WO2021184665A1 - 自动工作系统、自动行走设备及其控制方法及计算机可读存储介质 - Google Patents

自动工作系统、自动行走设备及其控制方法及计算机可读存储介质 Download PDF

Info

Publication number
WO2021184665A1
WO2021184665A1 PCT/CN2020/109479 CN2020109479W WO2021184665A1 WO 2021184665 A1 WO2021184665 A1 WO 2021184665A1 CN 2020109479 W CN2020109479 W CN 2020109479W WO 2021184665 A1 WO2021184665 A1 WO 2021184665A1
Authority
WO
WIPO (PCT)
Prior art keywords
sub
image
regions
region
control method
Prior art date
Application number
PCT/CN2020/109479
Other languages
English (en)
French (fr)
Inventor
朱绍明
任雪
袁立超
Original Assignee
苏州科瓴精密机械科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 苏州科瓴精密机械科技有限公司 filed Critical 苏州科瓴精密机械科技有限公司
Priority to US17/768,028 priority Critical patent/US20240094739A1/en
Priority to EP20925859.9A priority patent/EP4123406A4/en
Publication of WO2021184665A1 publication Critical patent/WO2021184665A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
    • G05D1/02—Control of position or course in two dimensions
    • G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
    • G05D1/0231—Control of position or course in two dimensions specially adapted to land vehicles using optical position detecting means
    • G05D1/0246—Control of position or course in two dimensions specially adapted to land vehicles using optical position detecting means using a video camera in combination with image processing means
    • A—HUMAN NECESSITIES
    • A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
    • A01D—HARVESTING; MOWING
    • A01D34/00—Mowers; Mowing apparatus of harvesters
    • A01D34/006—Control or measuring arrangements
    • A01D34/008—Control or measuring arrangements for automated or remotely controlled operation
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
    • G05D1/02—Control of position or course in two dimensions
    • G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
    • G05D1/0231—Control of position or course in two dimensions specially adapted to land vehicles using optical position detecting means
    • G05D1/0238—Control of position or course in two dimensions specially adapted to land vehicles using optical position detecting means using obstacle or wall sensors
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
    • G05D1/02—Control of position or course in two dimensions
    • G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
    • G05D1/0255—Control of position or course in two dimensions specially adapted to land vehicles using acoustic signals, e.g. ultra-sonic singals
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/11—Region-based segmentation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/136—Segmentation; Edge detection involving thresholding
    • A—HUMAN NECESSITIES
    • A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
    • A01D—HARVESTING; MOWING
    • A01D2101/00—Lawn-mowers
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20024—Filtering details
    • G06T2207/20028—Bilateral filtering
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30248—Vehicle exterior or interior
    • G06T2207/30252—Vehicle exterior; Vicinity of vehicle
    • G06T2207/30261—Obstacle

Definitions

  • the invention relates to the field of intelligent control, in particular to an automatic working system, an automatic walking device, a control method thereof, and a computer-readable storage medium.
  • intelligent robots With the continuous advancement of computer technology and artificial intelligence technology, automatic walking equipment and automatic working systems of intelligent robots have slowly entered people's lives, such as intelligent sweeping robots and intelligent lawn mower robots.
  • smart robots are small in size, and are integrated with sensing devices, driving devices, batteries, etc., without manual manipulation, and can travel and work in a prescribed area.
  • the battery power when the battery power is insufficient, it can automatically return to the charging station, dock with the charging station and charge, and continue to travel and work after the charging is completed.
  • the working area of the existing automatic working system is large lawn, and the boundary is mostly energized equipment buried under the ground, so that the intelligent lawn mowing robot can sense .
  • the boundary line is buried under the ground, much manpower and material resources are required.
  • burying the boundary line requires certain requirements, for example, the angle of the corner cannot be less than 90 degrees, etc., which also limits the shape of the lawn for the intelligent lawnmower robot to work to a certain extent.
  • the present invention provides a control method of an autonomous walking device.
  • the control method includes: acquiring a photographed image; processing the photographed image to acquire a processed image; dividing the processed image into at least one sub-areas; respectively Calculate the size of each sub-region A n ; count the number of sub-regions A n > V in the processed image, and record it as the number of special sub-regions N b , where V is the preset threshold; if N b ⁇ 1, then judge The captured image belongs to a lawn area; if N b > 1, it is determined that the captured image belongs to an incomplete lawn area.
  • the step of “segmenting the processed image into at least one sub-region” specifically includes: obtaining the pixel value p of each pixel, preset at least one pixel value range Pn, and assigning the pixel values p in the processed image to the same
  • the pixel points of the pixel value range Pn are respectively grouped into a sub-region.
  • the step of "processing the captured image to obtain a processed image” includes: performing bilateral filtering on the captured image to generate a filtered image; normalizing the filtered image to generate a standard mode image; Perform image segmentation to generate segmented images; perform flood filling processing on the segmented images to obtain a filled image and record it as a processed image.
  • the pyramid mean shift algorithm is used for image segmentation.
  • the step of "respectively calculating the size A n of each sub-region" includes:
  • the step of "calculating the number of sub-regions A n > V in the processed image, and recording it as the number of special sub-regions N b , where V is the preset threshold" includes:
  • the step of "respectively calculating the size A n of each sub-region” includes: separately calculating the area F n of each sub-region;
  • the step of "calculating the number of sub-regions A n > V in the processed image, and recording it as the number of special sub-regions N b , where V is the preset threshold" includes:
  • the step "if N b ⁇ 1, then determine that the captured image belongs to the lawn area" includes:
  • the step of “segmenting the processed image into at least one sub-region” specifically includes: obtaining the pixel value p of each pixel, preset at least one pixel value range Pn, and assigning the pixel values p in the processed image to the same
  • the pixel points of the pixel value range Pn are respectively grouped into a sub-region.
  • the step of "processing the captured image to obtain a processed image” includes: performing bilateral filtering on the captured image to generate a filtered image; normalizing the filtered image to generate a standard mode image; Perform image segmentation to generate segmented images; perform flood filling processing on the segmented images to obtain a filled image and record it as a processed image.
  • the pyramid mean shift algorithm is used for image segmentation.
  • the step of "respectively calculating the size A n of each sub-region" includes:
  • the step "to count the number of sub-regions A n > V in the processed image and record it as the number of special sub-regions Nb, where V is the preset threshold" includes:
  • the step of "respectively calculating the size A n of each sub-region” includes: separately calculating the area Fn of each sub-region;
  • the step of "calculating the number of sub-regions A n > V in the processed image, and recording it as the number of special sub-regions N b , where V is the preset threshold" includes:
  • the step "if N b ⁇ 1, then determine that the captured image belongs to the lawn area" includes:
  • the present invention also provides an automatic working system, including: an automatic traveling device, which can work according to the control method as described above; The working area, the boundary extends upward from the ground.
  • the present invention also provides an automatic working system, including: an automatic walking device that can work according to the above-mentioned control method; a working area, where a non-working area is provided outside the edge of the working area, The geology of the working area and the non-working area are different and form a boundary.
  • the present invention also provides an automatic walking device, including a main body, a walking module, a power supply module, and a memory and a processor arranged in the main body.
  • the memory stores a computer that can run on the processor.
  • the autonomous walking device further includes a camera arranged on the body, the shooting direction of the camera is facing the front side of the autonomous walking device along the traveling direction; when the processor executes the computer program, it can be implemented as described above The steps of the control method of autonomous walking equipment.
  • the autonomous walking equipment further includes an ultrasonic detector, and the ultrasonic detector is installed on the body.
  • the present invention also provides a computer-readable storage medium that stores a computer program, which when executed by a processor, can realize the steps in the control method of the autonomous walking device as described above. .
  • the present invention can process and analyze the photographed image taken by the autonomous walking device, and divide the processed image to obtain at least one sub-region. If a certain sub-region is larger, it can be Determined as a special sub-area; if there are more than one special sub-areas, it can be determined that the captured image belongs to an incomplete lawn area, and it can be determined that there are large obstacles or borders in the captured image, which requires the automatic walking device to retreat, Turning and other operations to avoid; if the number of the special sub-areas does not exceed one, it can be determined that the captured image is entirely a lawn area, even if there are some obstacles such as fallen leaves, small rocks, etc. Therefore, by analyzing the captured images, it is possible to analyze whether the automatic walking encounters boundaries or obstacles, which is more convenient and makes the control more sensitive and effective.
  • Figure 1 is a schematic diagram of the structure of the automatic working system of the present invention
  • Fig. 3 is a schematic flowchart of a second embodiment of the present invention.
  • the self-propelled equipment of the present invention may be an automatic lawn mower, or an automatic vacuum cleaner, etc., which automatically walks in the work area for mowing and vacuuming.
  • the self-propelled equipment is a lawn mower as an example.
  • the working area may be a lawn.
  • the self-propelled equipment is not limited to lawn mowers and vacuum cleaners, but can also be other equipment, such as spraying equipment, snow removal equipment, monitoring equipment, etc., suitable for unattended equipment.
  • a control method of an autonomous walking device 1 is provided, and the control method includes:
  • N b > 1 it is determined that the captured image belongs to an incomplete lawn area.
  • the captured image captured by the autonomous vehicle 1 can be processed and analyzed, and the processed image can be divided to obtain at least one subregion. If a certain subregion is larger, it can be determined as a special subregion. Area; if there are more than one special sub-areas, it can be determined that the captured image belongs to an incomplete lawn area, and it can be determined that there are large obstacles or boundaries 2 in the captured image, which requires the automatic walking device 1 to retreat, turn, etc. Operate to avoid; if the number of the special sub-areas does not exceed one, it can be determined that the captured image is entirely a lawn area, even if there are some obstacles such as fallen leaves, small rocks, etc. Therefore, by analyzing the captured images, it is possible to analyze whether the automatic walking encounters the boundary 2 or obstacles, which is more convenient and makes the control more sensitive and effective.
  • the step of "dividing the processed image into at least one sub-region” specifically includes: obtaining the pixel value p of each pixel, preset at least one pixel value range Pn, and dividing the pixel value p in the processed image into the same pixel value range Pn.
  • the pixels are grouped into a sub-region respectively.
  • the sub-regions are divided according to the pixel value p. Collecting the pixel points with the pixel value p in the same pixel value range Pn into a sub-area, that is, collecting the pixels with similar or the same color in the processed image into a sub-area. Therefore, the pixel value range Pn may be a pixel value interval, and of course, it may also refer to a few specific pixel values.
  • step of "obtaining the pixel value p of each pixel" includes:
  • the processed image includes at least one sub-region, and the pixel values of the pixels in each sub-region are the same.
  • the present invention provides a specific embodiment, which can process the captured image to form the processed image as described above.
  • the step of "processing the captured image to obtain a processed image” includes:
  • the segmented image meanImage is filled with water, and the filled image fillImage is obtained and recorded as the processed image.
  • bilateral filtering is a non-linear filtering method that combines the spatial proximity of the image and the similarity of pixel values. It also considers the spatial information and gray-scale similarity to achieve the purpose of edge preservation and denoising. . It is simple, non-iterative, and partial.
  • the normalization process refers to the process of performing a series of standard processing transformations on the image and transforming the image into a fixed standard form.
  • Image segmentation is a crucial preprocessing in image recognition and computer vision.
  • Image segmentation is based on the brightness and color of pixels in the image, and artificial intelligence is introduced to correct errors in segmentation caused by uneven lighting, shadows, unclear images, or noise.
  • image segmentation can be roughly processed into an image composed of several different area color blocks.
  • image segmentation can adopt a variety of methods, such as a threshold-based segmentation method, a region-based segmentation method, an edge-based segmentation method, and a specific theory-based segmentation method.
  • the pyramid mean shift algorithm is used for image segmentation.
  • the autonomous walking device 1 in the present invention since the autonomous walking device 1 in the present invention usually walks on the grass, it needs to identify large obstacles or boundaries 2. It has been filled with flooded water.
  • the flood filling process refers to filling the connected areas by color, and the purpose of flood filling is achieved by setting the upper and lower limits of the connectable pixels, and the more similar pixel areas can be connected into a whole.
  • the captured image can be processed into a processed image for subsequent analysis.
  • the sub-region size determination A n provides two specific embodiments.
  • the step of "respectively calculating the size A n of each sub-region" includes:
  • the step of "calculating the number of sub-regions A n > V in the processed image, and recording it as the number of special sub-regions N b , where V is the preset threshold" includes:
  • the need to pre-set threshold number V num, the number of V num and pixel sub-region threshold number N p are compared, if the number of subregion pixels of N p exceeds the threshold number V num, It means that the number of pixels in the sub-region is large, and it can be judged that the sub-region occupies a larger area in the processed image; if the number of pixels in the sub-region N p is less than the number threshold V num , it can be determined that the sub-region The number of pixels in the middle is small, and it can be judged that this sub-region occupies a small area in the processed image.
  • the sub-region is grass, and if the number of sub-regions occupying a larger area in the processed image is at least two, it is obvious that one of them is A certain sub-area is a large obstacle or boundary 2.
  • the step of "respectively calculating the size A n of each sub-region” includes: separately calculating the area F n of each sub-region;
  • the step of "calculating the number of sub-regions A n > V in the processed image, and recording it as the number of special sub-regions N b , where V is the preset threshold" includes:
  • the area threshold V F needs to be preset first, and the area threshold V F is compared with the area F n in the sub-region. If the area F n of the sub-region exceeds the area threshold V F , it can be determined The sub-region occupies a larger area in the processed image. Therefore, if the number of sub-regions occupying a larger area in the processed image is one, it is clear that the sub-region is grass, and if the number of sub-regions occupying a larger area in the processed image is at least two, it is obvious that one of them is A certain sub-area is a large obstacle or boundary 2.
  • the number threshold V num and the area threshold V F are related to the size of the captured image, and the larger the size of the captured image, the greater the value of the number threshold V num and the area threshold V F.
  • the number of pixels N p1 in one sub-region is 8284
  • the number of pixels N p2 in another sub-region is 10658
  • the number threshold V num is set to 250, then it can be judged as incomplete lawn.
  • step "if N b ⁇ 1, then determine that the captured image belongs to the lawn area" includes:
  • the sub-area A n ⁇ V it means that the sub-area may be small stones, shadows, bare soil, leaves and other obstacles on the lawn that will not affect the movement of the automatic walking equipment, and these obstacles are not Is the actual obstacle. Therefore, ultrasonic detectors are also installed on the automatic traveling equipment, and the ultrasonic detectors are used to detect the distance between these sub-areas and the automatic traveling equipment. If the distance is large and exceeds the preset distance threshold V S , it means that the sub-area cannot Cause actual obstacles; if the distance is small and not greater than the preset distance threshold V S , it means that the sub-area is a small obstacle, such as a tall stone, and it will affect the travel of the automatic walking device.
  • control method includes:
  • N>1 calculate the size A n of each sub-region separately, count the number of sub-regions A n > V in the processed image, and record it as the number of special sub-regions N b , where V is the preset number threshold; If N b ⁇ 1, it is determined that the captured image belongs to a lawn area; if N b > 1, it is determined that the captured image belongs to an incomplete lawn area.
  • the difference from the above-mentioned first embodiment is that after analyzing the processed image to obtain several sub-regions, the number of sub-regions is first judged. If it is judged that the number of sub-regions is only one, it is directly judged that the captured image belongs to the lawn area, and no other analysis is performed; if it is judged that the number of sub-regions is more than one, then the number of special sub-regions N b is further determined. judge.
  • the step of "dividing the processed image into at least one sub-region" specifically includes: obtaining the pixel value p of each pixel, preset at least one pixel value range Pn, and assigning the pixel value p in the processed image to the same pixel value range Pn
  • the pixels of are grouped into a sub-area.
  • the pixel points with the pixel value p belonging to the same pixel value range Pn are grouped into a sub-region, that is, the pixels of similar or the same color in the processed image are grouped into a sub-region. Therefore, the pixel value range Pn may be a pixel value interval, and of course, it may also refer to a few specific pixel values.
  • the step of "obtaining the pixel value p of each pixel" includes:
  • the step "process the captured image to obtain the processed image” includes:
  • the segmented image meanImage is filled with water, and the filled image fillImage is obtained and recorded as the processed image.
  • the pyramid mean shift algorithm is used for image segmentation.
  • the steps of "calculating the size of each sub-region A n separately " include:
  • the step of "calculating the number of sub-regions A n > V in the processed image, and recording it as the number of special sub-regions N b , where V is the preset threshold" includes:
  • the step of "respectively calculating the size A n of each sub-region” includes: separately calculating the area F n of each sub-region;
  • the step of "calculating the number of sub-regions A n > V in the processed image, and recording it as the number of special sub-regions N b , where V is the preset threshold" includes:
  • step "if N b ⁇ 1, then determine that the captured image belongs to the lawn area" includes:
  • the present invention also provides an automatic working system, which includes:
  • the self-propelled equipment 1 can work according to the above-mentioned control method
  • the boundary 2 is enclosed in a ring shape and used to define the working area of the autonomous vehicle 1, and the boundary 2 extends upward from the ground.
  • the autonomous vehicle 1 obtains the photographed image, and then processes and analyzes the photographed image to obtain the area that the autonomous vehicle 1 travels. Therefore, the boundary 2 of the automatic working system of the present invention must extend upward from the ground. Only then can it be photographed and recognized by the autonomous walking device 1.
  • the present invention also provides an automatic working system, including:
  • the self-propelled equipment 1 can work according to the above-mentioned control method
  • the working area is provided with a non-working area outside the edge of the working area, and the geology of the working area and the non-working area are different and form a boundary 2.
  • the grass is the working area.
  • the non-working area can be bare soil, floor, cement board, etc., and its geology is larger than that of grass. Its color is also quite different from that of grass. Therefore, due to the obvious difference in geology, a boundary 2 will be naturally formed between the working area and the non-working area. The boundary 2 is not artificially set but formed naturally. However, due to the obvious color difference between the working area and the non-working area and the formation of the boundary 2, the control method of the present invention can also be applied.
  • the present invention also provides an autonomous walking device 1, including a main body, a walking module, a power supply module, and a memory and a processor arranged in the main body.
  • the memory stores a computer program that can run on the processor.
  • the walking device 1 also includes a camera arranged on the body, and the shooting direction of the camera faces the front side of the automatic walking device 1 along the traveling direction; the processor can realize the automatic walking as described above when the computer program is executed.
  • the body of the autonomous vehicle 1 in the present invention is provided with a camera, so that the photographed image can be photographed and acquired.
  • the photographing direction of the camera is toward the front side of the autonomous vehicle 1 in the traveling direction, so that the camera captures a view of the front side of the autonomous vehicle 1. Therefore, it is possible to analyze the subsequent movement trajectory of the autonomous vehicle 1 according to the photographed images taken by the autonomous vehicle 1: If it is determined that the photographed image belongs to the lawn area, the autonomous vehicle 1 is controlled to further walk and work; if; if it is judged to be taken If the image belongs to a non-complete lawn area, the automatic walking device 1 is controlled to stop, turn, or retreat.
  • the autonomous walking device 1 further includes an ultrasonic detector, and the ultrasonic detector is installed on the body.
  • the ultrasound probe for detecting the size of a small sub-area sub-area A n S n and the distance between the self-propelled device 1. If it is judged that there is a small obstacle in the captured image, when the small obstacle is reached, the automatic traveling device 1 is controlled to stop, turn or retreat; if it is determined that there is no small obstacle in the captured image, the automatic traveling device can be controlled 1 Travel freely.
  • the present invention also provides a computer-readable storage medium, which stores a computer program, and when the computer program is executed by a processor, it can realize the steps in the control method of the autonomous walking device 1 as described above. That is to say, when the processor executes the computer program, the steps of the control method of any embodiment of the autonomous walking device 1 described above can be implemented.
  • the present invention provides an automatic working system, an automatic traveling device 1 and a control method thereof, and a computer readable storage medium.
  • the control method of the automatic traveling device 1 two implementations are used to It is judged whether the image belongs to the lawn area or the incomplete lawn area. Therefore, by analyzing the captured images, it is possible to analyze whether the automatic walking encounters the boundary 2 or obstacles, which is more convenient and makes the control more sensitive and effective.
  • the present invention also proposes a specific embodiment for processing the captured image, which is mainly by performing bilateral filtering processing, normalization processing, image segmentation and flood filling processing on the captured image, thereby processing the processed image into It includes at least one sub-region, and the pixel values of the pixels in each sub-region are the same or similar.
  • the pyramid mean shift algorithm is used for image segmentation, so that the processing result of the captured image can more satisfy the purpose of the present invention.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Automation & Control Theory (AREA)
  • Aviation & Aerospace Engineering (AREA)
  • Theoretical Computer Science (AREA)
  • Electromagnetism (AREA)
  • Acoustics & Sound (AREA)
  • Multimedia (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Environmental Sciences (AREA)
  • Image Processing (AREA)
  • Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)

Abstract

一种自动工作系统、自动行走设备（1）及其控制方法及计算机可读存储介质，控制方法包括：获取拍摄图像；对拍摄图像进行处理获取处理图像；将处理图像分割为至少一个子区域；分别计算每一个子区域的大小A n；统计处理图像中A n＞V的子区域的数量，并记为特殊子区域的数量N b，其中V为预设数量阈值；若N b≤1，则判断拍摄图像属于草坪区域；若N b＞1，则判断拍摄图像属于非完全草坪区域。若判断为非完全草坪区域时，可确定具有大型障碍物或边界（2）等。通过分析拍摄图像可分析出自动行走设备（1）是否遇到障碍物或边界（2），更加方便，使得控制也更灵敏、有效。

Description

自动工作系统、自动行走设备及其控制方法及计算机可读存储介质 技术领域
本发明涉及智能控制领域，特别是涉及一种自动工作系统、自动行走设备及其控制方法及计算机可读存储介质。
背景技术
随着计算机技术和人工智能技术的不断进步，智能机器人的自动行走设备、自动工作系统已经慢慢进入人们的生活，例如智能扫地机器人、智能割草机器人等。通常的，此类智能机器人体积较小，且集成有传感装置、驱动装置、电池等，无需人工操控，并可在规定的区域内行进并工作。并且，在电池电量不够时，可自动返回充电站，与充电站对接并充电，充电完成后继续行进和工作。
针对现有的智能割草机器人来说，现有的自动工作系统的工作区域均为较大的草坪，并且边界大多是为埋设在地面下的可通电设备，从而可使得智能割草机器人感应到。但是，若在地面下埋设边界线，需要花费较多的人力和物力。并且埋设边界线需要一定的要求，例如拐角的角度不能小于90度等，因而也在一定程度上限制了供智能割草机器人工作的草坪的形状。
因此，必须设计一种较为方便、可以搭建在地面上的自动工作系统，及相应的自动行走设备及其控制方法及计算机可读存储介质。
发明内容
为解决上述问题之一，本发明提供了一种自动行走设备的控制方法，所述控制方法包括：获取拍摄图像；对拍摄图像进行处理获取处理图像；将处理图像分割为至少一个子区域；分别计算每一个子区域的大小A n；统计处理图像中A n＞V的子区域的数量，并记为特殊子区域的数量N b，其中V为预设阈值；若N b≤1，则判断该拍摄图像属于草坪区域；若N b＞1，则判断该拍摄图像属于非完全草坪区域。
作为本发明的进一步改进，步骤“将处理图像分割为至少一个子区域”具体为：获取每个像素点的像素值p，预设至少一个像素值范围Pn，将处理图像中像素值p属于同一像素值范围Pn的像素点分别集合为一个子区域。
作为本发明的进一步改进，步骤“获取每个像素点的像素值p”包括：预设系数a、b、c；分别获取每个像素点的R、G、B值；计算每个像素点的像素值p，p＝a*R+b*G+c*B。
作为本发明的进一步改进，步骤“对拍摄图像进行处理获取处理图像”包括：对拍摄图像进行双边滤波处理，生成滤波图像；对滤波图像进行归一化处理，生成标准模式图像；对标准模式图像进行图像分割，生成分割图像；对分割图像进行漫水填充处理，获得填充图像并记为处理图像。
作为本发明的进一步改进，步骤“对标准模式图像进行图像分割”中，采用金字塔均值漂移算法进行图像分割。
作为本发明的进一步改进，步骤“分别计算每一个子区域的大小A n”包括：
分别计算每一个子区域内像素点的数量N p；
步骤“统计处理图像中A n＞V的子区域的数量，并记为特殊子区域的数量N b，其中V为预设阈值”包括：
统计处理图像中N p大于V num的子区域的数量，并记为特殊子区域的数量N b，其中V num为预设数量阈值；
或者，
步骤“分别计算每一个子区域的大小A n”包括：分别计算每一个子区域的面积F n；
步骤“统计处理图像中A n＞V的子区域的数量，并记为特殊子区域的数量N b，其中V为预设阈值”包括：
统计处理图像中F n大于V F的子区域的数量，并记为特殊子区域的数量N b，其中V F为预设面积阈值。
作为本发明的进一步改进，步骤“若N b≤1，则判断该拍摄图像属于草坪区域”之后包括：
超声波探测A n≤V的子区域与自动行走设备之间的距离S n；
统计S n≤V S的子区域的数量，并记为小型障碍物的数量N S，其中，V S为预设距离阈值；
若N S＞0，则说明该拍摄图像中具有小型障碍物；
若N S＝0，则说明该拍摄图像中无小型障碍物。
为解决上述问题之一，本发明提供了一种自动行走设备的控制方法，所述控制方法包括：获取拍摄图像；对拍摄图像进行处理获取处理图像；将处理图像分割为至少一个子区域；计算子区域的数量N；若N＝1，则判断该拍摄图像属于草坪区域；若N＞1，则分别计算每一个子区域的大小A n，统计处理图像中A n＞V的子区域的数量，并记为特殊子区域的数量N b，其中V为预设阈值；若N b≤1，则判断该拍摄图像属于草坪区域；若N b＞1，则判断该拍摄图像属于非完全草坪区域。
作为本发明的进一步改进，步骤“将处理图像分割为至少一个子区域”具体为：获取每个像素点的像素值p，预设至少一个像素值范围Pn，将处理图像中像素值p属于同一像素值范围Pn的像素点分别集合为一个子区域。
作为本发明的进一步改进，步骤“获取每个像素点的像素值p”包括：预设系数a、b、c；分别获取每个像素点的R、G、B值；计算每个像素点的像素值p，p＝a*R+b*G+c*B。
作为本发明的进一步改进，步骤“对拍摄图像进行处理获取处理图像”包括：对拍摄图像进行双边滤波处理，生成滤波图像；对滤波图像进行归一化处理，生成标准模式图像；对标准模式图像进行图像分割，生成分割图像；对分割图像进行漫水填充处理，获得填充图像并记为处理图像。
作为本发明的进一步改进，步骤“对标准模式图像进行图像分割”中，采用金字塔均值漂移算法进行图像分割。
作为本发明的进一步改进，步骤“分别计算每一个子区域的大小A n”包括：
分别计算每一个子区域内像素点的数量N p；
步骤“统计处理图像中A n＞V的子区域的数量，并记为特殊子区域的数量 Nb，其中V为预设阈值”包括：
统计处理图像中N p大于V num的子区域的数量，并记为特殊子区域的数量N b，其中V num为预设数量阈值；
或者，
步骤“分别计算每一个子区域的大小A n”包括：分别计算每一个子区域的面积Fn；
步骤“统计处理图像中A n＞V的子区域的数量，并记为特殊子区域的数量N b，其中V为预设阈值”包括：
统计处理图像中F n大于V F的子区域的数量，并记为特殊子区域的数量N b，其中V F为预设面积阈值。
作为本发明的进一步改进，步骤“若N b≤1，则判断该拍摄图像属于草坪区域”之后包括：
超声波探测A n≤V的子区域与自动行走设备之间的距离S n；
统计S n≤V S的子区域的数量，并记为小型障碍物的数量N S，其中，V S为预设距离阈值；
若N S＞0，则说明该拍摄图像中具有小型障碍物；
若N S＝0，则说明该拍摄图像中无小型障碍物。
为解决上述问题之一，本发明还提供了一种自动工作系统，包括：自动行走设备，可按照如上述所述的控制方法工作；边界，围设呈环状并形成用以限定自动行走设备的工作区域，所述边界自地面向上延伸。
为解决上述问题之一，本发明还提供了一种自动工作系统，包括:自动行走设备，可按照如上述所述的控制方法工作；工作区域，所述工作区域边沿外侧设置有非工作区域，所述工作区域和非工作区域的地质不同且形成边界。
为解决上述问题之一，本发明还提供了一种自动行走设备，包括本体、行走模块、电源模块及设置于本体内的存储器和处理器，所述存储器存储有可在处理器上运行的计算机程序，所述自动行走设备还包括设置于本体上的摄像头，所述摄像头的拍摄方向朝向该自动行走设备沿行进方向的前侧；所述处理器执 行所述计算机程序时可实现如上述所述的自动行走设备的控制方法的步骤。
作为本发明的进一步改进，所述自动行走设备还包括有超声波探测器，所述超声波探测器安装于所述本体上。
为解决上述问题之一，本发明还提供了一种计算机可读存储介质，其存储有计算机程序，该计算机程序被处理器执行时可实现如上述所述的自动行走设备的控制方法中的步骤。
与现有技术相比，本发明中，可对自动行走设备所拍摄到的拍摄图像进行处理和分析，并将处理图像分割从而获得至少一个子区域，则，若某个子区域较大，则可判定为特殊子区域；若该特殊子区域设置多于一个，则可判断该拍摄图像属于非完全草坪区域，可确定该拍摄图像中具有大型障碍物或边界等，从而需要自动行走设备进行后退、转弯等操作以进行避让；若该特殊子区域数量不超过一个，则可确定该拍摄图像中完全为草坪区域，即使具有一些障碍物如落叶、小型石块等。从而，通过分析拍摄图像可分析出自动行走是否遇到边界或障碍物，更加方便，使得控制也更加灵敏、有效。
附图说明
图1为本发明自动工作系统的结构示意图；
图2为本发明第一种实施方式的流程示意图；
图3为本发明第二种实施方式的流程示意图。
具体实施例
为了使本技术领域的人员更好地理解本发明中的技术方案，下面将结合本发明实施例中的附图，对本发明实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本发明一部分实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都应当属于本发明保护的范围。
在本申请的各个图示中，为了便于图示，结构或部分的某些尺寸会相对于 其它结构或部分夸大，因此，仅用于图示本申请的主题的基本结构。
本发明的自动行走设备可以是自动割草机，或者自动吸尘器等，其自动行走于工作区域以进行割草、吸尘工作，本发明具体示例中，以自动行走设备为割草机为例做具体说明，相应的，所述工作区域可为草坪。当然，自动行走设备不限于割草机和吸尘器，也可以为其它设备，如喷洒设备、除雪设备、监视设备等等适合无人值守的设备。
如图1和图2所示，在本发明的第一种实施方式中，提供了一种自动行走设备1的控制方法，所述控制方法包括：
获取拍摄图像；
对拍摄图像进行处理获取处理图像；
将处理图像分割为至少一个子区域；
分别计算每一个子区域的大小A n；
统计处理图像中A n＞V子区域的数量，并记为特殊子区域的数量N b，其中V为预设阈值；
若N b≤1，则判断该拍摄图像属于草坪区域；
若N b＞1，则判断该拍摄图像属于非完全草坪区域。
因此，本发明中，可对自动行走设备1所拍摄到的拍摄图像进行处理和分析，并将处理图像分割从而获得至少一个子区域，则，若某个子区域较大，则可判定为特殊子区域；若该特殊子区域设置多于一个，则可判断该拍摄图像属于非完全草坪区域，可确定该拍摄图像中具有大型障碍物或边界2等，从而需要自动行走设备1进行后退、转弯等操作以进行避让；若该特殊子区域数量不超过一个，则可确定该拍摄图像中完全为草坪区域，即使具有一些障碍物如落叶、小型石块等。从而，通过分析拍摄图像可分析出自动行走是否遇到边界2或障碍物，更加方便，使得控制也更加灵敏、有效。
其中，步骤“将处理图像分割为至少一个子区域”具体为：获取每个像素点的像素值p，预设至少一个像素值范围Pn，将处理图像中像素值p属于同一像素值范围Pn的像素点分别集合为一个子区域。本发明中，是依据像素值p对 子区域进行划分的。将像素值p属于同一像素值范围Pn的像素点集合为一个子区域，即是将处理图像中颜色相近或相同的像素点集合为一个子区域。因而，该像素值范围Pn可以为一段像素值区间，当然也可指具体的某几个像素值。
进一步的，步骤“获取每个像素点的像素值p”包括：
预设系数a、b、c；
分别获取每个像素点的R、G、B值；
计算每个像素点的像素值p，p＝a*R+b*G+c*B。
在本发明中，需要对拍摄图像进行处理分析后获得处理图像，进而再在处理图像中进行像素点的像素值的计算及子区域的划分。并且，系数a、b、c可由草坪的颜色所决定。其中，显然的，通过对拍摄图像进行处理，获得的所述处理图像中包括至少一块子区域，且每块子区域内的像素的像素值相同。
本发明中提供了一种具体实施例，可将拍摄图像处理形成如上述所述的处理图像。具体的，步骤“对拍摄图像进行处理获取处理图像”包括：
对拍摄图像进行双边滤波处理，生成滤波图像filterImage；
对滤波图像filterImage进行归一化处理，生成标准模式图像norImage；
对标准模式图像norImage进行图像分割，生成分割图像meanImage；
对分割图像meanImage进行漫水填充处理，获得填充图像fillImage并记为处理图像。
其中，双边滤波处理是一种非线性的滤波方法，是结合图像的空间邻近度和像素值相似度的一种折中处理，同时考虑空域信息和灰度相似性，达到保边去噪的目的。具有简单、非迭代、局部的特点。
而，归一化处理是指对图像进行一系列标准的处理变换、将图像变换为一固定标准形式的过程。
图像分割是图像识别和计算机视觉中至关重要的预处理。图像分割的依据为图像中像素的亮度及颜色，并通过引入人工智能的方法来纠正分割中由于光照不均匀、阴影、图像不清晰、或存在噪声等造成的错误。通过图像分割，可大致将图像处理成由若干个不同区域色块组成的图像。在本发明中从而将标准 图像转换为类似于处理图像的图像。并且，图像分割可采用多种方式，例如基于阈值的分割方法、基于区域的分割方法、基于边缘的分割方法以及基于特定理论的分割方法等。在本发明中，具体的，步骤“对标准模式图像norImage进行图像分割”中，采用金字塔均值漂移算法进行图像分割。
进一步的，由于本发明中的自动行走设备1通常在草地上行走，因而需要识别大型障碍物或边界2、草地即可，因而为了使得最终的处理图像更容易分析，本发明中还对分割图像进行了漫水填充处理。漫水填充处理是指通过颜色填充连通的区域，通过设置可连通像素的上下限来达到漫水填充的目的，并可将较为类似的像素区域连通为一体。
因此，通过上述方法，可将拍摄图像处理形成处理图像，以进行后续的分析。
进一步的，本发明中，对判断子区域的大小A n提供了两种具体实施例。
第一种实施例中，步骤“分别计算每一个子区域的大小A n”包括：
分别计算每一个子区域内像素点的数量N p；
步骤“统计处理图像中A n＞V的子区域的数量，并记为特殊子区域的数量N b，其中V为预设阈值”包括：
统计处理图像中N p大于V num的子区域的数量，并记为特殊子区域的数量N b，其中V num为预设数量阈值。
在该实施例中，需要先预设数量阈值V num，将数量阈值V num与子区域中的像素点的数量N p进行比较，则若子区域中像素点的数量N p超过数量阈值V num，则说明子区域中像素点的数量较多，则可判断该子区域在处理图像中占据较大面积；若子区域中像素点的数量N p少于该数量阈值V num，则可说明该子区域中像素点的数量较少，可判断该子区域在处理图像中占据较小面积。从而，若在处理图像中占据较大面积的子区域的数量为一个，则显然该子区域为草地，若在处理图像中占据较大面积的子区域的数量为至少两个，则显然其中的某个子区域为大型障碍物或边界2。
或者，另一种实施例中，步骤“分别计算每一个子区域的大小A n”包括： 分别计算每一个子区域的面积F n；
步骤“统计处理图像中A n＞V的子区域的数量，并记为特殊子区域的数量N b，其中V为预设阈值”包括：
统计处理图像中F n大于V F的子区域的数量，并记为特殊子区域的数量N b，其中V F为预设面积阈值。
同样的，在本实施例中，需要先预设面积阈值V F，将面积阈值V F与子区域中面积F n进行比较，则若子区域的面积F n超过面积阈值V F，则可判断该子区域在处理图像中占据较大面积。从而，若在处理图像中占据较大面积的子区域的数量为一个，则显然该子区域为草地，若在处理图像中占据较大面积的子区域的数量为至少两个，则显然其中的某个子区域为大型障碍物或边界2。
当然，该数量阈值V num和面积阈值V F与拍摄图像的尺寸相关，且拍摄图像的尺寸越大，数量阈值V num及面积阈值V F的值越大。例如在一具体实施例中，一个子区域的像素点个数N p1为8284，另一个子区域的像素点个数N p2为10658，设置数量阈值V num为250，则即可判断为非完全草坪。
另外，步骤“若N b≤1，则判断该拍摄图像属于草坪区域”之后包括：
超声波探测A n≤V的子区域与自动行走设备之间的距离S n；
统计S n≤V S的子区域的数量，并记为小型障碍物的数量N S，其中，V S为预设距离阈值；
若N S＞0，则说明该拍摄图像中具有小型障碍物；
若N S＝0，则说明该拍摄图像中无小型障碍物。
本发明中，若子区域的大小A n≤V，则说明该子区域可能为草坪上的小石块、阴影、裸露土壤、叶片等不会影响自动行走设备行进的障碍物，而这些障碍物并非为实际障碍物。因此，还在自动行走设备上设置超声波探测器，采用超声探测器探测这些子区域与自动行走设备之间的距离，若距离大，且超过了预设距离阈值V S，则说明该子区域无法造成实际的阻碍；若距离小，且不大于预设距离阈值V S，则说明该子区域上为小型障碍物，例如较高的石块，且会影响自动行走设备的行进。
如图1和图3所示，在本发明的第二种实施方式中，提供了另一种自动行走设备1的控制方法，所述控制方法包括：
获取拍摄图像；
对拍摄图像进行处理获取处理图像；
将处理图像分割为至少一个子区域；
计算子区域的数量N；
若N＝1，则判断该拍摄图像属于草坪区域；
若N＞1，则分别计算每一个子区域的大小A n，统计处理图像中A n＞V的子区域的数量，并记为特殊子区域的数量N b，其中V为预设数量阈值；若N b≤1，则判断该拍摄图像属于草坪区域；若N b＞1，则判断该拍摄图像属于非完全草坪区域。
与上述第一种实施方式中不同的是，在对处理图像进行分析获得若干个子区域后，先对子区域的数量进行判断。若判断子区域的数量仅为一个，那直接判断该拍摄图像属于草坪区域，且不在进行其他的分析；若判断子区域的数量不止为一个，则再进一步的进行特殊子区域的数量N b的判断。
进一步的，步骤“将处理图像分割为至少一个子区域”具体为：获取每个像素点的像素值p，预设至少一个像素值范围Pn，将处理图像中像素值p属于同一像素值范围Pn的像素点分别集合为一个子区域。同样的，将像素值p属于同一像素值范围Pn的像素点集合为一个子区域，即是将处理图像中颜色相近或相同的像素点集合为一个子区域。因而，该像素值范围Pn可以为一段像素值区间，当然也可指具体的某几个像素值。
其中，步骤“获取每个像素点的像素值p”包括：
预设系数a、b、c；
分别获取每个像素点的R、G、B值；
计算每个像素点的像素值p，p＝a*R+b*G+c*B。
步骤“对拍摄图像进行处理获取处理图像”包括：
对拍摄图像进行双边滤波处理，生成滤波图像filterImage；
对滤波图像filterImage进行归一化处理，生成标准模式图像norImage；
对标准模式图像norImage进行图像分割，生成分割图像meanImage；
对分割图像meanImage进行漫水填充处理，获得填充图像fillImage并记为处理图像。
步骤“对标准模式图像进行图像分割”中，采用金字塔均值漂移算法进行图像分割。
步骤“分别计算每一个子区域的大小A n”包括：
分别计算每一个子区域内像素点的数量N p；
步骤“统计处理图像中A n＞V的子区域的数量，并记为特殊子区域的数量N b，其中V为预设阈值”包括：
统计处理图像中N p大于V num的子区域的数量，并记为特殊子区域的数量N b，其中V num为预设数量阈值；
或者，
步骤“分别计算每一个子区域的大小A n”包括：分别计算每一个子区域的面积F n；
步骤“统计处理图像中A n＞V的子区域的数量，并记为特殊子区域的数量N b，其中V为预设阈值”包括：
统计处理图像中F n大于V F的子区域的数量，并记为特殊子区域的数量N b，其中V F为预设面积阈值。
另外，步骤“若N b≤1，则判断该拍摄图像属于草坪区域”之后包括：
超声波探测A n≤V的子区域与自动行走设备之间的距离S n；
统计S n≤V S的子区域的数量，并记为小型障碍物的数量N S，其中，V S为预设距离阈值；
若N S＞0，则说明该拍摄图像中具有小型障碍物；
若N S＝0，则说明该拍摄图像中无小型障碍物。
以上均与第一种实施方式中相同，在此不再赘述。
本发明中还提供了一种自动工作系统，其包括：
自动行走设备1，可按照如上述所述的控制方法工作；
边界2，围设呈环状并用以限定自动行走设备1的工作区域，所述边界2自地面向上延伸。
由于本发明中，所述自动行走设备1通过获取拍摄图像，再对拍摄图像进行处理分析后获得自动行走设备1行进的区域，因而，本发明自动工作系统的边界2必须自地面向上延伸，从而才可以被自动行走设备1拍摄并识别到。
另外，本发明中还提供了一种自动工作系统，包括:
自动行走设备1，可按照如上述所述的控制方法工作；
工作区域，所述工作区域边沿外侧设置有非工作区域，所述工作区域和非工作区域的地质不同且形成边界2。
由于本发明中的自动行走设备1应用于割草机中，则草地即为工作区域，显然的，非工作区域中可为裸露土壤、地板、水泥板等等，其地质均和草地有较大的不同，其颜色也和草地有较大的不同。因而，由于地质的明显差距会自然的在工作区域和非工作区域之间形成边界2，该边界2不是通过人为进行设置而是自然形成的。但是，由于工作区域和非工作区域之间明显的颜色差别及边界2的形成，也还是可以应用本发明中的控制方法。
本发明中还提供了一种自动行走设备1，包括本体、行走模块、电源模块及设置于本体内的存储器和处理器，所述存储器存储有可在处理器上运行的计算机程序，所述自动行走设备1还包括设置于本体上的摄像头，所述摄像头的拍摄方向朝向该自动行走设备1沿行进方向的前侧；所述处理器执行所述计算机程序时可实现如上述所述的自动行走设备1的控制方法的步骤。也就是说，所述处理器执行所述计算机程序时可实现如上述所述的自动行走设备1的任何一种实施例的控制方法的步骤。
如上述所述，本发明中的自动行走设备1的本体上设置有摄像头，从而可以拍摄并获取拍摄图像。并且，摄像头的拍摄方向朝向自动行走设备1沿行进方向的前侧，从而所述摄像头拍摄获得的是自动行走设备1前侧的景象。因而，可以根据自动行走设备1所拍摄获得的拍摄图像来分析自动行走设备1接下来 的运动轨迹：若判断拍摄图像属于草坪区域，则控制自动行走设备1进一步行走及工作；若；若判断拍摄图像属于非完全草坪区域，则控制自动行走设备1进行停止或转向或后退等操作。
所述自动行走设备1还包括有超声波探测器，所述超声波探测器安装于所述本体上。所述超声波探测器用以检测子区域大小A n较小的子区域和自动行走设备1之间的距离S n。若判断该拍摄图像中具有小型障碍物，则在到达该小型障碍物时控制自动行走设备1进行停止或转向或后退等操作；若判断该拍摄图像中无小型障碍物，则可控制自动行走设备1自由行进。
同样的，本发明中还提供了一种计算机可读存储介质，其存储有计算机程序，该计算机程序被处理器执行时可实现如上述所述的自动行走设备1的控制方法中的步骤。也就是说，所述处理器执行所述计算机程序时可实现如上述所述的自动行走设备1的任何一种实施例的控制方法的步骤。
综上所述，本发明中提供了一种自动工作系统、自动行走设备1及其控制方法及计算机可读存储介质，其中，自动行走设备1的控制方法中，通过两种实施方式，对拍摄图像是属于草坪区域还是非完全草坪区域进行判断。从而，通过分析拍摄图像可分析出自动行走是否遇到边界2或障碍物，更加方便，使得控制也更加灵敏、有效。进一步的，本发明中还提出了一种处理拍摄图像的具体实施例，其主要是通过对拍摄图像进行双边滤波处理、归一化处理、图像分割及漫水填充处理，从而将处理图像处理为包括至少一块子区域，且每块子区域内的像素的像素值相同或近似。并特别的采用金字塔均值漂移算法进行图像分割，使得拍摄图像处理结果更加能够满足本发明的目的。
此外，应当理解，虽然本说明书按照实施方式加以描述，但并非每个实施方式仅包含一个独立的技术方案，说明书的这种叙述方式仅仅是为清楚起见，本领域技术人员应当将说明书作为一个整体，各实施方式中的技术方案也可以经适当组合，形成本领域技术人员可以理解的其他实施方式。
上文所列出的一系列的详细说明仅仅是针对本发明的可行性实施方式的具体说明，并非用以限制本发明的保护范围，凡未脱离本发明技艺精神所作的等 效实施方式或变更均应包含在本发明的保护范围之内。

Claims (19)

  1. 一种自动行走设备的控制方法，其特征在于，所述控制方法包括：
    获取拍摄图像；
    对拍摄图像进行处理获取处理图像；
    将处理图像分割为至少一个子区域；
    分别计算每一个子区域的大小A n；
    统计处理图像中A n＞V的子区域的数量，并记为特殊子区域的数量N b，其中V为预设阈值；
    若N b≤1，则判断该拍摄图像属于草坪区域；
    若N b＞1，则判断该拍摄图像属于非完全草坪区域。
  2. 根据权利要求1所述的控制方法，其特征在于，步骤“将处理图像分割为至少一个子区域”具体为：获取每个像素点的像素值p，预设至少一个像素值范围Pn，将处理图像中像素值p属于同一个像素值范围Pn的像素点分别集合为一个子区域。
  3. 根据权利要求2所述的控制方法，其特征在于，步骤“获取每个像素点的像素值p”包括：
    预设系数a、b、c；
    分别获取每个像素点的R、G、B值；
    计算每个像素点的像素值p，p＝a*R+b*G+c*B。
  4. 根据权利要求1所述的控制方法，其特征在于，步骤“对拍摄图像进行处理获取处理图像”包括：
    对拍摄图像进行双边滤波处理，生成滤波图像；
    对滤波图像进行归一化处理，生成标准模式图像；
    对标准模式图像进行图像分割，生成分割图像；
    对分割图像进行漫水填充处理，获得填充图像并记为处理图像。
  5. 根据权利要求4所述的控制方法，其特征在于，步骤“对标准模式图像进行图像分割”中，采用金字塔均值漂移算法进行图像分割。
  6. 根据权利要求1所述的控制方法，其特征在于，步骤“分别计算每一个子区域的大小A n”包括：
    分别计算每一个子区域内像素点的数量N p；
    步骤“统计处理图像中A n＞V的子区域的数量，并记为特殊子区域的数量N b，其中V为预设阈值”包括：
    统计处理图像中N p大于V num的子区域的数量，并记为特殊子区域的数量N b，其中V num为预设数量阈值；
    或者，
    步骤“分别计算每一个子区域的大小A n”包括：分别计算每一个子区域的面积F n；
    步骤“统计处理图像中A n＞V的子区域的数量，并记为特殊子区域的数量N b，其中V为预设阈值”包括：
    统计处理图像中F n大于V F的子区域的数量，并记为特殊子区域的数量N b，其中V F为预设面积阈值。
  7. 根据权利要求1所述的控制方法，其特征在于，步骤“若N b≤1，则判断该拍摄图像属于草坪区域”之后包括：
    超声波探测A n≤V的子区域与自动行走设备之间的距离S n；
    统计S n≤V S的子区域的数量，并记为小型障碍物的数量N S，其中，V S为预设距离阈值；
    若N S＞0，则说明该拍摄图像中具有小型障碍物；
    若N S＝0，则说明该拍摄图像中无小型障碍物。
  8. 一种自动行走设备的控制方法，其特征在于，所述控制方法包括：
    获取拍摄图像；
    对拍摄图像进行处理获取处理图像；
    将处理图像分割为至少一个子区域；
    计算子区域的数量N；
    若N＝1，则判断该拍摄图像属于草坪区域；
    若N＞1，则分别计算每一个子区域的大小A n，统计处理图像中A n＞V的子区域的数量，并记为特殊子区域的数量N b，其中V为预设阈值；若N b≤1，则判断该拍摄图像属于草坪区域；若N b＞1，则判断该拍摄图像属于非完全草坪区域。
  9. 根据权利要求8所述的控制方法，其特征在于，步骤“将处理图像分割为至少一个子区域”具体为：获取每个像素点的像素值p，预设至少一个像素值范围Pn，将处理图像中像素值p属于同一像素值范围Pn的像素点集合为一个子区域。
  10. 根据权利要求9所述的控制方法，其特征在于，步骤“获取每个像素点的像素值p”包括：
    预设系数a、b、c；
    分别获取每个像素点的R、G、B值；
    计算每个像素点的像素值p，p＝a*R+b*G+c*B。
  11. 根据权利要求8所述的控制方法，其特征在于，步骤“对拍摄图像进行处理获取处理图像”包括：
    对拍摄图像进行双边滤波处理，生成滤波图像；
    对滤波图像进行归一化处理，生成标准模式图像；
    对标准模式图像进行图像分割，生成分割图像；
    对分割图像进行漫水填充处理，获得填充图像并记为处理图像。
  12. 根据权利要求11所述的控制方法，其特征在于，步骤“对标准模式图像进行图像分割”中，采用金字塔均值漂移算法进行图像分割。
  13. 根据权利要求8所述的控制方法，其特征在于，步骤“分别计算每一个子区域的大小A n”包括：
    分别计算每一个子区域内像素点的数量N p；
    步骤“统计处理图像中A n＞V的子区域的数量，并记为特殊子区域的数量N b，其中V为预设阈值”包括：
    统计处理图像中N p大于V num的子区域的数量，并记为特殊子区域的数量N b，其中V num为预设数量阈值；
    或者，
    步骤“分别计算每一个子区域的大小A n”包括：分别计算每一个子区域的面积F n；
    步骤“统计处理图像中A n＞V的子区域的数量，并记为特殊子区域的数量N b，其中V为预设阈值”包括：
    统计处理图像中F n大于V F的子区域的数量，并记为特殊子区域的数量N b，其中V F为预设面积阈值。
  14. 根据权利要求8所述的控制方法，其特征在于，步骤“若N b≤1，则判断该拍摄图像属于草坪区域”之后包括：
    超声波探测A n≤V的子区域与自动行走设备之间的距离S n；
    统计S n≤V S的子区域的数量，并记为小型障碍物的数量N S，其中，V S为预设距离阈值；
    若N S＞0，则说明该拍摄图像中具有小型障碍物；
    若N S＝0，则说明该拍摄图像中无小型障碍物。
  15. 一种自动工作系统，其特征在于，包括：
    自动行走设备，可按照如权利要求1至14中的任意一项所述的控制方法工作；边界，围设呈环状并形成用以限定自动行走设备的工作区域，所述边界自地面向上延伸。
  16. 一种自动工作系统，其特征在于，包括:
    自动行走设备，可按照如权利要求1至14中的任意一项所述的控制方法工作；工作区域，所述工作区域边沿外侧设置有非工作区域，所述工作区域和非工作区域的地质不同且形成边界。
  17. 一种自动行走设备，包括本体、行走模块、电源模块及设置于本体内的存储器和处理器，所述存储器存储有可在处理器上运行的计算机程序，其特征在于，所述自动行走设备还包括设置于本体上的摄像头，所述摄像头的拍摄方向朝向该自动行走设备沿行进方向的前侧；所述处理器执行所述计算机程序时可实现权利要求1至14中任意一项所述的自动行走设备的控制方法的步骤。
  18. 根据权利要求17所述的自动行走设备，其特征在于，所述自动行走设备还 包括有超声波探测器，所述超声波探测器安装于所述本体上。
  19. 一种计算机可读存储介质，其存储有计算机程序，其特征在于，该计算机程序被处理器执行时可实现权利要求1至14中任意一项所述的自动行走设备的控制方法中的步骤。
PCT/CN2020/109479 2020-03-19 2020-08-17 自动工作系统、自动行走设备及其控制方法及计算机可读存储介质 WO2021184665A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
US17/768,028 US20240094739A1 (en) 2020-03-19 2020-08-17 Automatic Working System, Automatic Walking Device, and Method for Controlling Same, and Computer-Readable Storage Medium
EP20925859.9A EP4123406A4 (en) 2020-03-19 2020-08-17 AUTOMATICALLY OPERATING SYSTEM, AUTOMATICALLY MOVING DEVICE AND ASSOCIATED CONTROL METHOD, AND COMPUTER READABLE STORAGE MEDIUM

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010198015.9A CN113495553A (zh) 2020-03-19 2020-03-19 自动工作系统、自动行走设备及其控制方法及计算机可读存储介质
CN202010198015.9 2020-03-19

Publications (1)

Publication Number Publication Date
WO2021184665A1 true WO2021184665A1 (zh) 2021-09-23

Family

ID=77769976

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/109479 WO2021184665A1 (zh) 2020-03-19 2020-08-17 自动工作系统、自动行走设备及其控制方法及计算机可读存储介质

Country Status (4)

Country Link
US (1) US20240094739A1 (zh)
EP (1) EP4123406A4 (zh)
CN (1) CN113495553A (zh)
WO (1) WO2021184665A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115471504A (zh) * 2022-11-14 2022-12-13 深圳市富安娜家居用品股份有限公司 一种基于纺织物的线头自动识别方法

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20010022506A1 (en) * 1995-05-30 2001-09-20 Ehud Peless Navigation method and system for autonomous machines with markers defining the working area
US20110150348A1 (en) * 2009-12-17 2011-06-23 Noel Wayne Anderson Automated tagging for landmark identification
CN104111460A (zh) * 2013-04-22 2014-10-22 苏州宝时得电动工具有限公司 自动行走设备及其障碍检测方法
CN107463167A (zh) * 2016-06-03 2017-12-12 苏州宝时得电动工具有限公司 自动行走设备及目标区域识别方法
CN107564071A (zh) * 2016-07-01 2018-01-09 苏州宝时得电动工具有限公司 一种图像识别草地方法及装置
CN109584258A (zh) * 2018-12-06 2019-04-05 南京苏美达智能技术有限公司 草地边界识别方法及应用其的智能割草装置
CN110399840A (zh) * 2019-05-22 2019-11-01 西南科技大学 一种快速的草坪语义分割及边界检测方法

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2011002512A1 (en) * 2009-07-01 2011-01-06 Mtd Products Inc Visual segmentation of lawn grass
CN104111653A (zh) * 2013-04-22 2014-10-22 苏州宝时得电动工具有限公司 自动行走设备及其工作区域判断方法
CN103413308B (zh) * 2013-08-01 2016-07-06 东软集团股份有限公司 一种障碍物检测方法和装置
CN107454969B (zh) * 2016-12-19 2019-10-29 深圳前海达闼云端智能科技有限公司 障碍物检测方法及装置
CN109345539B (zh) * 2018-10-08 2022-03-01 浙江农林大学 基于图像抽象的自适应Mean-Shift立木图像分割方法

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20010022506A1 (en) * 1995-05-30 2001-09-20 Ehud Peless Navigation method and system for autonomous machines with markers defining the working area
US20110150348A1 (en) * 2009-12-17 2011-06-23 Noel Wayne Anderson Automated tagging for landmark identification
CN104111460A (zh) * 2013-04-22 2014-10-22 苏州宝时得电动工具有限公司 自动行走设备及其障碍检测方法
CN107463167A (zh) * 2016-06-03 2017-12-12 苏州宝时得电动工具有限公司 自动行走设备及目标区域识别方法
CN107564071A (zh) * 2016-07-01 2018-01-09 苏州宝时得电动工具有限公司 一种图像识别草地方法及装置
CN109584258A (zh) * 2018-12-06 2019-04-05 南京苏美达智能技术有限公司 草地边界识别方法及应用其的智能割草装置
CN110399840A (zh) * 2019-05-22 2019-11-01 西南科技大学 一种快速的草坪语义分割及边界检测方法

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP4123406A4 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115471504A (zh) * 2022-11-14 2022-12-13 深圳市富安娜家居用品股份有限公司 一种基于纺织物的线头自动识别方法
CN115471504B (zh) * 2022-11-14 2023-02-10 深圳市富安娜家居用品股份有限公司 一种基于纺织物的线头自动识别方法

Also Published As

Publication number Publication date
EP4123406A4 (en) 2024-03-13
CN113495553A (zh) 2021-10-12
US20240094739A1 (en) 2024-03-21
EP4123406A1 (en) 2023-01-25

Similar Documents

Publication Publication Date Title
WO2022021630A1 (zh) 自动行走设备及其控制方法和系统及可读存储介质
WO2021169190A1 (zh) 自动工作系统、自动行走设备及其控制方法及计算机可读存储介质
CN111104933B (zh) 地图处理方法、移动机器人及计算机可读存储介质
WO2016045593A1 (zh) 自移动机器人
WO2021169192A1 (zh) 自动工作系统、自动行走设备及其控制方法及计算机可读存储介质
BR102014027364B1 (pt) sistema de controle e de medição da altura de corte para uma ceifeira de base, método, e, colheitadeira
KR102347511B1 (ko) 잔디경계선에 대한 영상처리를 이용한 잔디깍기 로봇 및 이의 제어 방법
CN111353431A (zh) 自动工作系统、自动行走设备及其控制方法及计算机可读存储介质
WO2014173290A1 (zh) 自动行走设备及其工作区域判断方法
CN110262487B (zh) 一种障碍物检测方法、终端及计算机可读存储介质
WO2021184665A1 (zh) 自动工作系统、自动行走设备及其控制方法及计算机可读存储介质
CN107463167B (zh) 自动行走设备及目标区域识别方法
WO2021184664A1 (zh) 自动工作系统、自动行走设备及其控制方法及计算机可读存储介质
WO2021184663A1 (zh) 自动工作系统、自动行走设备及其控制方法及计算机可读存储介质
CN112558596A (zh) 自动工作系统、自动行走设备及其控制方法及计算机可读存储介质
EP4044114A1 (en) Systems and methods for determining residue length within a field
WO2022000878A1 (zh) 自动行走设备及其控制方法及可读存储介质
Ali et al. Drivable area segmentation in deteriorating road regions for autonomous vehicles using 3D LiDAR sensor
WO2017206950A1 (zh) 自动行走设备及其控制行走方法
CN113223034A (zh) 一种道路边沿检测跟踪方法
WO2021042486A1 (zh) 自动工作系统、自动行走设备及其控制方法及计算机可读存储介质
KR20200108949A (ko) 높이 센싱 알고리즘을 구비한 자율주행 트랙터
WO2021031406A1 (zh) 自动工作系统、自动行走设备及其控制方法
EP4042851A1 (en) Systems and methods for soil clod detection
EP4044128A1 (en) Systems and methods for terrain variation detection

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20925859

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 17768028

Country of ref document: US

NENP Non-entry into the national phase

Ref country code: DE

ENP Entry into the national phase

Ref document number: 2020925859

Country of ref document: EP

Effective date: 20221019