With the development of science and technology, sweepers have gradually entered thousands of households. In the past, sweepers
were nicknamed "artificial intellectual disability" due to poor obstacle avoidance ability, hitting walls everywhere during work, and
after several years of development, sweepers have been able to solve the problem of obstacle avoidance well, which is worthy of
the title of "artificial intelligence". In fact, this progress is precisely the progress of core obstacle avoidance function module technology.
(1) Infrared tube obstacle avoidance (short-range sensing and no ranging function)
(2) Single and binocular obstacle avoidance (high definition, but depth information cannot be directly obtained)
(3) Structured light obstacle avoidance (long distance, can be measured, under certain circumstances, it is interfered by ambient light)
(4) Infrared PSD sensing obstacle avoidance (short distance, can be measured, but the accuracy is not high)
(5) 3D TOF (more used for sweeper top LIDAR bias navigation and two-dimensional plane modeling, auxiliary obstacle avoidance)
In addition to infrared tube obstacle avoidance and infrared PSD sensing obstacle avoidance, let's talk about the other 4 mainstream
obstacle avoidance technologies on the market at present.
Visible-based monocular cameras and binocular cameras
3D TOF based on invisible-infrared light and 3D structured light-two-line laser
Monocular obstacle avoidance, the shape of the obstacle is recorded by a single camera, and the algorithm recognizes the size and
distance of the current object, but it is impossible to judge the object that does not exist in the database. In addition to the unique
advantages of visual recognition, the binocular obstacle avoidance system uses the method of parallel shooting to identify the
difference in the images recorded by the two cameras through the information recorded by the dual cameras, which can provide
more data for identifying the type of obstacle. In general, from the perspective of practicality, sweeping robots using binocular
recognition technology can better adapt to the working environment, and at the same time, in terms of obstacle avoidance performance,
they can also identify more kinds of obstacles, and then choose the response through databases and algorithms.
TOF technology can roughly understand its principle from its name - time of flight, flight time, mainly by emitting light, light encounters
impenetrable objects will be reflected, by recording the time when the reflected light reaches the receiver, because the speed of light
and light wavelength are known, to calculate the distance between the light source and the object. This technology uses more
single-point TOF scheme, which is used for LIDAR on the top of the sweeper to obtain environmental depth information by turning
to scanning to assist obstacle avoidance, but it is more biased towards the final effect of navigation and two-dimensional plane modeling.
3D structured light scheme, the principle is to use infrared light source, the emitted light through a certain code projection on the object,
these patterns through the surface of the object reflected back, with the distance of the object will occur different deformation, the image
sensor will take the deformed pattern. Then, by calculating the amount of distortion of each pixel in the captured pattern, the corresponding
parallax is obtained, and then the depth value is obtained. However, the ranging range of the structured light scheme is affected by
the spot pattern, and the obstacle avoidance range is small. In addition, the performance in the strong light environment is poor, and it is
easily affected by light.
3D structured light
Two crossed line lasers work together to create a stereoscopic map by moving the sweeper. The advantage of dual-line laser obstacle
avoidance is that the ability to resist ambient light interference is relatively strong; In the face of dark light environment, when the light
becomes dark, the objects in the camera are basically invisible, and under the role of infrared light, the sweeper can still clearly see the
road ahead, in fact, it can be understood that the sweeping robot has been holding a flashlight to illuminate the front, and the light out
of this flashlight, our human eyes cannot see. For strong light environments, the infrared light in the ambient light is relatively strong,
and the infrared light in the machine is required to be strong enough to weaken the interference in the ambient light.
The line laser used in the sweeper is not continuously output, but works at a certain frequency. In the same unit of time, as long as the
line laser module works more frequently, the sweeper can get more information, thereby helping the sweeper judge obstacles faster,
so whether it is facing sudden people, pets or other debris, the sweeper can judge obstacles and avoid them.
Two-line laser
LEMON Photonics Technologies Co., Ltd. is a high-tech enterprise jointly founded by the professional team of returnees in the
semiconductor laser industry, venture capital and industrial investment institutions, committed to the industrialization of high-end
semiconductor laser chips and laser projection light source modules. The company has the world's leading all-round semiconductor
laser technology and process, including the exclusive new generation of laser chip technology. Based on independent core technology,
the company is committed to creating a one-stop supplier of semiconductor laser chips and light source solutions for core light source
components.
