from continuous positive values to continuous negative values or reverses, and the absolute values of the buckling distribution difference on both sides of the change reach the preset distribution difference, the first positive-negative boundary position is determined as the waist-pelvis distribution turning point, and the corresponding candidate buckling segment is determined as the analysis segment.
[0011] The lumbar electromyography signal within the analysis segment is rectified and the envelope is extracted to obtain the lumbar muscle activity amplitude sequence. Within the preset search range before and after the lumbar-pelvic junction transition position, the average amplitude of the lumbar muscle activity amplitude sequence within the preset time period at the beginning of the analysis segment is used as the benchmark. The position where the sequence first enters a state that is continuously lower than the benchmark and the decrease reaches the preset decrease is determined as the lumbar muscle activity reduction position. If it is not identified, a reduction missing marker is generated.
[0012] When determining the location of reduced lumbar muscle activity, the lumbar flexion amount corresponding to the location and the lumbar-pelvic distribution turning point is divided by the lumbar flexion amount at the end of the analysis segment, and the signed process offset is obtained by subtracting the second ratio from the first ratio.
[0013] Analysis segments with signed process offsets exceeding the preset offset range or with reduced missing markers are identified as monitoring candidate segments. Each monitoring candidate segment is associated with its lumbar-pelvic allocation turning point, signed process offset, or reduced missing marker in the order of appearance to obtain the lumbar-pelvic change monitoring sequence.
[0014] Secondly, this application provides a sitting lumbar change monitoring system based on lumbar flexion distribution transition, used to implement the above-mentioned sitting lumbar change monitoring method based on lumbar flexion distribution transition, including:
[0015] The multi-source acquisition module is used to acquire lumbar electromyography signals, pelvic tilt sequence, and lumbar curvature sequence of the subject in a continuous sitting posture, and align them by time.
[0016] The flexion quantification module is used to extract the posterior tilt change of the pelvic tilt sequence and the lordosis reduction of the lumbar curvature sequence, which are normalized into posterior tilt component and flattening component, respectively. The two components are added together at each time step to obtain the total change sequence. The interval where the total change sequence is continuously positive and the cumulative value reaches the preset flexion threshold is determined as the candidate flexion segment. The lumbar flexion amount is obtained by accumulating the total change sequence from the starting point of the segment.
[0017] The buckling distribution transition module is used to calculate the difference in the proportion of the backward tilt component and the flattening component within a preset window in the candidate buckling segment, forming a buckling distribution difference sequence. When the buckling distribution difference sequence changes from continuous positive values to continuous negative values or reverses, and the absolute values of the buckling distribution differences on both sides of the transition reach the preset distribution difference, the first positive-negative boundary position is determined as the waist-pelvis distribution transition position, and the corresponding candidate buckling segment is determined as the analysis segment.
[0018] The electromyography response module is used to rectify and extract the envelope of the lumbar electromyography signal within the analysis segment to obtain the lumbar muscle activity amplitude sequence. Within the preset search range before and after the lumbar-pelvic junction transition position, the average amplitude of the lumbar muscle activity amplitude sequence within the preset time period at the beginning of the analysis segment is used as the benchmark. The position where the sequence first enters a state that is continuously lower than the benchmark and the decrease reaches the preset decrease is determined as the lumbar muscle activity reduction position. If it is not identified, a reduction missing marker is generated.
[0019] The process offset module is used to determine the location of reduced lumbar muscle activity by dividing the lumbar flexion amount corresponding to the location and the lumbar-pelvic assignment turning point by the lumbar flexion amount at the end of the analysis segment, and subtracting the latter ratio from the former ratio to obtain the signed process offset.
[0020] The monitoring sequence module is used to identify analysis segments with signed process offsets exceeding the preset offset range or with reduced missing markers as monitoring candidate segments. Each monitoring candidate segment is associated with its waist-pelvis allocation turning position, signed process offset, or reduced missing marker in the order of appearance to obtain the waist-pelvis change monitoring sequence.
[0021] Thirdly, this application provides a computer device, the computer device including a memory and a processor, the memory storing code, the processor being configured to acquire the code and execute the above-described method for monitoring changes in lumbar posture based on lumbar flexion distribution transition.
[0022] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for monitoring changes in lumbar posture based on lumbar flexion distribution.
[0023] Compared with the prior art, the beneficial effects of the present invention are as follows:
[0024] This application solves the problem that existing continuous monitoring methods, which rely on whether electromyography amplitude decreases or the overall degree of flexion, make it difficult to distinguish between different sitting posture changes with similar overall flexion but different relative changes in the pelvis and lumbar spine and different muscle activity timings. This is achieved by separating the contributions of pelvic tilt and lumbar flattening within the same flexion process and adaptively determining the lumbar-pelvic distribution turning position based on the difference in their proportions. The application also converts the position of reduced lumbar muscle activity into a signed process offset relative to that position using the amount of lumbar flexion.
[0025] This application addresses the problem in existing schemes that discard cases where the motor structure has changed but the electromyography has not decreased accordingly, by generating a reduced missing marker when the lumbar-pelvic junction transition has occurred but no decrease in lumbar muscle activity has occurred in the vicinity, and by associating this marker with the transition location and the sequence of the stage process as a type of monitoring information. This ensures that the corresponding missing situation is retained as a type of monitoring record that can be inherited over time, rather than being excluded from the monitoring results. Attached Figure Description
[0026] The disclosure of this invention is illustrated with reference to the accompanying drawings. It should be understood that the drawings are for illustrative purposes only and are not intended to limit the scope of protection of this invention. Wherein:
[0027] Figure 1 This is a flowchart of the sitting posture lumbar change monitoring method based on lumbar pelvic flexion distribution transition proposed in this invention;
[0028] Figure 2 This is a structural block diagram of the sitting posture lumbar change monitoring system based on lumbar pelvic flexion distribution transition proposed in this invention. Detailed Implementation
[0029] It is readily understood that, based on the technical solution of this invention, those skilled in the art can propose various interchangeable structural methods and implementations without altering the essential spirit of the invention. Therefore, the following detailed embodiments and accompanying drawings are merely illustrative examples of the technical solution of this invention and should not be considered as the entirety of the invention or as limitations or restrictions on the technical solution of this invention.
[0030] For ease of understanding, the following explanation uses a continuous sitting posture monitoring of a subject who has been working at a desk for extended periods as an example. Assume the subject wore a back and waist surface sensor for two consecutive hours of work, with a sampling rate of 100 Hz. Multiple posture adjustments occurred during this time. This example selects one instance of flexion, from a relatively upright sitting posture to a hunched-over sitting posture, as a typical case. This flexion lasted approximately 12 seconds, with the pelvis gradually tilting backward from approximately 5 degrees to approximately 18 degrees, and the lumbar lordosis angle gradually decreasing from approximately 35 degrees to approximately 12 degrees. This method is applicable to continuous monitoring of the lumbar region in sedentary individuals (office workers, students, and those sitting at home) and during lumbar pain rehabilitation. The monitoring results are provided for clinical or rehabilitation personnel to interpret in conjunction with symptoms, identifying noteworthy lumbar changes during prolonged sitting, assisting in lumbar pain assessment, and tracking the rehabilitation process.
[0031] like Figure 1 As shown, in step S1, the lumbar electromyography signal, pelvic tilt sequence and lumbar curvature sequence of the subject in a continuous sitting posture are acquired and aligned by time.
[0032] It should be noted that the sources and acquisition methods of the three signals are as follows: the lumbar electromyography (EMG) signal is collected by surface EMG electrodes attached to the erector spinae muscles of the lower back of the subject, reflecting the electrical activity intensity of the active muscles in the lumbar region, and the original signal is a continuous voltage sequence over time; the pelvic tilt sequence is output by an inertial measurement unit fixed to the sacrum or the posterior side of the pelvis, taking the sequence of its pitch angle in the sagittal plane over time, with the posterior pelvic tilt direction as positive; the lumbar curvature sequence is obtained by calculating the relative posture of a flexible bending sensor placed on the surface of the lumbar spine or two inertial measurement units placed in the thoracolumbar and sacral segments, characterizing the magnitude of lumbar lordosis, with the direction of decreasing lordosis (flattening of the lumbar spine) as positive; each of the three signals is time-stamped based on the clock of the same acquisition device, and the three signals are resampled to a common sampling time grid based on this unified clock, so that at any given time, the values of EMG amplitude, pelvic tilt angle, and lumbar curvature are simultaneously available. It should be noted that if a signal experiences packet loss or abrupt changes at certain times, it will be bridged by linear interpolation using its adjacent valid sampling points; if the continuous missing time exceeds a preset duration (e.g., more than 0.5 seconds), the time period will be marked as invalid and will not be included in subsequent analysis to avoid contaminating the subsequent allocation relationship calculation.
[0033] To further clarify, the pelvic tilt angle and lumbar curvature signals were selected instead of the single signal of overall trunk tilt angle because subsequent steps require distinguishing the contributions of the pelvis and lumbar spine in the same flexion. If only the overall trunk tilt angle is collected, it is impossible to separate the relative changes of the pelvis and lumbar spine within a single flexion, and the subsequent distribution transitions will be impossible to calculate.
[0034] In step S2, the posterior tilt change of the pelvic tilt sequence and the lordosis reduction of the lumbar curvature sequence are extracted and normalized into posterior tilt component and flattening component, respectively. The two components are added together at each time step to obtain the total change sequence. The interval where the total change sequence is continuously positive and the cumulative value reaches the preset flexion threshold is determined as the candidate flexion segment. The lumbar pelvic flexion amount is obtained by accumulating the total change sequence from the starting point of the segment.
[0035] It should be noted that the method for obtaining the change in posterior tilt is as follows: the difference between adjacent sampling times of the pelvic tilt sequence is taken, where the increment in the positive direction of posterior tilt is recorded as the corresponding value of the change in posterior tilt, and the change in the direction of anterior tilt is recorded as zero; similarly, the decrease in lordosis is obtained by taking the difference in the lumbar curvature sequence in the direction of decreasing lordosis, and the change in the direction of increasing lordosis is recorded as zero; the normalization of both is used to eliminate the difference in dimensions and amplitude scales. Although both pelvic tilt and lumbar curvature are measured in degrees, their range of motion in one flexion is different (the range of motion of posterior pelvic tilt is usually more than ten degrees, while the decrease in lumbar lordosis can reach more than twenty degrees). If they are directly added together, the one with the larger range of motion will dominate.
[0036] Specifically, the difference between the maximum and minimum values of the pelvic tilt sequence during the current data collection period is used as the posterior tilt normalization scale. The posterior tilt component is obtained by dividing the change in posterior tilt at each time point by the posterior tilt normalization scale. The difference between the maximum and minimum values of the lumbar curvature sequence during the current data collection period is used as the flattening normalization scale. The flattening component is obtained by dividing the decrease in lordosis at each time point by the flattening normalization scale. Both the posterior tilt component and the flattening component are dimensionless quantities and can be added together at the same scale to form a total change sequence. The total change sequence represents the speed of the overall lumbar and pelvic flexion progression. When any normalization scale is zero, the corresponding component is set to zero.
[0037] The preset buckling threshold is a cumulative threshold used to determine whether a continuous buckling segment is worthy of being considered as a candidate. It means the lower limit of the cumulative value of the total change sequence within a continuously positive interval. Since too small a cumulative buckling does not constitute a meaningful posture adjustment and may easily misjudge breathing or micro-movement as buckling, while too large a threshold may miss buckling with small amplitude but intact structure, this threshold is set to filter out micro-fluctuations. The value range is 0.2 to 0.5 of the normalized cumulative total change, with a typical value of 0.3. The method for obtaining the lumbar pelvic buckling amount is as follows: starting from the starting point of the candidate buckling segment, the total change sequence is accumulated moment by moment to obtain a cumulative value sequence that monotonically increases with time. This sequence serves as a common benchmark for subsequently converting each key position into a progress position.
[0038] Using the previous example, the cumulative change in posterior pelvic tilt during this flexion, after normalization, is approximately 0.7, and the cumulative decrease in lumbar lordosis, after normalization, is approximately 0.9. The total change sequence obtained by adding them up moment by moment is positive for approximately 12 seconds, with a cumulative value of approximately 1.6, which exceeds the preset flexion threshold of 0.3. Therefore, it is identified as a candidate flexion segment. The lumbar-pelvic flexion amount accumulated from the beginning of the segment reaches approximately 1.6 at the end of the segment.
[0039] In step S3, the difference between the proportions of the backward tilt component and the flattening component is calculated according to a preset window within the candidate buckling segment to form a buckling distribution difference sequence. When the buckling distribution difference sequence changes from continuous positive values to continuous negative values or reverses, and the absolute values of the continuous differences on both sides of the change reach the preset distribution difference, the first positive-negative boundary position is determined as the waist-pelvis distribution turning point, and the corresponding candidate buckling segment is determined as the analysis segment.
[0040] It should be noted that the flexion distribution difference sequence characterizes which of the two—posterior pelvic tilt and lumbar flattening—is dominant at each stage of flexion progression: a positive sequence indicates that posterior pelvic tilt is dominant at the current stage, while a negative sequence indicates that lumbar flattening is dominant. The lumbar-pelvic distribution turning point is the point where this dominant relationship reverses, and its physical meaning is the turning point of a flexion from a stage dominated by pelvic movement to a stage dominated by lumbar movement (or vice versa).
[0041] Furthermore, a significant potential problem during implementation is that a large number of false positive-negative boundaries can be generated when the flexion distribution difference sequence exhibits slight jitter near zero. Therefore, this scheme sets a condition that the absolute values of the continuous differences on both sides of the transition must reach a preset distribution difference. This preset distribution difference is an amplitude threshold used to confirm that a dominant relationship reversal is a genuine structural transition rather than noise jitter. Since reversals smaller than this amplitude are mostly caused by sensor noise or micro-movements and do not represent a substantial exchange of dominant relationships between the pelvis and lumbar spine, the absolute values of the distribution differences on both sides of the reversal must reach this threshold to be confirmed as a lumbar-pelvic distribution transition position. The value range is 0.1 to 0.3 of the normalized proportional difference, with a typical value of 0.15. When there are no positive-negative boundaries within a candidate flexion segment that meet this condition, the candidate flexion segment is determined not to have formed a lumbar-pelvic distribution transition position, is not identified as an analysis segment, and subsequent processing of the candidate flexion segment is terminated.
[0042] In step S4, the lumbar electromyography signal within the analysis segment is rectified and the envelope is extracted to obtain the lumbar muscle activity amplitude sequence. Within the preset search range before and after the lumbar-pelvic junction transition position, the average amplitude of the lumbar muscle activity amplitude sequence within the preset time period at the beginning of the analysis segment is used as the benchmark. The position where the sequence first enters a state that is continuously lower than the benchmark and the decrease reaches the preset decrease is determined as the lumbar muscle activity reduction position. If it is not identified, a reduction missing marker is generated.
[0043] It should be noted that rectification and envelope extraction are standard procedures in this field: rectification uses the full-wave rectified value of the lumbar electromyography signal; envelope extraction uses a sliding root mean square or low-pass filter to obtain a sequence of lumbar muscle activity amplitudes reflecting the change in lumbar muscle activity intensity over time. The benchmark is obtained by taking the average value of the lumbar muscle activity amplitude sequence within a preset time period (e.g., 0.5 to 1 second from the beginning of the segment) as a reference for the level of lumbar muscle activity at the start of the flexion. The reason for taking the first segment's average instead of the entire segment's average is that what needs to be determined subsequently is the decrease in muscle activity relative to the start of flexion; if the entire segment's average is used as the benchmark, the decrease will be overlooked. The inclusion of a baseline weakens the sensitivity of the judgment; the preset search range is a process interval extending forward and backward around the lumbar-pelvic distribution transition position, used to limit the search for reduced lumbar muscle activity within which flexion process near the lumbar-pelvic distribution transition position; for example, with the lumbar-pelvic distribution transition position as the center, the search span is 20% of the lumbar-pelvic flexion amount at the end of the analysis segment forward and backward respectively, and the part exceeding the start or end of the analysis segment is truncated to the corresponding segment boundary, thus forming the preset search range. Using a search span based on the lumbar-pelvic flexion amount can keep the process meaning consistent under different flexion durations.
[0044] To further clarify, the determination of the location of decreased lumbar muscle activity adopts the condition of continuous decrease relative to the baseline rather than instantaneous low point, so as to avoid misjudging transient electromyographic fluctuations as decreased muscle activity; when no location that meets the condition of continuous decrease is found after traversing the preset search range, a decrease missing marker is generated, which indicates that a lumbar-pelvic distribution transition occurred during this flexion but no corresponding decrease in lumbar muscle activity was observed in its vicinity.
[0045] It should be noted that the reduction in loss itself is retained as a monitoring information and passed on to subsequent steps, rather than being discarded as an invalid result. Even if a turning point has occurred in the motor structure but muscle activity has not decreased accordingly, this corresponding loss situation is also valuable for judging the condition of the waist.
[0046] In step S5, when determining the position of reduced lumbar muscle activity, the lumbar flexion amount corresponding to the position and the lumbar-pelvic distribution turning point are respectively divided by the lumbar flexion amount at the end of the analysis segment. The signed process offset is obtained by subtracting the latter ratio from the former ratio.
[0047] It should be noted that the lumbar pelvic flexion amount at the end of the analysis segment, i.e., the lumbar pelvic flexion amount obtained in step S2, is used as the total advancement amount of this flexion to convert each key position on the time axis into a process position independent of time. Since the same flexion may last for different durations, directly measuring the order of the relative transition of lumbar muscle activity reduction by the time difference will be affected by the speed of flexion. Therefore, the difference is calculated by the ratio of the lumbar pelvic flexion amount corresponding to each position to the total flexion amount (i.e., the process position). The signed process offset represents whether the lumbar muscle activity reduction is ahead or behind the lumbar pelvic distribution transition in the flexion process, regardless of the time. A positive signed process offset indicates that the lumbar muscle activity reduction is behind the lumbar pelvic distribution transition in the process, and a negative one indicates that it is ahead. When there is a reduction missing marker, this conversion is not performed and no signed process offset is generated. This situation is propagated along the reduction missing marker.
[0048] Using the previous example, the lumbar-pelvic distribution transition point of this flexion corresponds to a lumbar-pelvic flexion amount of approximately 1.0 (from pelvic dominance to lumbar dominance), the lumbar muscle activity reduction point corresponds to a lumbar-pelvic flexion amount of approximately 1.3, and the lumbar-pelvic flexion amount at the end of the analysis segment is approximately 1.6. The process ratio corresponding to the lumbar-pelvic distribution transition point is approximately 1.0 divided by 1.6, which is approximately 0.63, and the process ratio corresponding to the lumbar muscle activity reduction point is approximately 1.3 divided by 1.6, which is approximately 0.81. Subtracting the previous ratio from the latter yields a signed process offset of approximately 0.18 (positive value), indicating that the lumbar muscle activity reduction lags behind the lumbar-pelvic distribution transition by approximately 0.18 process units in the flexion process.
[0049] In step S6, analysis segments with signed process offsets exceeding the preset offset range or with reduced missing markers are identified as monitoring candidate segments. Each monitoring candidate segment is associated with its waist-pelvis allocation turning position, signed process offset, or reduced missing marker in the order of appearance to obtain a waist-pelvis change monitoring sequence.
[0050] It should be noted that the preset offset range is used to filter whether the marked process offset falls outside the common correspondence. It means that the offset between the decrease in lumbar muscle activity and the lumbar-pelvic distribution transition is considered to be the upper and lower bounds of the regular correspondence in the process. Since the offset falling within this range corresponds to the decrease in lumbar muscle activity and the lumbar-pelvic distribution transition being roughly in tandem in the process and is a common situation, while the situation that exceeds this range (excessively leading or excessively lagging) and the decrease is missing are more worthy of being included in the monitoring, this range is used as the screening boundary, with the lower bound of the value range being -0.2 to -0.1 and the upper bound being 0.2 to 0.4, and the typical range being -0.15 to 0.3. The field structure of the lumbar-pelvic change monitoring sequence is as follows: arranged according to the order of appearance of each monitoring candidate segment in continuous sitting posture, each monitoring candidate segment is associated with its lumbar-pelvic distribution transition position, and the marked process offset (when determining the position of the decrease in lumbar muscle activity) or the decrease missing mark (when the decrease is missing), and the start time of the monitoring candidate segment is used as its time label.
[0051] To further clarify, the lumbar pelvic change monitoring sequence is the continuous monitoring result output by this method to the user. It organizes a series of noteworthy lumbar changes during prolonged sitting, allowing clinicians or rehabilitation personnel to interpret the changing trend of the lumbar condition in conjunction with the test subject's symptoms. This method does not directly provide a judgment on whether there is pain.
[0052] Through the above technical solution, this embodiment separates the contributions of pelvic tilt and lumbar flattening within a single flexion, adaptively locates the lumbar-pelvic distribution transition position by flipping it according to their dominant relationship, and uses this transition as the process anchor point to measure the signed process offset of reduced lumbar muscle activity, thereby filtering the chaotic sitting posture changes into a time-organized lumbar-pelvic change monitoring sequence. Within the process of the same flexion, it processes the two-layer relationship of the relative change relationship between the pelvis and lumbar spine and the timing of muscle activity changes, so that sitting posture changes with similar overall flexion degree but different internal structures or response timings can be distinguished. The lumbar-pelvic change monitoring sequence obtained in this section, as well as the lumbar-pelvic distribution transition position, signed process offset or reduction missing marker associated with each monitoring candidate segment, are used as the basic data for subsequent refinement processing and sequence organization.
[0053] Based on step S3 above, the difference in the proportion of the backward tilt component and the flattening component is calculated within a preset window within the candidate buckling segment to form a buckling distribution difference sequence, specifically including:
[0054] It should be noted that the preset window is constructed based on the advance of lumbar-pelvic buckling, rather than a fixed time length. The starting point of the candidate buckling segment is recorded as the first process boundary. From this starting point, candidate buckling segments are read along the increasing direction of lumbar-pelvic buckling. Whenever the lumbar-pelvic buckling is increased by a preset buckling increment compared to the lumbar-pelvic buckling corresponding to the previous process boundary, the corresponding time point is recorded as the next process boundary. A process unit is formed by the time interval between two adjacent process boundaries. The preset buckling increment is used to determine the buckling advance covered by each process unit. Since an increment that is too small will result in an inflated number of process units, and an increment that is too large will make the subsequent dominant relationship statistics too coarse, its value range is 1 / 10 to 1 / 5 of the lumbar-pelvic buckling at the end of the candidate buckling segment, with a typical value of 1 / 8.
[0055] To further explain, a preset window is formed by grouping consecutive preset window units into a single preset window, and each time the window moves one process unit towards the end of the segment to form the next preset window. The number of preset window units is used to determine the span of the flexion process covered each time the relative contribution of the pelvis and lumbar spine is statistically analyzed. Since a span that is too small will cause the window-by-window ratio to be unstable due to local fluctuations, and a span that is too large will mask the changes in the dominant relationship near the turning point, its value ranges from 2 to 4 process units, with a typical value of 3 process units. Within each preset window, the tilt component and the flattening component are accumulated to obtain the window tilt amount and the window flattening amount, respectively. The two are added together to obtain the window total measurement. The tilt proportion is obtained by dividing the window tilt amount by the window total measurement, and the flattening proportion is obtained by dividing the window flattening amount by the window total measurement. The sum of the two proportions is one. The tilt proportion minus the flattening proportion is used as the flexion distribution difference value for the corresponding preset window. The values are arranged according to the movement order of the preset windows to obtain the flexion distribution difference sequence. Since the two proportions originate from the same window and their difference is dimensionless, the values can be directly compared along the window order without needing to be normalized.
[0056] For example, the following is a numerical implementation process throughout this step: using the example of approximately 1.6 lumbar-pelvic buckling at the end of the aforementioned candidate buckling segment, taking a preset buckling increment of 0.2, approximately 8 process units can be formed; taking a preset window unit number of 3, approximately 6 preset windows can be formed by moving one process unit at a time. Within a pre-defined window in the anterior flexion phase, the window tilt is approximately 0.42, the window flattening is approximately 0.18, and the total window value is approximately 0.60. This corresponds to a tilt ratio of approximately 0.70% and a flattening ratio of approximately 0.30%, with a flexion distribution difference of approximately +0.40, indicating that posterior pelvic tilt dominates in this stage. Within a pre-defined window in the posterior flexion phase, the window tilt is approximately 0.21, the window flattening is approximately 0.39, and the total window value is approximately 0.60, corresponding to a tilt ratio of approximately 0.35% and a flattening ratio of approximately 0.65%, with a flexion distribution difference of approximately -0.30, indicating that lumbar flattening dominates in this stage. The resulting flexion distribution difference sequence shows a transition from the positive value region through the vicinity of zero to the negative value region.
[0057] Through the above technical solution, this embodiment constructs a flexion distribution difference sequence by using the difference in the proportion of the two components of the local window of the process, and quantifies the dominant relationship between posterior pelvic tilt and lumbar flattening within the same flexion as the process evolves into a directed sequence with interpretable positive and negative boundaries; this flexion distribution difference sequence is used as the direct input for determining the lumbar-pelvic distribution turning point in step S3.
[0058] In an optional embodiment, before determining the corresponding candidate buckling segment as the analysis segment, the method further includes determining the post-buckling observation process, specifically including:
[0059] It should be noted that the post-turn observation process determination directly uses the process units formed based on the preset flexion increment, without re-dividing the candidate flexion segment. The execution process is as follows: First, locate which process unit the lumbar-pelvic allocation turn position falls into. Starting from the next process unit after that, count the number of process units formed along the direction of increasing lumbar-pelvic flexion until the end of the candidate flexion segment, obtaining the number of units after the turn. Compare the number of units after the turn with the preset number of observation units. When the preset number of observation units is reached, it indicates that there is still sufficient flexion advancement after the turn for observing the reduction in lumbar muscle activity, and the candidate flexion segment is determined as the analysis segment; otherwise, the candidate flexion segment is not determined as the analysis segment. The preset number of observation units is the lower limit of the number of units to confirm whether the post-turn observation is sufficient. Since if the turn is too close to the end of the candidate flexion segment, there is insufficient time to observe the corresponding reduction in lumbar muscle activity, this lower limit is set, with a value range of 2 to 4 process units, and a typical value of 3 process units. It should be emphasized that this determination only requires sufficient observable progress after the transition, and does not require that a decrease in lumbar muscle activity necessarily occur thereafter; if the observable progress is sufficient but no decrease in lumbar muscle activity occurs in the analysis segment, a decrease missing marker will still be generated in step S4 as usual.
[0060] Through the above technical solution, this embodiment measures the observable advance after the lumbar-pelvic distribution transition using a process unit with equal flexion amount and sets a lower limit for the number of observation units. Before determining the analysis segment, candidate flexion segments with transitions too close to the end of the segment and insufficient for observing the reduction of lumbar muscle activity are excluded. The key difference between this and the existing approach that only sets thresholds based on duration or amplitude is that the sufficiency of observation is determined by the flexion advance amount rather than time, and objects are not excluded due to lack of electromyographic response. The analysis segment determined by this determination is used as the processing object for locating the position of reduced lumbar muscle activity in step S4.
[0061] Based on step S4 above, the position where the lumbar muscle activity amplitude sequence first enters a state where it is continuously lower than the baseline and the decrease reaches a preset decrease value is determined as the lumbar muscle activity reduction position. If it is not identified, a reduction loss marker is generated, specifically including:
[0062] It should be noted that the determination of a state where the amplitude of lumbar muscle activity is continuously lower than the benchmark and the reduction reaches the preset reduction is as follows: starting from the preset search range, the amplitude sequence of lumbar muscle activity is traversed along time. When the amplitude at a certain sampling position is lower than the benchmark and the magnitude of the reduction reaches the preset reduction, the position is recorded as a candidate starting point of low activity. From the candidate starting point of low activity, the amplitude sequence of lumbar muscle activity is required to be continuously lower than the benchmark and maintain the reduction target. At the same time, the increase in lumbar pelvic flexion during this continuous period is required to reach the preset reduction in continuous flexion amount. Only then is the candidate starting point of low activity confirmed as the position of reduced lumbar muscle activity. The preset reduction amplitude is the threshold for determining whether the electromyographic amplitude has substantially decreased. Since small fluctuations do not represent a substantial decrease in muscle activity, its value ranges from 15% to 35% of the baseline amplitude, with a typical value of 25%. The preset reduction duration of flexion is the threshold that requires the reduced state to be sustained for a sufficient length of flexion propulsion to exclude instantaneous descent. Since the duration is measured by the amount of flexion propulsion rather than time, it is not affected by the speed of flexion, so its value ranges from 1 / 10 to 1 / 4 of the lumbar-pelvic flexion at the end of the analysis segment, with a typical value of 1 / 6.
[0063] To further clarify, the anomaly and branching process is as follows: When a candidate low-activity starting point is identified, if the lumbar muscle activity amplitude sequence returns to the baseline or the decrease no longer meets the target before the flexion amount increases to reach the preset reduction in sustained flexion amount, the candidate low-activity starting point is determined to have exited midway, is cancelled, and the search continues along time for the next candidate low-activity starting point; if the lumbar muscle activity reduction location is not identified after traversing the preset search range, a reduction missing marker is generated. Since the location that first meets the sustainability condition can avoid interference from subsequent occasional secondary electromyographic fluctuations, the first identified location is taken as the lumbar muscle activity reduction location.
[0064] Through the above technical solution, this embodiment confirms the location of lumbar muscle activity reduction by using the continuous decrease relative to the segment head reference and adding the continuous condition measured by flexion propulsion, distinguishing instantaneous electromyographic fluctuations from true continuous muscle activity reduction, while retaining the situation where the movement transition has occurred but the electromyographic activity has not decreased accordingly by using a reduction missing marker; the determined location of lumbar muscle activity reduction is used as the input for calculating the signed process offset in step S5, and the conversion of the signed process offset is not performed when generating the reduction missing marker, and the reduction missing marker is directly passed backward and participates in the determination of monitoring candidate segments.
[0065] In an optional embodiment, after identifying the corresponding candidate flexion segment as the analysis segment, the method further includes lumbar spine involvement status identification, specifically including:
[0066] It should be noted that this identification method continues to process the analysis segment window by window using the aforementioned preset windows composed of continuous process units: within each preset window, the tilt component and the flattening component are accumulated to obtain the window tilt amount and the window flattening amount, and the two are added together to obtain the window summation; when both the window tilt amount and the window flattening amount are positive, the corresponding preset window is marked as a common advancement state, indicating that the pelvic tilt and lumbar spine flattening in this window continue to advance along the flexion direction; when both the window summation and the window tilt amount are positive and the window flattening amount is not greater than zero, the corresponding preset window is marked as a disengagement state, indicating that the pelvic tilt is still advancing in this window but no further flattening of the lumbar spine is observed; the remaining preset windows are marked as transition states; the state markings are arranged in window order to obtain the lumbar spine participation state sequence.
[0067] To further clarify, the determination of the lumbar vertebral dislocation location is as follows: A preset window corresponding to the lumbar-pelvic flexion transition position is located. Starting from the next preset window, the lumbar vertebral participation state sequence is read chronologically. When the number of consecutive common propulsion states reaches the preset number of common propulsion windows, and the number of consecutive dislocation propulsion states that occur afterward reaches the preset number of dislocation propulsion windows, the starting position of the preset window corresponding to the first dislocation propulsion state is determined as the lumbar vertebral dislocation location. This indicates that the lumbar vertebrae have transitioned from flexion with the pelvis to no longer participating in flexion propulsion. The preset number of common propulsion windows and the preset number of dislocation propulsion windows are used to confirm the continuity of the states before and after the transition. Because insufficient continuity can easily lead to misjudging occasional windows as state transitions, both preset windows are set to 2 to 4, with a typical value of 3 preset windows. If the lumbar vertebral dislocation location is still not determined at the end of the analysis segment, a non-dislocation marker is generated. The non-dislocation marker indicates that no lumbar vertebral dislocation location satisfying the above continuous state transition conditions has been observed, and does not mean that the lumbar vertebrae have always been in a common propulsion state throughout the entire analysis segment.
[0068] Through the above technical solution, this embodiment divides the lumbar spine participation state by using the dual component ratio of a preset window on the analysis segment and locates the lumbar spine disengagement position where the lumbar spine changes from joint propulsion to disengagement propulsion, thus characterizing the change in the degree of lumbar spine participation during a single flexion. The obtained lumbar spine disengagement position or non-disengagement marker is used as the basis for dividing the process before and after disengagement in step S6.
[0069] In an optional embodiment, after identifying analysis segments with signed process offsets exceeding a preset offset range or with reduced missing markers as monitoring candidate segments, the process further includes stage process construction, specifically including:
[0070] It should be noted that the phase process construction is performed on each monitoring candidate segment. Based on the sequence of the location of lumbar muscle activity reduction and the lumbar-pelvic junction transition, whether there is a reduction absence marker, whether there is a non-dislocation marker, and whether the lumbar dislocation location is between the transition and the process endpoint, the monitoring candidate segment is divided into several stages and normalized into a phase process sequence. The phase process sequence is a sequence composed of one or more process quantities after normalization of the lumbar-pelvic flexion at the end of the analysis segment, arranged sequentially. It is used to express the process structure within the monitoring candidate segment segment segmented by the transition, electromyographic reduction, and lumbar dislocation. All process quantities involved are obtained by dividing the difference in lumbar-pelvic flexion between corresponding two locations by the lumbar-pelvic flexion at the end of the analysis segment, and are dimensionless.
[0071] To further clarify, the determination of the process endpoint and branching process are as follows: For monitoring candidate segments where the location of lumbar muscle activity reduction is determined, if the location of lumbar muscle activity reduction is earlier than the lumbar-pelvic allocation transition point, the difference between the corresponding lumbar-pelvic flexion amounts is normalized and used as the preceding process quantity, which constitutes a separate stage process sequence; for monitoring candidate segments where the location of lumbar muscle activity reduction is determined, if the location of lumbar muscle activity reduction is not earlier than the lumbar-pelvic allocation transition point, the location of lumbar muscle activity reduction is used as the process endpoint; for monitoring candidate segments with a reduction loss marker, the end of the analysis segment is used as the process endpoint. For monitoring candidate segments where the location of lumbar muscle activity reduction is not earlier than the lumbar-pelvic allocation transition point or has a reduction loss marker, the corresponding process endpoint is obtained, and segmentation based on lumbar dislocation location continues; for monitoring candidate segments where the location of lumbar muscle activity reduction is earlier than the lumbar-pelvic allocation transition point, the current stage process construction ends after forming a stage process sequence containing the preceding process quantity.
[0072] It should also be noted that the segmented normalization after obtaining the process endpoint is as follows: when there is an undislodged marker, or when the lumbar vertebral dislodged position is not located between the lumbar-pelvic distribution transition position and the process endpoint, the difference in lumbar-pelvic flexion from the lumbar-pelvic distribution transition position to the process endpoint is normalized and taken as a single segment process quantity, which constitutes the stage process sequence alone; when the lumbar vertebral dislodged position is located between the lumbar-pelvic distribution transition position and the process endpoint, the difference in lumbar-pelvic flexion from the lumbar-pelvic distribution transition position to the lumbar vertebral dislodged position and from the lumbar vertebral dislodged position to the process endpoint is normalized and taken as the pre-dislodged process quantity and the post-dislodged process quantity, which are arranged in chronological order to constitute the stage process sequence.
[0073] Through the above technical solution, this embodiment divides each monitoring candidate segment into different scenarios and normalizes them into a phase process sequence based on the location of lumbar muscle activity reduction, the location of lumbar-pelvic distribution transition, and the location of lumbar vertebral dislocation. The relative layout of the transition, electromyographic response, and lumbar vertebral dislocation within a single flexion is expressed as a comparable process structure. The key difference lies in the different phase process construction paths adopted according to the sequential relationship between the location of lumbar muscle activity reduction and the location of lumbar-pelvic distribution transition, as well as the cases of reduction absence. This ensures that lumbar muscle activity reduction before the transition, lumbar muscle activity reduction not before the transition, and reduction absence can all form phase process sequences corresponding to their actual positional relationship. The obtained phase process sequence is used as a component of the monitoring process record generated in step S6.
[0074] In an optional embodiment, each monitoring candidate segment is associated with its lumbar-pelvic assignment transition position, signed process offset, or reduced missing marker in the order of appearance to obtain a lumbar-pelvic change monitoring sequence, specifically including:
[0075] It should be noted that the conversion of each process position is as follows: extract the lumbar pelvic flexion amount corresponding to the lumbar pelvic allocation turning point in each monitoring candidate segment, and divide it by the lumbar pelvic flexion amount at the end of the analysis segment to obtain the turning process position; when determining the lumbar vertebral dislocation position, divide the lumbar pelvic flexion amount corresponding to the lumbar vertebral dislocation position by the lumbar pelvic flexion amount at the end of the analysis segment to obtain the dislocation process position; for monitoring candidate segments that determine the position of lumbar muscle activity reduction, divide the lumbar pelvic flexion amount corresponding to the position of lumbar muscle activity reduction by the lumbar pelvic flexion amount at the end of the analysis segment to obtain the reduction process position; each process position is a dimensionless quantity, so that monitoring candidate segments with different durations and different total flexion amounts can be compared on the same process scale.
[0076] To further clarify, the monitoring process record is composed of one of two types of monitoring candidate segments: for monitoring candidate segments that determine the location of decreased lumbar muscle activity, the turning point, decreasing process, departure from process or not departure marker, signed process offset, and stage process sequence are associated as a monitoring process record; for monitoring candidate segments with a missing decreasing marker, the turning point, decreasing missing marker, departure from process or not departure marker, and stage process sequence are associated as a monitoring process record. This case does not include the signed process offset because step S5 does not generate it. Thus, the fields of the two types of records are clearly defined and do not conflict with each other.
[0077] It should also be noted that the lumbar and pelvic change monitoring sequence is arranged as follows: the start time of the analysis segment corresponding to each monitoring candidate segment is used as the recording time of its monitoring process record. All monitoring process records are arranged in chronological order of recording time to obtain the lumbar and pelvic change monitoring sequence organized by time. For example, the monitoring process record corresponding to a monitoring candidate segment that determines the position of decreased lumbar muscle activity is as follows: turning point position approximately 0.63, decreasing point position approximately 0.81, detachment point approximately 0.72, marked process offset approximately 0.18, the stage process sequence consists of a process amount before detachment approximately 0.09 and a process amount after detachment approximately 0.09, and the recording time is the start time of the analysis segment; the monitoring process record corresponding to a monitoring candidate segment with a decreasing missing marker is as follows: turning point position approximately 0.55, decreasing missing marker, no detachment marker, the stage process sequence consists of a single-segment process amount approximately 0.45, and the recording time is the start time of its analysis segment.
[0078] Through the above technical solution, this embodiment converts the turning points, descents, and detachments of each monitoring candidate segment into process positions and selects one of two cases to form a monitoring process record. Then, it arranges the segments by their starting time to form a lumbar and pelvic change monitoring sequence, forming a continuous monitoring result that is organized by time and has a complete internal structure. The key difference is that separate recording fields are set for the two cases of determining the descent position and the descent absence, so that the monitoring sequence retains the descent absence information while maintaining strict consistency in the field caliber. The resulting lumbar and pelvic change monitoring sequence is the continuous monitoring result output by this method, which can be used by clinical or rehabilitation personnel to interpret the changing trend of the lumbar condition in conjunction with symptoms.
[0079] See Figure 2 As shown, this solution proposes a sitting lumbar change monitoring system based on lumbar flexion distribution transition, used to implement the aforementioned sitting lumbar change monitoring method based on lumbar flexion distribution transition, including:
[0080] The multi-source acquisition module is used to acquire lumbar electromyography signals, pelvic tilt sequence, and lumbar curvature sequence of the subject in a continuous sitting posture, and align them by time.
[0081] The flexion quantification module is used to extract the posterior tilt change of the pelvic tilt sequence and the lordosis reduction of the lumbar curvature sequence, which are normalized into posterior tilt component and flattening component, respectively. The two components are added together at each time step to obtain the total change sequence. The interval where the total change sequence is continuously positive and the cumulative value reaches the preset flexion threshold is determined as the candidate flexion segment. The lumbar flexion amount is obtained by accumulating the total change sequence from the starting point of the segment.
[0082] The buckling distribution transition module is used to calculate the difference in the proportion of the backward tilt component and the flattening component within a preset window in the candidate buckling segment, forming a buckling distribution difference sequence. When the buckling distribution difference sequence changes from continuous positive values to continuous negative values or reverses, and the absolute values of the buckling distribution differences on both sides of the transition reach the preset distribution difference, the first positive-negative boundary position is determined as the waist-pelvis distribution transition position, and the corresponding candidate buckling segment is determined as the analysis segment.
[0083] The electromyography response module is used to rectify and extract the envelope of the lumbar electromyography signal within the analysis segment to obtain the lumbar muscle activity amplitude sequence. Within the preset search range before and after the lumbar-pelvic junction transition position, the average amplitude of the lumbar muscle activity amplitude sequence within the preset time period at the beginning of the analysis segment is used as the benchmark. The position where the sequence first enters a state that is continuously lower than the benchmark and the decrease reaches the preset decrease is determined as the lumbar muscle activity reduction position. If it is not identified, a reduction missing marker is generated.
[0084] The process offset module is used to determine the location of reduced lumbar muscle activity by dividing the lumbar flexion amount corresponding to the location and the lumbar-pelvic assignment turning point by the lumbar flexion amount at the end of the analysis segment, and subtracting the latter ratio from the former ratio to obtain the signed process offset.
[0085] The monitoring sequence module is used to identify analysis segments with signed process offsets exceeding the preset offset range or with reduced missing markers as monitoring candidate segments. Each monitoring candidate segment is associated with its waist-pelvis allocation turning position, signed process offset, or reduced missing marker in the order of appearance to obtain the waist-pelvis change monitoring sequence.
[0086] In another embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above embodiments.
[0087] In one embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps described above.
[0088] In one embodiment, a computer program product or computer program is provided, the computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and executes the computer instructions, causing the computer device to perform the steps described above.
[0089] The technical scope of this invention is not limited to the content described above. Those skilled in the art can make various modifications and variations to the above embodiments without departing from the technical concept of this invention, and all such modifications and variations should fall within the protection scope of this invention.
Claims
1. A method for monitoring lumbar changes in seated posture based on lumbar-pelvic flexion distribution transition, characterized in that, The method includes: Acquire lumbar electromyography (EMG) signals, pelvic tilt sequence, and lumbar curvature sequence of subjects in a continuous sitting posture, and align them by time. The posterior tilt change of the pelvic tilt sequence and the lordosis reduction of the lumbar curvature sequence are extracted and normalized into posterior tilt component and flattening component, respectively. The two components are added together at each time step to obtain the total change sequence. The interval where the total change sequence is continuously positive and the cumulative value reaches the preset flexion threshold is determined as the candidate flexion segment. The lumbar pelvic flexion amount is obtained by accumulating the total change sequence from the starting point of the segment. Within the candidate buckling segment, the difference in the proportion of the backward tilt component and the flattening component is calculated according to a preset window to form a buckling distribution difference sequence. When the buckling distribution difference sequence changes from continuous positive values to continuous negative values or reverses, and the absolute values of the buckling distribution difference on both sides of the change reach the preset distribution difference, the first positive-negative boundary position is determined as the waist-pelvis distribution turning point, and the corresponding candidate buckling segment is determined as the analysis segment. The lumbar electromyography signal within the analysis segment is rectified and the envelope is extracted to obtain the lumbar muscle activity amplitude sequence. Within the preset search range before and after the lumbar-pelvic junction transition position, the average amplitude of the lumbar muscle activity amplitude sequence within the preset time period at the beginning of the analysis segment is used as the benchmark. The position where the sequence first enters a state that is continuously lower than the benchmark and the decrease reaches the preset decrease is determined as the lumbar muscle activity reduction position. If it is not identified, a reduction missing marker is generated. When determining the location of reduced lumbar muscle activity, the lumbar flexion amount corresponding to the location and the lumbar-pelvic distribution turning point is divided by the lumbar flexion amount at the end of the analysis segment, and the signed process offset is obtained by subtracting the second ratio from the first ratio. Analysis segments with signed process offsets exceeding the preset offset range or with reduced missing markers are identified as monitoring candidate segments. Each monitoring candidate segment is associated with its lumbar-pelvic allocation turning point, signed process offset, or reduced missing marker in the order of appearance to obtain the lumbar-pelvic change monitoring sequence.
2. The method according to claim 1, characterized in that, Within the candidate buckling segment, the difference in the proportions of the backward tilt component and the flattening component is calculated according to a preset window to form a buckling distribution difference sequence, specifically including: Starting from the beginning of the candidate buckling segment, the lumbar-pelvic buckling amount is accumulated. Whenever the accumulated increase reaches a preset buckling increment, the corresponding time point is recorded as the process boundary, and the time period between two adjacent process boundaries is used to form a process unit. The backward tilt component and the flattening component in each process unit are accumulated respectively to obtain the unit backward tilt amount and the unit flattening amount. The continuous process units are combined into a preset window according to the preset number of window units, and one process unit is moved backward each time. The unit backtilt and unit flattening within each preset window are accumulated to obtain the window backtilt and window flattening. The two are added together to obtain the window total measurement. The difference between the proportion of the window backtilt and window flattening to the window total measurement is calculated, and a buckling distribution difference sequence is formed according to the preset window movement order.
3. The method according to claim 2, characterized in that, The first positive-negative boundary position is determined as the turning point of the lumbar-pelvic distribution. The corresponding candidate flexion segment is determined before the analysis segment. The process of observation after the turning point is also included, specifically: Locate the process unit where the lumbar-pelvic distribution transition point is located. Starting from the next process unit after this process unit, count the process units formed before the end of the candidate buckling segment in sequence along the direction of increasing lumbar-pelvic buckling amount to obtain the number of units after the transition. If the number of units after the turning point reaches the preset number of observation units, the corresponding candidate buckling segment will be determined as the analysis segment; otherwise, the corresponding candidate buckling segment will not be determined as the analysis segment.
4. The method according to claim 1, characterized in that, The location where the lumbar muscle activity amplitude sequence first enters a state where it is consistently below the baseline and the decrease reaches a preset decrease value is defined as the location of decreased lumbar muscle activity. If it is not identified, a decrease loss marker is generated, specifically including: Read the lumbar muscle activity amplitude sequence in chronological order within the preset search range. Whenever the lumbar muscle activity amplitude changes from not reaching the preset reduction to being below the baseline and the reduction reaches the preset reduction, record the corresponding position as the candidate starting point of low activity. Simultaneously read the lumbar muscle activity amplitude sequence and lumbar pelvic flexion from the low activity candidate starting point, and calculate the increase in the current lumbar pelvic flexion relative to the lumbar pelvic flexion corresponding to the low activity candidate starting point; When the amplitude of lumbar muscle activity is consistently lower than the baseline and the decrease reaches the preset decrease, and the increase in lumbar pelvic flexion reaches the preset decrease in continuous flexion, the candidate starting point of low activity is determined as the position of lumbar muscle activity reduction. When the amplitude of lumbar muscle activity exits the reduced state before the increase in lumbar pelvic flexion reaches the preset reduction in continuous flexion, the current low activity candidate starting point is canceled and the search for the next low activity candidate starting point continues. If the position of lumbar muscle activity reduction is still not determined after traversing the preset search range, a reduction missing marker is generated.
5. The method according to claim 2, characterized in that, After identifying the first positive-negative boundary location as the lumbar-pelvic distribution turning point and determining the corresponding candidate flexion segment as the analysis segment, the analysis also includes lumbar spine participation status identification, specifically including: The backward tilt component and the flattening component within the preset window are accumulated separately to obtain the window backward tilt amount and the window flattening amount. The two are added together to obtain the window total amount. When both the window tilt and the window flattening are positive, the corresponding window is marked as the joint propulsion state; when both the window summation and the window tilt are positive and the window flattening is not greater than zero, the corresponding window is marked as the disengagement propulsion state; the remaining windows are marked as transition states, and the state markings are arranged in window order to obtain the lumbar spine participation state sequence. The preset window corresponding to the lumbar-pelvic distribution turning point is located. Starting from the next preset window, the lumbar spine participation state sequence is read in chronological order. When the number of consecutive common propulsion states reaches the preset number of common propulsion windows, and the number of consecutive disengagement propulsion states that occur afterward reaches the preset number of disengagement propulsion windows, the starting position of the window corresponding to the first disengagement propulsion state is determined as the lumbar spine disengagement position. If the location of lumbar vertebral detachment is still not determined when the analysis segment is read to the end, a non-detachment marker is generated.
6. The method according to claim 5, characterized in that, After identifying segments with signed process offsets exceeding the preset offset range or with reduced missing markers as monitoring candidate segments, the process also includes stage process construction, specifically including: For the candidate monitoring segments where the lumbar muscle activity decreases, the order of the lumbar muscle activity decrease position and the lumbar-pelvic distribution transition position is compared. When the lumbar muscle activity decrease position is earlier than the lumbar-pelvic distribution transition position, the difference between the corresponding lumbar-pelvic flexion amounts is calculated. The lumbar-pelvic flexion amount at the end of the analysis segment is normalized to obtain the preceding process amount. The preceding process amount is used as a single element to form the stage process sequence. For monitoring candidate segments that determine the location of decreased lumbar muscle activity, if the location of decreased lumbar muscle activity is not earlier than the lumbopelvic junction, the location of decreased lumbar muscle activity is determined as the process endpoint; if there is a missing marker for decreased activity, the end of the analysis segment is determined as the process endpoint. When there is no detachment marker, or the lumbar detachment position is not located between the lumbar-pelvic distribution turning point and the process endpoint, calculate the difference in lumbar-pelvic flexion from the lumbar-pelvic distribution turning point to the process endpoint, and normalize the lumbar-pelvic flexion at the end of the analysis segment to obtain the single-segment process quantity. Use the single-segment process quantity as a single element to form the stage process sequence. When the lumbar vertebral dislocation is located between the lumbar-pelvic distribution turning point and the end of the process, the difference in lumbar-pelvic flexion between the lumbar-pelvic distribution turning point and the lumbar vertebral dislocation point and between the lumbar vertebral dislocation point and the end of the process is calculated respectively. The difference is then normalized with the lumbar-pelvic flexion at the end of the analysis segment to obtain the pre-dislocation process amount and the post-dislocation process amount. The pre-dislocation process amount and the post-dislocation process amount are arranged in chronological order to form a stage process sequence.
7. The method according to claim 6, characterized in that, Each candidate segment is correlated with its lumbar-pelvic allocation transition position, signed process offset, or reduced missing marker according to the order of appearance, to obtain the lumbar-pelvic change monitoring sequence, which specifically includes: The lumbar pelvic flexion amount corresponding to the lumbar pelvic flexion position in each monitoring candidate segment is extracted separately. The lumbar pelvic flexion amount at the end of the analysis segment is normalized to obtain the position of the flexion process. For the candidate monitoring segments where the lumbar muscle activity is reduced, the lumbar pelvic flexion corresponding to the reduced lumbar muscle activity position is normalized to the lumbar pelvic flexion at the end of the analysis segment to obtain the position of the reduction process. When determining the location of lumbar vertebral detachment, the lumbar pelvic flexion corresponding to the lumbar detachment location is divided by the lumbar pelvic flexion at the end of the analysis segment to obtain the location of the detachment process. For candidate monitoring segments where the lumbar muscle activity is reduced, the monitoring process record is formed by associating the turning process position, the reduction process position, the departure process position or the non-departure mark, the marked process offset, and the stage process sequence. For monitoring candidate segments with reduced missing markers, the turning point process position, reduced missing marker, exit process position or not exit marker, and stage process sequence are associated to form a monitoring process record; The start time of the corresponding monitoring candidate segment is used as the recording time of each monitoring process. The records of each monitoring process are arranged in chronological order of recording time to obtain the waist and pelvis change monitoring sequence.
8. A sitting posture lumbar change monitoring system based on lumbar pelvic flexion distribution transition, characterized in that, A method for monitoring changes in lumbar posture based on lumbar flexion distribution transition as described in any one of claims 1-7, comprising: The multi-source acquisition module is used to acquire lumbar electromyography signals, pelvic tilt sequence, and lumbar curvature sequence of the subject in a continuous sitting posture, and align them by time. The flexion quantification module is used to extract the posterior tilt change of the pelvic tilt sequence and the lordosis reduction of the lumbar curvature sequence, which are normalized into posterior tilt component and flattening component, respectively. The two components are added together at each time step to obtain the total change sequence. The interval where the total change sequence is continuously positive and the cumulative value reaches the preset flexion threshold is determined as the candidate flexion segment. The lumbar flexion amount is obtained by accumulating the total change sequence from the starting point of the segment. The buckling distribution transition module is used to calculate the difference in the proportion of the backward tilt component and the flattening component within a preset window in the candidate buckling segment, forming a buckling distribution difference sequence. When the buckling distribution difference sequence changes from continuous positive values to continuous negative values or reverses, and the absolute values of the buckling distribution differences on both sides of the transition reach the preset distribution difference, the first positive-negative boundary position is determined as the waist-pelvis distribution transition position, and the corresponding candidate buckling segment is determined as the analysis segment. The electromyography response module is used to rectify and extract the envelope of the lumbar electromyography signal within the analysis segment to obtain the lumbar muscle activity amplitude sequence. Within the preset search range before and after the lumbar-pelvic junction transition position, the average amplitude of the lumbar muscle activity amplitude sequence within the preset time period at the beginning of the analysis segment is used as the benchmark. The position where the sequence first enters a state that is continuously lower than the benchmark and the decrease reaches the preset decrease is determined as the lumbar muscle activity reduction position. If it is not identified, a reduction missing marker is generated. The process offset module is used to determine the location of reduced lumbar muscle activity by dividing the lumbar flexion amount corresponding to the location and the lumbar-pelvic assignment turning point by the lumbar flexion amount at the end of the analysis segment, and subtracting the latter ratio from the former ratio to obtain the signed process offset. The monitoring sequence module is used to identify analysis segments with signed process offsets exceeding the preset offset range or with reduced missing markers as monitoring candidate segments. Each monitoring candidate segment is associated with its waist-pelvis allocation turning position, signed process offset, or reduced missing marker in the order of appearance to obtain the waist-pelvis change monitoring sequence.
9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the method described in any one of claims 1-7.
10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1-7.
