Mirror surface obstacle detection system and robot

Abstract

The utility model discloses a mirror surface obstacle detection system and a robot, the mirror surface obstacle detection system comprises an optical camera and / or a laser radar arranged on a robot body and used for collecting first characteristic information in a current scene; the feature presentation component is arranged on the robot body and used for presenting preset second feature information; and the processor is used for comparing the first feature information acquired by the optical camera and / or the laser radar with the second feature information through imaging information of mirror surface obstacle imaging, and determining that a mirror surface obstacle is detected when judging that the first feature information corresponds to the imaging information. By the adoption of the technical scheme, the problem that in the prior art, a robot is difficult to recognize and detect mirror surface obstacles (including high-reflection objects such as glass), and therefore positioning mapping and obstacle avoidance of the robot are seriously affected is solved, and the mapping quality and the obstacle avoidance effect of the robot are greatly improved.

Description

TECHNICAL FIELD

[0001] The utility model relates to the field of artificial intelligence, specifically, a mirror surface obstacle detection system and robot. BACKGROUND

[0002] In the current robot application scene, the identification and detection of mirror surfaces (including glass and other objects with high reflectivity) are more difficult problems in the robot mapping positioning process, and also seriously affect the robot obstacle avoidance. Because the mirror surface reflects the environment, the glass can transmit light, and the robot perception device based on optics cannot effectively perceive and build obstacles for objects such as glass doors, glass fences, floor-to-ceiling windows, and floor mirrors, ultimately affecting the mapping quality. Accurate and efficient identification of mirror surface obstacles is a key problem that needs to be solved in the robot industry.

[0003] In the prior art, ultrasonic sensors or other non-optical detection methods are usually used to detect mirror surface obstacles that cannot be identified by optical sensors. However, the above mirror surface obstacle detection method has two obvious disadvantages: (1) it increases the cost and structure of the sensor; (2) because the detection accuracy and stability of other non-optical sensors are low, the reliability of the mapping is poor.

[0004] Therefore, without increasing the cost and structure of the sensor, how to accurately and efficiently detect and identify mirror surface obstacles is a problem that needs to be solved at present. UTILITY MODEL CONTENT

[0005] The main purpose of the utility model is to disclose a mirror surface obstacle detection system and robot to at least solve the problem of difficulty in identifying and detecting mirror surface obstacles (including glass and other high-reflective objects) by robots in related technologies.

[0006] According to one aspect of the utility model, a mirror surface obstacle detection system is provided.

[0007] The mirror surface obstacle detection system according to the utility model comprises: an optical camera and / or a laser radar, which is arranged on a robot body and is used to collect first feature information in a current scene; a feature presentation component, which is arranged on the robot body and is used to present predetermined second feature information; and a processor, which is used to compare the first feature information collected by the optical camera and / or the laser radar with the imaging information of the mirror surface obstacle imaging of the second feature information, and determine that a mirror surface obstacle is detected when it is judged that the first feature information corresponds to the imaging information.

[0008] Preferably, the second feature information presented by the feature presentation component includes asymmetric feature information.

[0009] Preferably, the feature presenting component comprises at least one of:

[0010] a component for presenting predetermined pattern information;

[0011] a component for presenting predetermined shape configuration information;

[0012] a component for presenting predetermined color information;

[0013] a component for presenting a signal that flashes in a predetermined time sequence;

[0014] a component for presenting a signal that flashes in a predetermined layout manner.

[0015] Preferably, the system further comprises a light emitting device arranged on the body of the robot for actively emitting full-band light or specific-band detection light, wherein the detection light is used to assist the optical camera and / or the laser radar in collecting the imaging information.

[0016] Preferably, the light emitting device comprises at least one of a line laser sensor and a laser radar.

[0017] Preferably, the feature presenting component is covered with a protective layer through which full-band light or specific-band light can pass.

[0018] Preferably, the protective layer covering the feature presenting component comprises one of:

[0019] a filter arranged above the feature presenting device;

[0020] a filter coating covering the feature presenting device.

[0021] Preferably, the processor is further configured to determine distance information and direction information of the mirror surface obstacle relative to the robot based on the first feature information and the second feature information.

[0022] According to another aspect of the present application, a robot is provided.

[0023] The robot according to the present application comprises the mirror surface obstacle detection system according to any one of the above.

[0024] According to the present application, a mirror surface obstacle detection scheme is provided, and comparison between feature information collected by a robot optical camera or laser radar and feature information arranged on the body of the robot can accurately and efficiently detect and identify mirror surface obstacles, thereby solving the problem that it is difficult for robots to identify and detect mirror surface obstacles (including high-reflective objects such as glass), which seriously affects robot positioning mapping and obstacle avoidance, and greatly improves robot mapping quality and obstacle avoidance effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 is a structural block diagram of a system for detecting mirror obstacles according to an embodiment of the present utility model;

[0026] Figure 2 Schematic diagram of characteristic information set on the robot body according to the first embodiment of the present utility model;

[0027] Figure 3 This is a schematic diagram of characteristic information set on a robot body according to the second embodiment of the present utility model;

[0028] Figure 4 This is a schematic diagram of characteristic information set on a robot body according to the third embodiment of the present utility model;

[0029] Figure 5 1 is a structural block diagram of a mirror obstacle detection system according to a preferred embodiment of the present invention;

[0030] Figure 6 It is a schematic diagram of the positional relationship between the robot feature sequence, the mirror surface, and the mirror surface virtual image according to a preferred embodiment of the present utility model. DETAILED DESCRIPTION

[0031] The specific implementation of the present invention is described in detail below with reference to the accompanying drawings.

[0032] According to an embodiment of the present utility model, a system for detecting mirror obstacles is provided.

[0033] Figure 1 FIG is a structural block diagram of a system for detecting mirror obstacles according to an embodiment of the present utility model. Figure 1 As shown, the mirror obstacle detection system includes: an optical camera 10 and / or a laser radar 12 ( Figure 1 , which shows a situation in which both an optical camera 10 and a laser radar 12 are included), is arranged on the robot body, and is used to collect first feature information in the current scene; a feature presentation component 14 is arranged on the robot body, and is used to present predetermined second feature information; a processor 16 is used to compare the first feature information collected by the above-mentioned optical camera and / or the above-mentioned laser radar with the imaging information of the above-mentioned second feature information through the imaging of the mirror obstacle, and when it is determined that the above-mentioned first feature information corresponds to the above-mentioned imaging information, it is determined that a mirror obstacle is detected.

[0034] In the prior art, in the presence of large-area smooth mirror type obstacles (for example, including high-reflective objects such as glass), using ultrasonic sensors or other non-optical detection methods to detect mirror obstacles that optical sensors cannot recognize will increase sensor cost and structural cost, and the detection accuracy and stability of other sensors are low, which reduces the reliability of mapping. Adopting Figure 1 The mirror obstacle detection system shown in the figure, the optical camera and / or laser radar arranged on the robot body collect first feature information, the feature presentation component arranged on the robot body presents predetermined second feature information, and the processor compares the first feature information collected by the robot with the second feature information presented by the feature presentation component on the robot body. When it is judged that the above-mentioned first feature information corresponds to the above-mentioned imaging information, it is determined that the mirror obstacle is detected. It solves the problem that the recognition and detection of mirror obstacles (including high-reflective objects such as glass) by robots in the prior art are relatively difficult, which seriously affects the robot positioning and mapping and obstacle avoidance, greatly improving the robot mapping quality and obstacle avoidance effect.

[0035] Among them, the processor compares the first feature information collected by the robot with the second feature information presented by the feature presentation component on the robot body, and judges whether the above-mentioned first feature information corresponds to the above-mentioned imaging information. The recognition and comparison algorithm in the prior art can be used, for example, the processor recognizes the collected first feature information (pattern, shape structure, color, layout mode or time sequence of flashing signal, etc.), judges whether the recognized feature information is the reverse sequence combination of the second feature information presented by the feature presentation component on the robot body, and if the recognized feature information is the reverse sequence combination of the second feature information presented by the feature presentation component on the robot body, it can be determined that the above-mentioned first feature information corresponds to the imaging information of the mirror obstacle imaging of the second feature information presented by the feature presentation component on the robot body.

[0036] Preferably, the second feature information presented by the feature presentation component can include asymmetric feature information. Of course, in the specific implementation process, symmetric feature information can also be arranged on the robot body, but if there are multiple robots with the same feature information arranged on the body, the optical camera or laser radar of the current robot will misrecognize the feature information collected from other robot bodies as the imaging information of the feature information arranged on the current robot body through the mirror obstacle imaging, which will cause the misrecognition of the mirror obstacle and thus cause errors in mapping. Therefore, arranging asymmetric feature information on the above-mentioned robot body can avoid the misrecognition of the mirror obstacle in the above-mentioned multiple robot scenario.

[0037] Preferably, the feature presentation component 14 can include at least one of the following:

[0038] a component for presenting predetermined pattern information;

[0039] a component for presenting predetermined shape configuration information;

[0040] a component for presenting predetermined color information;

[0041] a component for presenting a signal that flashes in a predetermined time sequence (for example, the feature presentation component 14 includes a plurality of LED light beads that flash in a predetermined time sequence);

[0042] a component for presenting a signal that flashes in a predetermined layout manner (for example, the feature presentation component 14 includes a plurality of LED light beads that flash in a predetermined layout manner).

[0043] In the preferred implementation, the feature presentation component 14 can present information characterized by a form of information, or information characterized by a combination of multiple forms of information, for example, Figure 2 In the preferred implementation, the feature presentation component 14 presents information characterized by pattern information (including: shape information of a figure, light and shade information of a figure, layout information of a figure, etc.), Figure 3 In the preferred implementation, the feature presentation component 14 presents information characterized by pattern information (including: shape information of a figure, layout information of a figure, etc.), Figure 4 In the preferred implementation, the feature presentation component 14 presents information characterized by shape configuration information.

[0044] Preferably, as shown in Figure 5 The above system can further include a light-emitting device 18 disposed on the body of the robot, configured to actively emit full-band light or specific-band detection light, wherein the detection light is used to assist the optical camera and / or the laser radar to collect the imaging information.

[0045] In the preferred implementation, when the robot approaches a mirror surface obstacle, the high reflectivity of the mirror surface obstacle will reflect the detection light (for example, the detection light actively emitted by the line laser or the laser radar of the robot) emitted by the light-emitting device of the robot into the optical camera or the laser radar of the robot, and the optical camera and / or the laser radar of the robot collects the imaging information of the second feature information imaged through the mirror surface obstacle. Of course, without actively emitting detection light by the light-emitting device, when the robot approaches a mirror surface obstacle, the robot can also be passively imaged on the mirror surface by ambient light, and the high reflectivity of the mirror surface obstacle will reflect the ambient light into the optical camera or the laser radar of the robot, and the optical camera or the laser radar of the robot collects the imaging information of the second feature information imaged through the mirror surface obstacle.

[0046] Preferably, the feature presenting component is covered with a protective layer, and the full-band light or the specific-band light can pass through the protective layer.

[0047] The protective layer covering the feature presenting component can include one of the following:

[0048] A filter disposed above the feature presenting device;

[0049] A filter coating covering the feature presenting device.

[0050] In the preferred implementation, a protective layer, such as a filter medium (e.g., a filter or a filter coating), can be disposed on the second feature information, and the full-band light or the specific-band light can pass through the filter medium. For the specific-band light, the light of the non-visible band (e.g., infrared light) can be set to pass through the filter medium. When the light emitting device of the robot actively emits the detection light of the infrared light, the infrared light can pass through the filter and irradiate on the second feature information, which is imaged on the mirror surface obstacle under the irradiation of the infrared light. The optical camera (e.g., infrared camera) of the robot collects the imaging information of the second feature information imaged through the mirror surface obstacle. Since the visible light cannot pass through the filter medium, the filter medium blocks the second feature information, and the naked eye cannot see the second feature information. On the one hand, the second feature information is protected, such as being protected from being damaged and maintaining the integrity of the feature information. On the other hand, the appearance is beautified, and the appearance characteristics of the robot are not affected.

[0051] Preferably, the processor 16 is further configured to determine the distance information and the direction information of the mirror surface obstacle relative to the robot according to the first feature information and the second feature information.

[0052] Preferably, the processor 16 is further configured to determine the distance information and the direction information of the mirror surface obstacle relative to the robot according to the first feature information and the second feature information, which can further include the following processing:

[0053] determining an included angle a of the imaging information relative to the front direction of the robot according to the center point of the imaging information, and determining the direction information of the mirror surface obstacle relative to the robot according to the angle information of the included angle a;

[0054] The distance information d of the mirror surface obstacle relative to the robot is calculated by the following formula: d=k*w / 2(w' / cos alpha), wherein k is a field of view angle parameter of the optical camera or the laser radar, w is an actual distance between the center points of the two endmost feature information in the second feature information, and w' is a distance between the center points of the two endmost feature information collected in the first feature information of the optical camera or the laser radar.

[0055] Therefore, the robot can accurately and efficiently identify the position and angle relationship of the mirror surface obstacle relative to the robot according to the distance information and the direction information of the mirror surface obstacle relative to the robot determined based on the first feature information and the second feature information, and the mapping quality and the obstacle avoidance effect of the robot are greatly improved.

[0056] As shown in FIG. 1, Figure 6 When the robot approaches the mirror surface obstacle, the light and dark stripe pattern 1 is imaged by the mirror to obtain a light and dark stripe imaging 2 (virtual image), and the included angle alpha between the center point of the light and dark stripe imaging 2 (virtual image) and a straight line formed by the optical camera or the laser radar (for example, a camera lens) of the robot relative to the front direction of the robot is known as the included angle beta between the mirror surface of the mirror surface obstacle and the robot, wherein beta=alpha, and the direction information of the mirror surface obstacle relative to the robot can be determined according to the included angle beta between the mirror surface of the mirror surface obstacle and the robot.

[0057] According to the parameters of the measured virtual image, the corrected width w' / cos alpha of the light and dark stripe imaging 2 (virtual image) can be obtained (based on different optical image acquisition methods, which can be distinguished as the difference between the lens field of view angle and the laser point array resolution, which is expressed as w' occupying the imaging width, that is, w' is the distance between the center points of the two endmost feature information collected in the feature information of the optical camera or the laser radar of the robot).

[0058] The lens field of view angle parameter is k, the actual distance between the center points of the two endmost feature information arranged on the robot body is w, and the virtual image distance from the robot is 2d=k*w / (w' / cos alpha).

[0059] Therefore, the distance d of the mirror surface obstacle relative to the robot is k*w / 2(w' / cos alpha).

[0060] According to the robot provided in the embodiment of the present application, the mirror surface obstacle detection system can be arranged on the robot.

[0061] According to the robot provided in the embodiment of the present application, the mirror surface obstacle detection system can be arranged on the robot.

[0062] It should be noted that the specific details of the above robot can be referred to Figures 1 to 6 The corresponding related description and effect in the embodiment shown are understood, and will not be repeated here.

[0063] To sum up, by means of the above-mentioned embodiments of the utility model, the feature information (for example, asymmetric feature mark) is set on the robot body, the robot compares the feature information collected by the robot with the feature information set on the robot body after the robot collects the feature information, detects and identifies the mirror surface obstacle, and determines the distance information and direction information of the mirror surface obstacle relative to the robot by using the structural feature information, which can accurately and efficiently identify the position angle relationship of the mirror surface object relative to the robot, greatly improves the robot mapping quality and obstacle avoidance effect. The present application provides basic data by using optical camera or laser radar and other sensors, simplifies the sensor requirement, uses the original sensor, and does not need to deploy or rely on other equipment. The power requirement is reduced, the detection efficiency and accuracy are improved, the cost is reduced, in addition, it can be suitable for various optical sensors, suitable for mirror surface detection in various scenes, and provides high universality and practicality.

[0064] The above disclosure is only a few specific embodiments of the utility model, but the utility model is not limited to this, any change that can be thought of by any person skilled in the art should fall within the protection scope of the utility model.

Claims

1. A system for detecting mirror obstacles, characterized in that: include: an optical camera and / or a laser radar, disposed on the robot body, for collecting first feature information in the current scene; A feature presenting component, provided on the robot body, for presenting predetermined second feature information; The processor is used to compare the first feature information collected by the optical camera and / or the lidar with the imaging information of the second feature information through the imaging of the mirror obstacle, and when it is determined that the first feature information corresponds to the imaging information, determine that a mirror obstacle is detected.

2. The system according to claim 1, wherein: The second feature information presented by the feature presenting component includes: asymmetric feature information.

3. The system according to claim 1, wherein: The feature presentation component includes at least one of the following: A component for presenting predetermined pattern information; A component presenting predetermined shape configuration information; A component for presenting predetermined color information; A component that presents a signal that flashes in a predetermined time sequence; A component that presents a signal that flashes in a predetermined pattern.

4. The system according to claim 1, wherein: The system further comprises: A light-emitting device is provided on the body of the robot and is used to actively emit full-band light or detection light of a specific band, wherein the detection light is used to assist the optical camera and / or laser radar in collecting the imaging information.

5. The system according to claim 4, characterized in that The light emitting device includes at least one of the following: a line laser sensor, a laser radar.

6. The system according to claim 1, wherein: The characteristic presentation component is covered with a protective layer, wherein the full-band light or the specific-band light can pass through the protective layer.

7. The system according to claim 6, characterized in that The protective layer covering the feature presentation component includes one of the following: a filter disposed above the feature presentation device; A filter coating is covered on the feature presentation device.

8. The system according to any one of claims 1 to 7, characterized in that The processor is further configured to determine distance information and direction information of the mirror obstacle relative to the robot based on the first feature information and the second feature information.

9. A robot, characterized in that: include: A system for detecting mirror obstacles as claimed in any one of claims 1 to 8.

Images