A carpet recognition method based on laser reflection characteristics

Abstract

The present invention discloses a carpet recognition method based on laser reflection characteristics. The method adopts a recognition device that can be operated on the ground. The recognition device is equipped with a laser that can form a light spot on the ground and a camera component that can shoot the light spot image on the ground. The method uses laser to irradiate the ground and utilizes the different reflection characteristics of different ground materials to the laser, especially the static reflection characteristics and dynamic reflection characteristics of the carpet to the laser are greatly different from those of other ground materials to complete the carpet recognition. The method not only fully utilizes the characteristics of the laser such as good environmental adaptability and low cost, but also has easy recognition and high recognition rate. The method is simple, practical and accurate, can well meet the needs of real-time processing, can effectively improve user experience, and is very suitable for installation on related home appliances such as sweeping robots.

Description

Technical Field

[0001] The invention relates to the application field of sweeping robots, and in particular to a carpet recognition method based on laser reflection characteristics. Background Art

[0002] In recent years, with the continuous advancement of science and technology, various intelligent robots have begun to enter the world of ordinary people and have been widely used in daily life. Among them, sweeping robots can automatically complete the cleaning work of the ground through artificial intelligence technology, and have been loved by the majority of users.

[0003] However, more and more families have carpets on the floor, which requires the sweeping robot to use different modes for different floors. Especially for carpets, the sweeping robot needs to adjust the suction and sweeping force or avoid turning on the mopping mode. Manual settings require constant human intervention, which provides a very poor user experience. Therefore, the sweeping robot needs to be able to automatically identify carpets. Most of the current sweeping robots use ultrasonic sensors to identify carpets based on the different reflection characteristics of ultrasonic waves on the surfaces of objects made of different materials. However, due to the long wavelength of ultrasonic waves, they are easily affected by ambient temperature and ambient noise, which affects the carpet recognition rate. Summary of the invention

[0004] In view of the above-mentioned deficiencies in the prior art, the present invention provides a carpet recognition method based on laser reflection characteristics, which performs recognition according to the different performances of laser irradiation on carpets and other ground materials. The method is simple, practical and highly accurate, and can well meet the needs of real-time processing.

[0005] To achieve the above object, the present invention provides a carpet recognition method based on laser reflection characteristics, using a recognition device that can be operated on the ground, the recognition device is equipped with a laser that can form a light spot on the ground and a camera component that can capture the image of the light spot on the ground. The recognition method specifically includes the following steps:

[0006] Step 1, obtaining an image frame sequence including a plurality of ground spot images;

[0007] Step 2, judging the working mode of the recognition device at each sampling moment based on the image frame sequence, wherein the working mode includes a static mode and a dynamic mode;

[0008] Step 3: If the working mode of the recognition device is the static mode, then go to step 4; if the working mode of the recognition device is the dynamic mode, then go to step 5;

[0009] Step 4, performing carpet recognition based on whether the outline of the light spot in the image frame sequence is regular;

[0010] Step 5: Carpet recognition is performed based on the variation amplitude of the light spot in the image frame sequence.

[0011] Furthermore, in step 2, the working mode of the recognition device is determined based on the image frame sequence, specifically:

[0012] For an image frame sequence {RGB n Each image RGB in (x,y)|n=1,2,…,N} n (x, y), get its corresponding grayscale image P n , and then calculate the judgment parameter TH at the n+2 sampling time = |2P n+1 -P n+2 -P n |, when TH≤ε1, it is determined that the working mode of the identification device at the n+2th sampling moment is a static mode, otherwise it is a dynamic mode, where ε1 is a setting parameter.

[0013] Furthermore, in step 4, the carpet recognition is performed based on whether the outline of the light spot in the image sequence is regular, specifically:

[0014] Step 4.1, obtain the ground spot image RGB at the current sampling time n (x, y), and convert it from RGB color space to YUV color space to get the brightness component image Y n (x,y);

[0015] Step 4.2, obtain the brightness component image Y n The pixel with the highest brightness in (x, y) is (x0, y0) and is in the brightness component image Y n In (x, y), with the pixel point (x0, y0) as the center, the average brightness t1 of the neighborhood δ1 is obtained with a radius of 50, and the average brightness t2 of the neighborhood δ2 is obtained with a radius of 100;

[0016] Step 4.3, based on the brightness average value t1 and the brightness average value t2, a first segmentation threshold th1 is obtained, and the brightness component image Y is segmented according to the first segmentation threshold th1. n (x, y) is binarized to obtain the binary image EY n (x,y);

[0017] Step 4.4, in the binary image EY n On the straight line with pixel (x0, y0) as the center and tgθ as the slope in (x, y), search for the line that is farthest from pixel (x0, y0) and has EY n (x θ ,y θ )=1 θ ,y θ ), let θ={1,2,…,360}, we can get the set corresponding to 360 pixels, which is {(x l,y l )|l=1,2,…,360};

[0018] Step 4.5, calculate the pixel point set {(x l ,y l )|l=1,2,…,360}, and the distance between each pixel and the pixel (x0,y0) is obtained to obtain a one-dimensional vector DL=[d1,d2,…,d 360 ], and perform second-order difference on the one-dimensional vector DL ​​to obtain the one-dimensional vector DF1;

[0019] Step 4.6, counting the total number of elements s1 greater than or equal to 0 in the one-dimensional vector DF1, if s1>360×α1, it is determined that the carpet is recognized at the current sampling moment, otherwise the carpet is not recognized, where α1 is the first control parameter.

[0020] Furthermore, in step 4.2, the process of obtaining the brightness average value is:

[0021]

[0022]

[0023] Where M1 is the number of pixels in the neighborhood δ1, and M2 is the number of pixels in the neighborhood δ2.

[0024] Furthermore, in step 4.3, the process of obtaining the first segmentation threshold is:

[0025]

[0026] Where ε2 is the first adjustment parameter.

[0027] Furthermore, in step 5, the carpet recognition is performed based on the variation amplitude of the light spot in the image frame sequence, specifically:

[0028] Step 5.1: Get the ground spot image RGB at the previous sampling time and the current sampling time n (x,y), RGB n+1 (x, y), and convert it from RGB color space to YUV color space to get the brightness component image Y n (x,y),Y n+1 (x, y), and get the brightness component image Y n (x,y),Y n+1 The pixels with the highest brightness in (x,y) are (x1,y1) and (x2,y2);

[0029] Step 5.2, in the brightness component image Y nIn (x, y), with the pixel point (x1, y1) as the center, the average brightness t3 of the neighborhood δ3 is obtained with a radius of 50, and the average brightness t4 of the neighborhood δ4 is obtained with a radius of 100;

[0030] Step 5.3, in the brightness component image Y n+1 In (x, y), with the pixel (x2, y2) as the center, the average brightness t5 of the neighborhood δ5 is obtained with a radius of 50, and the average brightness t6 of the neighborhood δ6 is obtained with a radius of 100;

[0031] Step 5.4, based on the brightness average values ​​t3, t4, t5, t6, a second segmentation threshold th2 is obtained, and the brightness component image Y is segmented according to the segmentation threshold th2. n (x,y),Y n+1 (x, y) is binarized to obtain the binary image FY n (x,y),FY n+1 (x,y);

[0032] Step 5.5, based on the binary image FY n (x,y),FY n+1 (x, y), and the position relationship between the pixel point (x1, y1) and the pixel point (x2, y2) obtains the judgment threshold S. If S>α2, it is determined that the carpet is recognized at the current sampling moment, otherwise the carpet is not recognized, where α2 is the second control parameter.

[0033] Further, in step 5.4, the second segmentation threshold is specifically:

[0034]

[0035] Where ε3 is the second adjustment parameter.

[0036] Furthermore, in step 5.5, the binary image FY n (x,y),FY n+1 (x, y), and the position relationship between the pixel point (x1, y1) and the pixel point (x2, y2) to obtain the judgment threshold S, which is specifically:

[0037] To move the pixel (x1, y1) to the pixel (x2, y2), you need to move (a, b), where a = x2-x1, b = y2-y1;

[0038] Then the judgment threshold S is:

[0039] S=∑|FY n+1 (x+a,y+b)-FY n (x,y)|.

[0040] The present invention provides a carpet recognition method based on laser reflection characteristics. The method utilizes the advantages of laser such as good adaptability to environmental conditions and low cost, adopts laser to illuminate the ground, and utilizes the different reflection characteristics of laser to different ground materials, especially the static reflection characteristics and dynamic reflection characteristics of carpet to laser are greatly different compared with other ground materials. The method is easy to recognize and has a high recognition rate. The method is simple, practical and accurate, and can well meet the needs of real-time processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the structures shown in these drawings without paying creative work.

[0042] Figure 1 Flow chart of the identification method in an embodiment of the present invention.

[0043] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0044] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] It should be noted that all directional indications in the embodiments of the present invention (such as up, down, left, right, front, back, etc.) are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0046] In addition, in the present invention, descriptions such as "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0047] In the present invention, unless otherwise clearly specified and limited, the terms "connection", "fixation", etc. should be understood in a broad sense. For example, "fixation" can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection, an electrical connection, a physical connection, or a wireless communication connection; it can be a direct connection, or an indirect connection through an intermediate medium, or it can be the internal connection of two elements or the interaction relationship between two elements, unless otherwise clearly defined. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0048] In addition, the technical solutions between the various embodiments of the present invention can be combined with each other, but it must be based on the fact that ordinary technicians in the field can implement it. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0049] This embodiment discloses a carpet recognition method based on laser reflection characteristics, using a recognition device that can be operated on the ground, the recognition device is equipped with a laser that can form a light spot on the ground and a camera component that can capture the ground light spot image, the laser generates an infrared laser and generates a light spot on the ground, and then the ground light spot image is acquired by the camera component. Taking advantage of the characteristics of infrared lasers such as short wavelength, not easily affected by ambient temperature and ambient noise, good environmental adaptability, and low cost, laser is used to illuminate the ground, and different ground materials have different reflection characteristics of lasers, especially the static reflection characteristics and dynamic reflection characteristics of carpets to lasers are quite different from other ground materials, so the recognition is easy and has a high recognition rate. It is simple, practical, and highly accurate, and can well meet the needs of real-time processing.

[0050] The carpet recognition method in this embodiment is mainly applied to a sweeping robot, so a lens and a CMOS image sensor can be directly installed on the sweeping robot as a camera assembly. The lens is installed at the center of the bottom of the sweeping robot, and the CMOS image sensor is installed at the center of the bottom of the lens, and the angle of the CMOS image sensor is vertical to ensure that the CMOS can shoot images externally. The laser is installed at a non-center position at the bottom of the sweeping robot, and a small hole is opened at this position to ensure that the laser can emit laser outward, and the laser is at a certain angle to the horizontal. The laser can be a visible light laser or an infrared laser. The laser can emit point lasers or other shapes such as linear or cross-shaped lasers.

[0051] Furthermore, the recognition device is also equipped with a CPU, which is installed at any position inside the fixed shell. The CMOS image sensor and the CPU are connected via a data cable. During the recognition process, a small laser emits a laser signal to the ground, and the CMOS image sensor takes continuous pictures of the ground and then transmits them to the CPU. The CPU is responsible for analyzing the static laser reflection characteristics (including shape, size, light intensity, etc.) and dynamic laser reflection characteristics (including changes in shape and size, changes in light intensity, etc.) of the ground material from the continuous pictures, and judging the ground material based on these characteristics.

[0052] refer to Figure 1 The identification method in this embodiment specifically includes the following steps 1-5.

[0053] Step 1: Obtain an image frame sequence {RGB n (x, y)|n=1, 2, ..., N}, where N is the total number of ground spot images in the image frame sequence.

[0054] Step 2: Determine the working mode of the recognition device at each sampling moment based on the image frame sequence, wherein the working mode includes a static mode and a dynamic mode. The specific implementation process is as follows:

[0055] For an image frame sequence {RGB n Each image RGB in (x,y)|n=1,2,…,N} n (x, y), get its corresponding grayscale image P n , and then calculate the judgment parameter TH at the n+2 sampling time = |2P n+1 -P n+2 -P n |, when TH≤ε1, it is determined that the working mode of the recognition device at the n+2th sampling moment is the static mode, otherwise it is the dynamic mode, where ε1 is a setting parameter. It should be noted that due to the determination parameter TH=|2P n+1 -P n+2 -P n |, that is, the recognition process in this embodiment can be performed only when there are more than three ground light spot images in the image frame sequence.

[0056] Step 3: If the working mode of the recognition device is the static mode, then go to step 4; if the working mode of the recognition device is the dynamic mode, then go to step 5.

[0057] Step 4: Carpet recognition is performed based on whether the outline of the light spot in the image frame sequence is regular. The specific implementation process is as follows:

[0058] Step 4.1, obtain the ground spot image RGB at the current sampling time n(x, y), and convert it from RGB color space to YUV color space as follows:

[0059]

[0060] Among them, the brightness component image Y n (x,y) and the other two chrominance component images U n (x,y),V n (x, y) is completely separated and more suitable for subsequent processing, so the ground spot image RGB is obtained n (x, y) corresponds to the brightness component image Y n (x,y).

[0061] Step 4.2, obtain the brightness component image Y n The pixel with the highest brightness in (x, y) is (x0, y0), that is, Y n (x0,y0)=maxY n (x, y); and in the brightness component image Y n In (x, y), with the pixel point (x0, y0) as the center and a radius of 50, the average brightness t1 of the neighborhood δ1 is obtained, and the average brightness t2 of the neighborhood δ2 is obtained with a radius of 100. Specifically:

[0062]

[0063]

[0064] Where M1 is the number of pixels in the neighborhood δ1, and M2 is the number of pixels in the neighborhood δ2;

[0065] Step 4.3, based on the brightness average value t1 and the brightness average value t2, the first segmentation threshold th1 is obtained as:

[0066]

[0067] Wherein, ε2 is the first adjustment parameter. In this embodiment, the value range of the first adjustment parameter ε2 is 0 to 0.1;

[0068] And according to the first segmentation threshold th1, the brightness component image Y n (x, y) is binarized to obtain the binary image EY n (x,y), is:

[0069] EY n (x,y)=Y n (x,y)>th1;

[0070] Step 4.4, in the binary image EY nOn the straight line with pixel (x0, y0) as the center and tgθ as the slope in (x, y), search for the line that is farthest from pixel (x0, y0) and has EY n (x θ ,y θ )=1 θ ,y θ ), let θ={1,2,…,360}, we can get the set corresponding to 360 pixels, which is {(x l ,y l )|l=1,2,…,360};

[0071] Step 4.5, calculate the pixel point set {(x l ,y l )|l=1,2,…,360} and the distance between each pixel and the pixel (x0,y0), denoted as {d l |l=1,2,…,360}, the calculation process is:

[0072]

[0073] And further obtain the one-dimensional vector DL ​​composed of distances = [d1, d2, ..., d 360 ], let B1=[1,-2,1], perform the second-order difference on the one-dimensional vector DL ​​composed of distances to obtain the one-dimensional vector DF1, which is:

[0074]

[0075] Step 4.6, count the total number of elements s1 greater than or equal to 0 in the one-dimensional vector DF1, if s1>360×α1, it is determined that the carpet is recognized at the current sampling moment, otherwise the carpet is not recognized, where α1 is the first control parameter with a value range of 0.8 to 1. This judgment is based on the fact that the light spot obtained by irradiating the laser on an ordinary floor is relatively regular, while the light spot obtained by irradiating the laser on a carpet has many burrs and is very irregular.

[0076] Step 5: Carpet recognition is performed based on the variation of the light spot in the image frame sequence. The specific implementation process is as follows:

[0077] Step 5.1: Get the ground spot image RGB at the previous sampling time and the current sampling time n (x,y), RGB n+1 (x, y), and convert it from RGB color space to YUV color space as follows:

[0078]

[0079] Among them, the brightness component image Y n (x, y) and the other two chrominance component images Un (x,y),V n (x, y) is completely separated and more suitable for subsequent processing, so the ground spot image RGB is obtained n (x,y), RGB n+1 (x, y) corresponds to the brightness component image Y n (x,y),Y n+1 (x,y);

[0080] And further get the brightness component image Y n (x,y),Y n+1 The pixel with the highest brightness in (x,y) is (x1,y1) and (x2,y2), that is, Y n (x1,y1)=maxY n (x,y),Y n+1 (x2,y2)=maxY n+1 (x,y);

[0081] Step 5.2, in the brightness component image Y n In (x, y), with the pixel point (x1, y1) as the center and a radius of 50, the average brightness t3 of the neighborhood δ3 is obtained, and the average brightness t4 of the neighborhood δ4 is obtained with a radius of 100, which is:

[0082]

[0083] Where M3 is the number of pixels in the neighborhood δ3, and M4 is the number of pixels in the neighborhood δ4;

[0084] Step 5.3, in the brightness component image Y n+1 In (x, y), with the pixel point (x2, y2) as the center and a radius of 50, the average brightness t5 of the neighborhood δ5 is obtained, and the average brightness t6 of the neighborhood δ6 is obtained with a radius of 100, which is:

[0085]

[0086] Where M5 is the number of pixels in the neighborhood δ5, and M6 is the number of pixels in the neighborhood δ6;

[0087] Step 5.4, based on the brightness average values ​​t3, t4, t5, t6, the second segmentation threshold th2 is obtained as:

[0088]

[0089] Wherein, ε3 is the second adjustment parameter. In this embodiment, the value range of the second adjustment parameter ε3 is 0 to 0.1;

[0090] And according to the segmentation threshold th2, the brightness component image Y n (x,y),Y n+1 (x, y) is binarized to obtain the binary image FY n (x,y),FY n+1 (x,y), is:

[0091]

[0092] Step 5.5, based on the binary image FY n (x,y),FY n+1 (x, y), and the position relationship between the pixel point (x1, y1) and the pixel point (x2, y2) to obtain the judgment threshold S, which is specifically:

[0093] To move the pixel (x1, y1) to the pixel (x2, y2), you need to move (a, b), where a = x2-x1, b = y2-y1;

[0094] Then the judgment threshold S is:

[0095] S=∑|FY n+1 (x+a,y+b)-FY n (x,y)|

[0096] If S>α2, it is determined that the carpet is recognized at the current sampling moment, otherwise the carpet is not recognized, where α2 is the second control parameter. This judgment is based on the fact that the change of the light spot obtained by the laser shining on the ordinary floor is relatively small during movement, while the change of the light spot obtained by the laser shining on the carpet is relatively large.

[0097] The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. All equivalent structural changes made by using the contents of the present invention specification and drawings under the inventive concept of the present invention, or directly / indirectly applied in other related technical fields are included in the patent protection scope of the present invention.

Claims

1. A carpet recognition method based on laser reflection characteristics, characterized in that: A recognition device that can be operated on the ground is used. The recognition device is equipped with a laser that can form a light spot on the ground and a camera component that can capture an image of the light spot on the ground. The recognition method specifically includes the following steps: Step 1, obtaining an image frame sequence including a plurality of ground spot images; Step 2, judging the working mode of the recognition device at each sampling moment based on the image frame sequence, wherein the working mode includes a static mode and a dynamic mode, specifically: For the image frame sequence { RGB n ( x , y )| n =1,2,…, N Each image in RGB n ( x , y ), and get its corresponding grayscale image P n , then calculate the n +2 Judgment parameters of sampling time TH =|2 P n+1 - P n+2 - P n |, when TH ≤ ε 1, this judgment is n +2 The working mode of the sampling time identification device is static mode, otherwise it is dynamic mode, in which, ε 1 is the setting parameter; Step 3: If the working mode of the recognition device is the static mode, then go to step 4; if the working mode of the recognition device is the dynamic mode, then go to step 5; Step 4, performing carpet recognition based on whether the outline of the light spot in the image frame sequence is regular; Step 5: Carpet recognition is performed based on the variation amplitude of the light spot in the image frame sequence.

2. The carpet recognition method based on laser reflection characteristics according to claim 1, characterized in that: In step 4, the carpet recognition is performed based on whether the outline of the light spot in the image frame sequence is regular, specifically: Step 4.1, obtain the ground spot image at the current sampling time RGB n ( x , y ) and change it from RGB Color space conversion to YUV Color space, get the brightness component image Y n ( x , y ); Step 4.2, obtain the brightness component image Y n ( x , y ) is the pixel with the largest brightness ( x 0, y 0), and in the brightness component image Y n ( x , y ) in pixels ( x 0, y 0) as the center and get the neighborhood with a radius of 50 δ The average brightness of 1 t 1, and the neighborhood with a radius of 100 δ The average brightness of 2 t 2; Step 4.3, based on the average brightness t 1 and average brightness t 2 Get the first segmentation threshold th 1, and according to the first segmentation threshold th 1 pair of luminance component images Y n ( x , y ) is binarized to obtain a binary image EY n ( x , y ); Step 4.4, in the binarized image EY n ( x , y ) in pixels ( x 0, y 0) is the center of the circle, tgθ On the straight line with a slope of x 0, y 0) Farthest and EY n ( x θ , y θ )=1 pixels( x θ , y θ ),make θ ={1,2,…,360}, that is, we can get a set corresponding to 360 pixels, which is {( x l , y l )| l =1,2,…,360}; Step 4.5, calculate the pixel point set {( x l , y l )| l =1,2,…,360} and the pixel points ( x 0, y 0) to obtain a one-dimensional vector composed of distances DL =[ d 1, d 2,…, d 360 ], and for the one-dimensional vector DL Take the second-order difference to get a one-dimensional vector DF 1; Step 4.6, count one-dimensional vectors DF The total number of elements greater than or equal to 0 in 1 s 1. If s 1>360× α 1, it is determined that the carpet is recognized at the current sampling moment, otherwise the carpet is not recognized, where: α 1 is the first control parameter.

3. The carpet recognition method based on laser reflection characteristics according to claim 2, characterized in that: In step 4.2, the brightness average t 1. Average brightness t The acquisition process of 2 is: In the formula, M 1 is the neighborhood δ 1 contains the number of pixels, M 2 is the neighborhood δ 2 contains the number of pixels.

4. The carpet recognition method based on laser reflection characteristics according to claim 2, characterized in that: In step 4.3, the process of obtaining the first segmentation threshold is: In the formula, ε 2 is the first adjustment parameter.

5. The carpet recognition method based on laser reflection characteristics according to claim 1, characterized in that: In step 5, the carpet recognition is performed based on the variation amplitude of the light spot in the image frame sequence, specifically: Step 5.1: Get the ground spot image at the previous sampling time and the current sampling time RGB n ( x , y ), RGB n+1 ( x , y ) and change it from RGB Color space conversion to YUV Color space, get the brightness component image Y n ( x , y ), Y n+1 ( x , y ) and obtain the brightness component image Y n ( x , y ), Y n+1 ( x , y ) are the pixels with the largest brightness, respectively ( x 1, y 1) ( x 2, y 2); Step 5.2, in the brightness component image Y n ( x , y ) in pixels ( x 1, y 1) is the center and the neighborhood is obtained with a radius of 50 δ 3 average brightness t 3, and get the neighborhood with a radius of 100 δ Average brightness of 4 t 4; Step 5.3, in the brightness component image Y n+1 ( x , y ) in pixels ( x 2, y 2) as the center, and get the neighborhood with a radius of 50 δ Average brightness of 5 t 5, and get the neighborhood with a radius of 100 δ Average brightness of 6 t 6; Step 5.4, based on the average brightness t 3. t 4. t 5. t 6 Get the second segmentation threshold th 2, and according to the segmentation threshold th 2 For the brightness component image Y n ( x , y ), Y n+1 ( x , y ) is binarized to obtain a binary image FY n ( x , y ), FY n+1 ( x , y ); Step 5.5, based on the binary image FY n ( x , y ), FY n+1 ( x , y ), and pixels ( x 1, y 1) and pixel ( x 2, y 2) The position relationship between them is used to obtain the judgment threshold S ,like S > α 2, it is determined that the carpet is recognized at the current sampling moment, otherwise the carpet is not recognized, where: α 2 is the second control parameter.

6. The carpet recognition method based on laser reflection characteristics according to claim 5, characterized in that: In step 5.4, the second segmentation threshold is specifically: In the formula, ε 3 is the second adjustment parameter.

7. The carpet recognition method based on laser reflection characteristics according to claim 5, characterized in that: In step 5.5, the binary image FY n ( x , y ), FY n+1 ( x , y ), and pixels ( x 1, y 1) and pixel ( x 2, y 2) The position relationship between them is used to obtain the judgment threshold S , specifically: The pixel point ( x 1, y 1) Move to the pixel ( x 2, y 2) Need to move ( a , b ),in, a = x 2- x 1. b = y 2- y 1; The judgment threshold S for: 。

Images