Computer systems and networks
Note! At the time of this writing, the master's work has not yet been completed. Estimated completion date: May 2021. The full text of the work and materials on the topic can be obtained from the author or his manager after that date.
Every day people move around the city, millions of cars around the world drive around cities and countries. And all this thanks to sidewalks, roads, highways. Since ancient times, the road has been a communication route for the movement of people, livestock, and transport. The most ancient roads date back to the 4th millennium BC [1]. A lot of time has passed since then, construction technologies have changed, but the load on the coating has also increased. In this regard, it is very important to monitor the condition of the road surface. The constant increase in the length of roads requires the involvement of a large number of specialists and equipment, which entails an increase in costs. But the risk of an inspection error also increases. In areas with severe meteorological conditions, road surveys must be carried out more often and more thoroughly, since in addition to the load from transport, the influence of temperature, precipitation, and ground movement affects. Modern technologies come to the rescue in such difficult situations.
A high-quality road is based on the so-called "road pie" or road clothes. "Road cake" is a multi-layer base, which consists of soil, sand and rubble. This base allows the road surface to last longer [2]. All stages of road construction have their own technology and purpose. The thickness of the sand and crushed stone layer depends on the expected loads and is calculated for a specific task. "Pie" is necessary to prevent premature destruction of the surface, road traffic and provides a speed limit and safe movement. Figure 1 shows the most common road pie structure.
Figure 1 - The structure of the "road pie"
Such a structure consists of a subgrade in which all communications are laid. Then comes the cushioning sand layer, which minimizes the likelihood of deep damage. Geotextiles can be placed between these two layers for separation. It acts as a drainage so that the road does not wash away and prevents the mixing of sand and soil. Then again comes the separator layer, geotextile or geogrid. A layer of crushed stone with a bitumen emulsion is laid on top. Thus, a strong foundation for the road surface is obtained, which can ensure the integrity, evenness and durability of the road [3].
All stages of work, as well as materials used for construction, are clearly regulated by the relevant GOSTs. Thus, GOST R 58349-2019 “Automobile roads for general use. Travel clothing. Methods for measuring the thickness of pavement layers ”applies to the construction, reconstruction, overhaul of pavements and establishes requirements for methods for measuring the thickness of base layers [4]. Accordingly, the equipment performing such work has certain parameters and characteristics. It also requires sufficient accuracy to work.
But even the unquestioning fulfillment of all the rules and compliance with standards does not guarantee the safety of the coating. Over time, various kinds of defects appear on the road surface. Pavement defects are deviations of the geometric parameters, texture and structure of the pavement from the standards. They are divided into the following types, depending on the nature, location and size [5]:
Any driver is familiar with the situation of potholes in the spring, after the snow melts. This occurs both from the effect of temperature and load on the pavement, and from a possible violation of the technology of laying the asphalt concrete pavement. Everyday use, heavy loads, ground water, ground movement all lead to destruction of the asphalt. To identify and further forecast the development of various defects, today, two groups of examination are used [6]:
Both of these surveys are regulated by GOST 33388-2015 “Public Roads. Requirements for diagnostics and certification” [7]. To determine various parameters, such GOSTs as GOST R 56925-2016 “Automobile roads and airfields. Methods for measuring unevenness of substrates and coatings” [8] and GOST 32825-2014 “Automobile roads for general use. Road surfaces. Methods for measuring the geometric dimensions of damage (Reissue)” [9]. The first standard applies to methods for measuring the unevenness of the road surface (as well as airfields) during construction and operation. The second standard regulates methods for measuring the geometric dimensions of road surface damage that affect road safety during its operation.
The use of cameras and computers has a positive effect on both financial costs and the quality of work. But the role of the human factor is still great. To solve this problem, it is proposed to create a system for analyzing the surface of an object, in this case a road surface. The main purpose of such a system is to determine the places of deformation of the road surface without human participation in this process, and in the future, the creation of completely independent mobile complexes for monitoring the condition of the roads. This system is based on the use of a laser projector, camera and data processing unit. The impetus for work on this topic was the development in 2013 by a group of specialists from Sichuan University [10]. Their idea was to create an LED bike projector that projects a square grid onto the road ahead. If any irregularities occur on the path, the strict geometry of the mesh changes, as shown in Figures 2 and 3.
Figure 2 - The principle of operation of Lumigrids
Figure 3 - Changing the mesh geometry when irregularities occur
As you can see from the figures above, the rays of the projector, when they hit any unevenness, change the state of the grid lines. Thus, the cyclist can notice in advance the presence of an obstacle on the road and adjust his trajectory or stop. And although the device did not go into the series, a similar solution, with some modifications, can be applied in the automotive sector.
To implement the system, it is proposed to use a laser projector to project a grid onto the road surface and a camera that will continuously shoot and send data to a data processing device. Such a device is a compact computer that analyzes the received frames. To process the resulting image and determine what is happening on the coating, a script written in Python using the OpenCV computer vision algorithms library is used [11].
It is proposed to use a Raspberry Pi 4 single-board computer as a platform for a data processing device [12]. This is one of the latest computer models.
As for the characteristics, there is a four-core 64-bit processor with a frequency of 1.5 GHz on board. You can choose from a different amount of RAM: 2, 4 or 8 Gb of LPDDR4 standard. The present USB 2.0 and 3.0 connectors will allow you to connect all the equipment you need for work: mouse, keyboard, camera. With two micro-HDMI ports, you can connect two monitors to keep track of what is happening. The board also has dual-band Wi-Fi and an Ethernet connector, which allows you to organize the transfer of processed data to the server for further work with them.
GPS technology is used to determine the location of the vehicle.
Any car can serve as a base for such a system. Outside it is installed a projector and a camera aimed at the road surface, and a computer is installed in the cabin.
Additionally, to obtain accurate data, it is possible to use a gyroscope and an accelerometer to fix the vehicle tilt angles and acceleration along the X/Y/Z axes. Such data will be able to provide a complete picture of what is happening at the place of detection of a road defect.
The algorithm of the system operation consists in continuous shooting, projected onto the road surface of the grid, by the camera while the car is moving. The grid has a square shape, the number of lines can be changed depending on the task. Figure 4 shows an example of a 10 by 10 grid.
Figure 4 - Example of a 10 by 10 grid
On any irregularities, the mesh geometry will certainly change. These changes are recorded by the camera and sent to the data analysis and processing device. Further, the resulting frame is passed through a Python script and using special computer vision algorithms, the position of the lines and their intersection angles are determined. As a result, based on the data obtained, a conclusion is made about the condition of the road surface: the presence of a depression or a hill. A request for GPS coordinates is made. Then, the received answer is written to a file with reference to GPS coordinates.
In the presence of additionally installed accelerometer and gyroscope, the sensors are polled and data are generated on the tilt angles at the current moment, as well as the latest data on vertical accelerations to complement the picture of the state of the road surface. Figure 5 shows a visual representation of the algorithm.
Thanks to the use of computer vision, the work efficiency and the area of the surveyed road are increased. The ease of installation of such a system allows it to be installed in any car and used in various fields.
Figure 5 - Visual representation of the work algorithm (animation: 11 frames, 5 repetition cycles, 50.8 kilobytes)
The proposed system will make it possible to quickly and accurately control the evenness of the road surface, optimize the financial costs for inspection of the surface and the determination of defects. Also identify potential locations of road defects and take timely action to eliminate them. And thanks to the flexible settings of the system, it can be adapted for use in other areas that require coverage analysis.
Resume
Abstract