Infrared imaging human posture recognition method and device

Abstract

The present invention relates to a method for human posture recognition using infrared imaging, comprising the following steps: S1: receiving images captured by a thermal imaging sensor, forming a human body thermal distribution image, and obtaining the highest temperature points in the human body thermal distribution image; S2: determining the human body posture based on the position of the highest temperature points in the human body thermal distribution image and the distribution pattern of the human body thermal distribution image. The infrared imaging human posture recognition method provided by the present invention achieves low-cost human posture recognition and is beneficial to the large-scale promotion of human body monitoring products.

Description

Technical Field

[0001] The present invention relates to the technical field of human posture recognition, and in particular to an infrared imaging human posture recognition method and device. Background Art

[0002] With my country's rapid population and economic growth, advances in science and technology, improved healthcare, and rising living standards, life expectancy has significantly increased. Beginning in 2010, the baby boomers born after the founding of the People's Republic of China gradually entered old age. By 2014, my country's population aged 60 and over reached 212 million, making it the world's largest. As of May 11, 2021, China's population aged 60 and over stood at 264.02 million, further deepening the aging of its population. With the aging of my country's population, providing timely care and assistance to the elderly in their daily lives has become a social issue that needs to be addressed and resolved. As the elderly age, the functions of various organs in the body deteriorate, and their body balance, coordination and stability are greatly reduced. The bones are fragile due to osteoporosis, and they are also affected and restricted by factors such as vision and reaction ability. They are very prone to falls and injuries. In addition, judging from the current elderly care situation in our country, the vast majority of elderly people rely on their children for their daily lives, and home care is the main form of elderly care. However, in real life, the children of the elderly often have a lot of things to do and cannot put all their energy on taking care of the elderly. The elderly are often in a state of living alone. When a fall occurs, it is difficult for the children to be informed of the accident in the first time and to rescue the elderly in time. This is also a problem that greatly troubles the current home-based elderly care life and has a great impact on the normal life of the elderly.

[0003] Infrared thermal imaging technology has rapidly advanced, and the cost of thermal imaging sensors has plummeted. This has enabled the technology to expand from high-end applications to more mass-market civilian applications. This shift has led to the proliferation of low-cost, low-resolution thermal imaging sensors on the market. Consequently, the question of how to effectively utilize these low-cost, low-resolution thermal imaging sensors to meet the needs of human posture monitoring has become a pressing issue. Summary of the Invention

[0004] In response to market demand, the present invention provides an infrared imaging human posture recognition method and device that realizes low-cost human posture recognition and is beneficial to the large-scale promotion of human monitoring products.

[0005] In order to achieve the purpose of the present invention, the present invention provides a method for infrared imaging human posture recognition, comprising the following steps:

[0006] S1: Receive images collected by the thermal imaging sensor, form a human body thermal distribution image, and obtain the high temperature point in the human body thermal distribution image;

[0007] S2: judging the human body posture according to the position of the high temperature point in the human body thermal distribution image and the distribution form of the human body thermal distribution image.

[0008] Preferably, in step S2: when the human body thermal distribution image is vertically distributed and the high temperature point is located at the upper part of the human body thermal distribution image, it is determined that the human body is in an upright state;

[0009] When the human body thermal distribution image is distributed in a horizontally long manner and the high temperature point is located at the left end or the right end of the human body thermal distribution image, it is determined that the human body is in a lying state.

[0010] Preferably, step S2 further includes the following sub-steps:

[0011] S21: determining the vertical distance between the upper and lower endpoints of the human body thermal distribution image as its height value, and determining the horizontal distance between the left and right endpoints of the human body thermal distribution image as its width value;

[0012] S22: Calculating the ratio of the height value to the width value of the human body thermal distribution image;

[0013] S23: judging whether the human body thermal distribution image is vertically distributed or horizontally distributed according to the ratio of the height value to the width value of the human body thermal distribution image.

[0014] Preferably, step S2 further includes the following sub-steps:

[0015] S221: Acquire one or more high temperature points from the human body thermal distribution image;

[0016] S222: Determine the distribution form of the human body thermal distribution image according to the relative position relationship between the high temperature point and the low temperature area.

[0017] Preferably, step S1 further includes the following sub-steps:

[0018] S11: The thermal imaging sensor collects the spatial temperature distribution, and smoothes the multi-frame data to eliminate sudden temperature fluctuations and form a stable temperature distribution image.

[0019] S12: using a differential algorithm on the multi-frame steady temperature distribution images to eliminate fixed constant temperature objects and form a protrusion temperature distribution image;

[0020] S13: Calculating the protrusion temperature distribution image surface, and eliminating the protrusion temperature distribution images below a preset value in the image surface;

[0021] S14: Scan the temperature distribution image of the protrusion to find the highest temperature point within the preset temperature range, which is the temperature high point. Then, with the temperature high point as the center, gradually lower the temperature value within the preset temperature range to form a thermal distribution contour line. According to the thermal distribution contour line, a thermal distribution image of the human body is obtained.

[0022] Preferably, the step S14 further includes the following sub-steps:

[0023] S141: Find the high temperature point, take the high temperature point as the center point, obtain the temperature values ​​of multiple adjacent points around the high temperature point, and if the temperature value of the adjacent point is within a preset temperature range, mark the adjacent point as a human body location point, otherwise mark it as a non-human body location point;

[0024] S142: Taking the marked human body position point as the center point, obtaining temperature values ​​of a plurality of unmarked adjacent points surrounding the center point; if the temperature value of the adjacent point is within a preset temperature range, marking the adjacent point as a human body position point; otherwise, marking the adjacent point as a non-human body position point;

[0025] S143: repeating step S142 until the temperature values ​​of all unmarked adjacent points surrounding the center point are not within the preset temperature range; or repeating step S142 until all data points of the protrusion temperature distribution image are traversed;

[0026] S144: Connect all continuous human body position points including the high temperature point to obtain a human body thermal distribution image.

[0027] Preferably, the present invention further provides an infrared imaging human posture recognition device, comprising:

[0028] Image analysis module: receives images collected by the thermal imaging sensor, forms a human body thermal distribution image, and obtains the highest temperature point in the human body thermal distribution image;

[0029] Posture judgment module: judges the posture of the human body according to the high temperature point obtained by the image processing module, the position of the high temperature point in the human body thermal distribution image and the distribution form of the thermal distribution image.

[0030] Preferably, in the posture judgment module:

[0031] When the human body thermal distribution image is vertically distributed and the high temperature point is located at the upper part of the human body thermal distribution image, it is determined that the human body is in an upright state;

[0032] When the human body thermal distribution image is distributed in a horizontally long manner and the high temperature point is located at the left end or the right end of the human body thermal distribution image, it is determined that the human body is in a lying state.

[0033] Preferably, the posture judgment module further includes a first posture judgment submodule 1, a first posture judgment submodule 2 and a first posture judgment submodule 3:

[0034] The first posture judgment submodule 1 determines the vertical distance between the upper and lower endpoints of the human body thermal distribution image as its height value, and determines the horizontal distance between the left and right endpoints of the human body thermal distribution image as its width value;

[0035] The first posture judgment submodule 2 calculates the ratio of the height value to the width value of the human body thermal distribution image;

[0036] The first posture judgment submodule three: judges whether the human body thermal distribution image is vertically distributed or horizontally distributed according to the ratio of the height value to the width value of the human body thermal distribution image.

[0037] Preferably, the posture judgment module further includes a second posture judgment submodule 1 and a second posture judgment submodule 2:

[0038] Second posture judgment submodule 1: obtaining one or more high temperature points from the human body thermal distribution image;

[0039] Second posture judgment submodule 2: determining the distribution form of the human body thermal distribution image according to the relative position relationship between the high temperature point and the low temperature area.

[0040] The beneficial effects of the present invention are as follows: the infrared imaging human posture recognition method and device provided by the present invention realize low-cost human posture recognition, which is beneficial to the large-scale promotion of human body monitoring products. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The above and other purposes, features and advantages of the present invention will become more apparent by describing in more detail the preferred embodiments of the present invention shown in the accompanying drawings. The same reference numerals indicate the same parts throughout the accompanying drawings, and the drawings are not intentionally scaled to actual size, but rather are intended to illustrate the subject matter of the present invention.

[0042] Figure 1 A specific flow chart of an infrared imaging human posture recognition method and device provided in an embodiment of the present invention;

[0043] Figure 2 A specific flow chart of step S1 in an infrared imaging human posture recognition method and device provided in an embodiment of the present invention;

[0044] Figure 3 A schematic diagram of a human body thermal contour distribution map provided by an embodiment of the present invention;

[0045] Figure 4 A schematic diagram of a first real-time human posture recognition provided by an embodiment of the present invention;

[0046] Figure 5 A schematic diagram of a second real-time human posture recognition method provided by an embodiment of the present invention; DETAILED DESCRIPTION

[0047] To facilitate understanding of the present invention, the present invention will be described more fully below with reference to the accompanying drawings.

[0048] It should be noted that when an element is considered to be "connected" to another element, it may be directly connected to the other element and integrated therewith, or there may be an intermediate element. The terms "mounted", "one end", "the other end" and similar expressions used herein are for illustrative purposes only.

[0049] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one skilled in the art to which this document pertains. The terms used in this specification are for the purpose of describing specific embodiments only and are not intended to limit the present invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0050] Please refer to Figure 1-5 The embodiment of the present invention provides a low-cost infrared imaging human posture recognition method, comprising the following steps:

[0051] S1: Receive images collected by the thermal imaging sensor, form a human body thermal distribution image, and obtain the high temperature point in the human body thermal distribution image;

[0052] S2: judging the human body posture according to the position of the high temperature point in the human body thermal distribution image and the distribution form of the human body thermal distribution image.

[0053] Please refer to Figure 1-5The infrared imaging human posture recognition method provided by the present invention has the following specific recognition method: first, data is collected from the corresponding area (bathroom, bedroom, kitchen, etc.) through a thermal imaging sensor, and then the collected data is arranged in a corresponding order to form a corresponding image, and the spatial temperature distribution in the area can be known through the collected image, and then the sudden temperature fluctuation is eliminated through smoothing operation of multiple frames of data to form a stable temperature distribution image, and then the difference between two frames of stable temperature distribution images is used to eliminate fixed constant temperature objects to form a protrusion temperature distribution image, and then multiple image surfaces in the protrusion temperature distribution image are calculated, and the protrusion temperature distribution image with a temperature value lower than a preset value (a temperature value that does not conform to the human body temperature) in the image surface is eliminated, and then the protrusion temperature distribution image is scanned to find the highest temperature point (high temperature point) that conforms to the human body temperature, and then with the high temperature point as the center, multiple adjacent points adjacent to the high temperature point are obtained around the high temperature point, and multiple adjacent points are obtained for multiple The adjacent points are judged. When the adjacent point is within the preset temperature range (the temperature range that matches the human body temperature), the adjacent point is marked and marked as a human body position point (the position of a part on the human body). When the adjacent point is not within the preset temperature range, the adjacent point is marked as a non-human body position point. Then, with the marked human body position point as the center, the temperature values ​​of multiple unmarked adjacent points surrounding the marked human body position point are obtained, and a judgment is made again to determine whether it is a human body position point. The specific judgment process is the same as the above judgment process. When the temperature values ​​of all the unmarked adjacent points around the marked human body position point are not within the preset temperature range, the temperature value acquisition process ends, and then all continuous human body position points containing the high temperature point are connected to form a human body thermal distribution image; or when traversing the temperature data points in the protrusion temperature distribution image, all continuous human body position points containing the high temperature point are connected to form a human body thermal distribution image. (The high temperature value is specifically the temperature value of the upper half of the human body)

[0054] After the thermal distribution contours are drawn, the specific location of the high temperature point in the thermal distribution contours is determined, and the human body posture is judged based on the human body thermal distribution image. The specific judgment is as follows:

[0055] When the human body thermal distribution image is vertically distributed and the high temperature point is located at the upper part of the human body thermal distribution image, it is determined that the human body is in an upright state;

[0056] When the body's thermal distribution image is horizontally long and the high temperature point is located at the left or right end of the body's thermal distribution image, it is determined that the body is in a lying state;

[0057] At the same time, the present invention can also judge the human body posture according to the specific ratio of the length and width of the human body thermal distribution image, and the specific judgment is as follows:

[0058] The vertical distance between the upper and lower endpoints of the human body thermal distribution image is determined as its height value, and the horizontal distance between the left and right endpoints of the human body thermal distribution image is determined as its width value. The ratio of the height value to the width value of the human body thermal distribution image is calculated. Based on the ratio of the height value to the width value of the human body thermal distribution image, it is judged whether the human body thermal distribution image is vertically distributed (standing upright, low posture, bent over, etc.) or horizontally distributed (lying down). The present invention can also judge the human body posture based on the relative position relationship between the high temperature point and the low temperature point of the human body thermal distribution image.

[0059] The beneficial effects of the present invention are as follows: the highest temperature point close to human body temperature is found according to the temperature distribution image of the protrusion, and then the temperature threshold is appropriately and gradually lowered with the high temperature point as the center to form a thermal distribution contour line. The present invention can judge the human posture through the specific shape of the thermal distribution contour line, the relative position between the high temperature point and the low temperature point, and the ratio of the length and width in the thermal distribution image of the human body, thereby realizing low-cost human posture recognition, which is beneficial to the large-scale promotion of human body monitoring products.

[0060] Please refer to Figure 1 In a preferred embodiment, in step S2:

[0061] When the human body thermal distribution image is vertically distributed and the high temperature point is located at the upper part of the human body thermal distribution image, it is determined that the human body is in an upright state;

[0062] When the human body thermal distribution image is horizontally long and the high temperature point is located at the left end or the right end of the human body thermal distribution image, it is determined that the human body is in a lying state.

[0063] Please refer to Figure 1 In a preferred embodiment, step S2 further includes the following sub-steps:

[0064] S21: determining the vertical distance between the upper and lower endpoints of the human body thermal distribution image as its height value, and determining the horizontal distance between the left and right endpoints of the human body thermal distribution image as its width value;

[0065] S22: Calculating the ratio of the height value to the width value of the human body thermal distribution image;

[0066] S23: judging whether the human body thermal distribution image is vertically distributed or horizontally distributed according to the ratio of the height value to the width value of the human body thermal distribution image.

[0067] The judgment is mainly based on the ratio of the specific length and width of the formed image, that is, the contour formed by the connection of the contour lines.

[0068] For example, if the height is h and the width is d, and the ratio of the height to the width is t, then t = h / d:

[0069] When t > 1.5, it is determined that the human body is in a standing state;

[0070] When 0.75 < t < 1.5, it is determined that the human body is in a low posture state;

[0071] When t < 0.75, it is determined that the human body is in a lying state; (This embodiment is only for illustration and does not represent specific values. At the same time, the ratio of the length to the width of the image can be adjusted, mainly to improve the accuracy of human posture judgment)

[0072] Please refer to Figure 1-3 , in a further preferred embodiment, step S2 further includes the following sub-steps:

[0073] S221: Obtain one or more temperature high points from the human body thermal distribution image;

[0074] S222: Determine the distribution form of the human body thermal distribution image according to the relative position relationship between the temperature high points and the low temperature region.

[0075] First, determine the temperature high points close to the human body temperature, which can be one or more. Then, obtain the low temperature region that conforms to the human body from the image (the region formed by multiple temperature low points. Assuming the temperature high points are 37, 37.5, 37.4, the temperature data of the low temperature region are 35.5, 36.5, 36, 6.7). Then, judge the human body posture according to the specific positions of the temperature high points and the low temperature region in the human body thermal distribution image. This judgment method can be more accurate.

[0076] The specific judgment steps are as follows:

[0077] When the temperature high points are in the upper part of the human body thermal distribution image and the low temperature region is in the lower part of the human body thermal distribution image, it can be determined that the human body is in a standing state;

[0078] When the temperature high points are in the lower part of the human body thermal distribution image and the low temperature region is in the lower part of the human body thermal distribution image, it can be determined that the human body is in a lying state;

[0079] When the temperature high points are in the middle part of the human body thermal distribution image and the low temperature region is in the lower part of the human body thermal distribution image, it can be determined that the human body is in a low posture state. The low posture state is a state such as bowing the head or bending the waist.

[0080] Please refer to Figure 1-3 , in a further preferred embodiment, step S1 further includes the following sub-steps:

[0081] S11: Collect the spatial temperature distribution through a thermal imaging sensor, and eliminate the sudden temperature fluctuations through the smoothing operation of multiple frames of data to form a stable temperature distribution image;

[0082] S12: using a differential algorithm on the multi-frame steady temperature distribution images to eliminate fixed constant temperature objects and form a protrusion temperature distribution image;

[0083] S13: Calculating the protrusion temperature distribution image surface, and eliminating the protrusion temperature distribution images below a preset value in the image surface;

[0084] The preset value is mainly a set value of the overall outline of the human body and can be adjusted. The image surface below the preset value is mainly the image outline of the non-human body and will be eliminated. The protrusion temperature distribution image surface is mainly the image outline of the human body shape.

[0085] S14: Scan the temperature distribution image of the protrusion to find the highest temperature point within the preset temperature range, which is the high temperature point. Then, with the high temperature point as the center, gradually reduce the temperature value within the preset temperature range to form a thermal distribution contour line. According to the thermal distribution contour line, a thermal distribution image of the human body is obtained.

[0086] The specific algorithm name is sliding average:

[0087] That is, the variable v at time t is recorded as vt, and θt is the value of the variable v at time t, that is, vt = θt. After using the sliding average model, the update formula of vt is as follows:

[0088] vt=β·vt-1+(1-β)·θt (1)

[0089] In the above formula, β∈[0,1]

[0090] To eliminate the initial bias, the estimate of the mean is corrected by dividing vt by (1-βt).

[0091] The update formulas for vt and v_biasedt are as follows:

[0092] vt=β·vt-1+(1-β)·θt

[0093] v_biasedt=vt1-βt (2)

[0094] By using formula (2) to calculate all temperature points, the mutation data can be eliminated. When β = 0.9, it is roughly equal to the average of the past 10 θ values; if β = 0.99, it is roughly equal to the average of the past 100 θ values.

[0095] The image formed by the smoothed temperature data can be considered a 32x32 matrix, where the data is the temperature. For fixed objects, the temperature does not change significantly between two frames of data. Through matrix subtraction, the unchanged data will cancel each other out, and the changed data will stand out. By setting an appropriate threshold, the data that is too small is reset to zero, thus clearing the background. The purpose of clearing the background is to remove interference from static objects such as tables, chairs, and benches. After clearing the background, the data is used to form a pseudo-color image using a rainbow image. In the specific object temperature distribution map, low-temperature areas will appear blue and high-temperature areas will appear red, highlighting the human body outline and enabling better recognition of human posture.

[0096] Please refer to Figure 3 In a preferred embodiment, step S14 further includes the following sub-steps:

[0097] S141: Find the high temperature point, take the high temperature point as the center point, obtain the temperature values ​​of multiple adjacent points around the high temperature point, and if the temperature value of the adjacent point is within a preset temperature range, mark the adjacent point as a human body position point, otherwise mark it as a non-human body position point;

[0098] S142: Taking the marked human body location point as the center point, obtaining the temperature values ​​of multiple unmarked adjacent points around the center point. If the temperature value of the adjacent point is within a preset temperature range, the adjacent point is marked as a human body location point; otherwise, it is marked as a non-human body location point.

[0099] S143: Repeat step S142 until the temperature values ​​of all unmarked adjacent points around the center point are not within the preset temperature range; or repeat step S142 until all data points of the protrusion temperature distribution image are traversed;

[0100] S144: Connect all continuous human body position points including the high temperature point to obtain a human body thermal distribution image.

[0101] The specific embodiments are:

[0102] First, extract the high temperature point from the data that meets the human body temperature value (assuming that the human body temperature value is 35-37° and the high temperature point is 36.9°). Then, take the high temperature point as the center and obtain the temperature value adjacent to the high temperature point. When the temperature value of the adjacent point is within the preset temperature range (assuming the preset temperature range is 35-37°), the adjacent point that meets the conditions is marked as the human body position point (the temperature point that meets the human body temperature), and the adjacent point that does not meet the conditions is marked as the non-human body position point (the temperature value of the constant temperature object or other non-human body). Then, take the human body position point as the center and judge the temperature value of the adjacent points around the human body position point. When the adjacent point is within the preset temperature range (assuming the preset temperature range is 35-37°), the adjacent point that meets the conditions is marked as the human body position point (the temperature point that meets the human body temperature). If the temperature value of a point is within a preset temperature range, the adjacent point is determined to be a human body position point, and then the judgment of the unmarked adjacent points is repeated. When none of the unmarked adjacent points meet the human body temperature value, a judgment can be made based on the image and the adjacent points, and the points that do not meet the human body temperature value are determined to be non-human body position points. Then, the adjacent points that meet the human body temperature value are connected with the high temperature point (specifically, for continuous human body position points, when determining adjacent points, there may be multiple image contours, such as the temperature image contours of two human bodies, or one person and multiple non-human bodies. The human body image is mainly determined by the contour formed by connecting multiple continuous adjacent points and the high temperature point);

[0103] When there are multiple image contours, there will be two images when judging adjacent points, and the two images are discontinuous. In this case, the main judgment condition (human body temperature value) is based on the continuous adjacent points formed with the high temperature point, and the images of adjacent points that do not meet the human body temperature value are eliminated accordingly.

[0104] Alternatively, the judgment of unmarked adjacent points is repeated, and each temperature point in the image is judged, the temperature points within the preset temperature range are marked, the temperature points outside the preset temperature range are excluded, and the temperature points within the preset temperature range are connected to form the human body image contour;

[0105] The above two methods are used to judge the human body contour, which is more convenient and simpler in recognizing human body images.

[0106] (When taking adjacent point values, there may be some errors, such as the influence of ambient temperature, and the specific values ​​can be pre-set)

[0107] For example, when the selected adjacent point is 22-25°, the temperature value of the selected adjacent point does not conform to the temperature value of the human body, so it can be excluded.

[0108] The specific steps are as follows: First, find the maximum value close to human body temperature in the data center, which can be completed by simply traversing 1024 data. The specific value acquisition process is as follows:

[0109]

[0110]

[0111] Determine the maximum value at the coordinates Max_x and Max_y, denoted as x and y.

[0112] Then it is considered that the range of t∈[maximum value-n, maximum value] is the reasonable temperature of the human body,

[0113] Compare the eight points (X+1, Y)(X+1, Y+1)(X-1, Y)(X-1, Y+1)(X, Y-1)(X, Y+1)(X-1, Y-1)(X+1, Y-1) around x, y to see if they meet t. If so, mark this point, and then check the eight points around this point with this point as the center until all points are marked or there are no points that meet t. The set of all marked points is a temperature contour plane. Different contour planes can be made for different n, that is, starting from a point in the area, by visiting the 8 adjacent points of the known point, all data points in the area are traversed under the premise of meeting t∈[maximum value-n, maximum value].

[0114] Please refer to Figure 1-5 In a further preferred embodiment, the present invention further provides an infrared imaging human posture recognition device, comprising:

[0115] Image analysis module: receives images collected by the thermal imaging sensor, forms a human body thermal distribution image, and obtains the highest temperature point in the human body thermal distribution image;

[0116] Posture judgment module: Based on the high temperature point obtained by the image processing module, the human body posture is judged according to the position of the high temperature point in the human body thermal distribution image and the distribution form of the thermal distribution image.

[0117] Please refer to Figure 1-5 In a preferred embodiment, in the posture determination module:

[0118] When the human body thermal distribution image is vertically distributed and the high temperature point is located at the upper part of the human body thermal distribution image, it is determined that the human body is in an upright state;

[0119] When the human body thermal distribution image is horizontally long and the high temperature point is located at the left end or the right end of the human body thermal distribution image, it is determined that the human body is in a lying state.

[0120] Please refer to Figure 1-5 In a preferred embodiment, the posture judgment module further includes a first posture judgment submodule:

[0121] The first submodule 1: determining the vertical distance between the upper and lower endpoints of the human body thermal distribution image as its height value, and determining the horizontal distance between the left and right endpoints of the human body thermal distribution image as its width value;

[0122] The first submodule 2: calculating the ratio of the height value to the width value of the human body thermal distribution image;

[0123] First submodule three: judging whether the human body thermal distribution image is vertically distributed or horizontally distributed according to the ratio of the height value to the width value of the human body thermal distribution image.

[0124] Please refer to Figure 1-5 In a preferred embodiment, the posture judgment module further includes a second posture judgment submodule:

[0125] Second submodule 1: obtaining one or more high temperature points from the human body thermal distribution image;

[0126] Second submodule 2: Determine the distribution form of the human body thermal distribution image according to the relative position relationship between the high temperature point and the low temperature area.

[0127] The beneficial effects of the present invention are as follows: the present invention provides an infrared imaging human posture recognition method and device which realizes low-cost human posture recognition and is beneficial to the large-scale promotion of human monitoring products.

[0128] In this specification, unless otherwise expressly specified or limited, when a first feature is "above" or "below" a second feature, it can mean that the first and second features are in direct contact, or that the first and second features are in indirect contact through an intermediary. Furthermore, when a first feature is "above," "above," or "above" a second feature, it can mean that the first feature is directly above or diagonally above the second feature, or simply means that the first feature is at a higher level than the second feature. When a first feature is "below," "below," or "below" a second feature, it can mean that the first feature is directly below or diagonally below the second feature, or simply means that the first feature is at a lower level than the second feature.

[0129] In the description of this specification, the description with reference to the terms "preferred embodiment", "further embodiment", "other embodiments" or "specific example" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine different embodiments or examples described in this specification and the features of different embodiments or examples, unless they are contradictory.

[0130] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations on the present application. Ordinary technicians in this field can change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A method for human body posture recognition using infrared imaging, characterized in that: The steps include: S11: Collecting spatial temperature distribution through a thermal imaging sensor, and performing smoothing calculations on multiple frames of data to eliminate sudden temperature fluctuations and form a stable temperature distribution image; S12: Using a differential algorithm on the multiple frames of stationary temperature distribution images to eliminate fixed constant-temperature objects, thereby forming a protrusion temperature distribution image; wherein the protrusion temperature distribution image is a 32x32 matrix; S13: Calculate the protrusion temperature distribution image surface, remove the protrusion temperature distribution image below the preset value in the image surface, and clear the background to remove the interference of static objects; the preset value is the set value of the overall outline of the human body; S14: Scanning the temperature distribution image of the protrusion to find the highest temperature point within the preset temperature range, i.e., the high temperature point. Then, with the high temperature point as the center, gradually lowering the temperature value within the preset temperature range to form a thermal distribution contour line, and obtaining a human body thermal distribution image based on the thermal distribution contour line; The specific method for obtaining the human body thermal distribution image includes: S141: Find the high temperature point, take the high temperature point as the center point, obtain the temperature values ​​of multiple adjacent points around the high temperature point, and if the temperature value of the adjacent point is within a preset temperature range, mark the adjacent point as a human body location point, otherwise mark it as a non-human body location point; S142: Taking the marked human body position point as the center point, obtaining temperature values ​​of a plurality of unmarked adjacent points surrounding the center point; if the temperature value of the adjacent point is within a preset temperature range, marking the adjacent point as a human body position point; otherwise, marking the adjacent point as a non-human body position point; S143: repeating step S142 until the temperature values ​​of all unmarked adjacent points surrounding the center point are not within the preset temperature range; or repeating step S142 until all data points of the protrusion temperature distribution image are traversed; S144: Connecting all consecutive human body position points including the high temperature point to obtain a human body thermal distribution image; When there are multiple image contours, the continuous adjacent points formed with the high temperature point are used as the judgment condition, and the images of the adjacent points that do not meet the human body temperature value are correspondingly eliminated; or the judgment of the unmarked adjacent points is repeated, and each temperature point in the image is judged to mark the temperature points within the preset temperature range, exclude the temperature points that do not fall within the preset temperature range, and connect the temperature points that fall within the preset temperature range to form the human body image contour; S2: judging the human body posture according to the position of the high temperature point in the human body thermal distribution image and the distribution form of the human body thermal distribution image.

2. The infrared imaging human posture recognition method according to claim 1, wherein: In step S2: When the human body thermal distribution image is vertically distributed and the high temperature point is located at the upper part of the human body thermal distribution image, it is determined that the human body is in an upright state; When the human body thermal distribution image is distributed in a horizontally long manner and the high temperature point is located at the left end or the right end of the human body thermal distribution image, it is determined that the human body is in a lying state.

3. The infrared imaging human posture recognition method according to claim 2, wherein: Step S2 further includes the following sub-steps: S21: determining the vertical distance between the upper and lower endpoints of the human body thermal distribution image as its height value, and determining the horizontal distance between the left and right endpoints of the human body thermal distribution image as its width value; S22: Calculating the ratio of the height value to the width value of the human body thermal distribution image; S23: judging whether the human body thermal distribution image is vertically distributed or horizontally distributed according to the ratio of the height value to the width value of the human body thermal distribution image.

4. The infrared imaging human posture recognition method according to claim 1, wherein: Step S2 further includes the following sub-steps: S221: Acquire one or more high temperature points from the human body thermal distribution image; S222: Determine the distribution form of the human body thermal distribution image according to the relative position relationship between the high temperature point and the low temperature area.

5. An infrared imaging human posture recognition device, characterized in that: include: Image analysis module: Receives images captured by the thermal imaging sensor, eliminates fixed constant-temperature objects, forms a protrusion temperature distribution image, removes protrusions below a preset value within the image plane, clears the background to remove interference from static objects, and gradually lowers the temperature threshold with the highest temperature point in the image as the center to form thermal distribution contours, forms a human body thermal distribution image, and obtains the highest temperature point and human body contour in the human body thermal distribution image; The forming of the human body thermal distribution image specifically includes: Find the high temperature point, take the high temperature point as the center point, obtain the temperature values ​​of multiple adjacent points around the high temperature point, and if the temperature value of the adjacent point is within a preset temperature range, mark the adjacent point as a human body position point, otherwise mark it as a non-human body position point; Taking the marked human body position point as the center point, obtain the temperature values ​​of multiple unmarked adjacent points around the center point. If the temperature value of the adjacent point is within the preset temperature range, mark the adjacent point as a human body position point, otherwise mark it as a non-human body position point; Repeat the above steps until the temperature values ​​of all unmarked adjacent points around the center point are not within the preset temperature range; or, until all data points of the protrusion temperature distribution image are traversed; Connecting all continuous human body position points including the high temperature point to obtain a human body thermal distribution image; Posture judgment module: judges the human body posture according to the high temperature point obtained by the image processing module, the position of the high temperature point in the human body thermal distribution image and the distribution form of the thermal distribution image.

6. The infrared imaging human posture recognition device according to claim 5, characterized in that: In the posture judgment module: When the human body thermal distribution image is vertically distributed and the high temperature point is located at the upper part of the human body thermal distribution image, it is determined that the human body is in an upright state; When the human body thermal distribution image is distributed in a horizontally long manner and the high temperature point is located at the left end or the right end of the human body thermal distribution image, it is determined that the human body is in a lying state.

7. The infrared imaging human posture recognition device according to claim 5, characterized in that: The posture judgment module further includes a first posture judgment submodule 1, a first posture judgment submodule 2 and a first posture judgment submodule 3: The first posture judgment submodule 1 determines the vertical distance between the upper and lower endpoints of the human body thermal distribution image as its height value, and determines the horizontal distance between the left and right endpoints of the human body thermal distribution image as its width value; The first posture judgment submodule 2 calculates the ratio of the height value to the width value of the human body thermal distribution image; The first posture judgment submodule three: judges whether the human body thermal distribution image is vertically distributed or horizontally distributed according to the ratio of the height value to the width value of the human body thermal distribution image.

8. The infrared imaging human posture recognition device according to claim 5, characterized in that: The posture judgment module further includes a second posture judgment submodule 1 and a second posture judgment submodule 2: Second posture judgment submodule 1: obtaining one or more high temperature points from the human body thermal distribution image; Second posture judgment submodule 2: determining the distribution form of the human body thermal distribution image according to the relative position relationship between the high temperature point and the low temperature area.

Images