The following is a proposal of using computer vision technique to address the issue of Rock Bolt Failure.
The purpose of this report is to propose an approach for detecting rock bolt failure by using a low cost Raspberry Pi based computer vision system which would allow operators to determine whether rock bolt is at the point of failure or whether the underground mine is at risk of collapse. This proposed solution would allow mining companies to effectively monitor the status of their mine by placing a small Linux powered computer device that would be connected to a video camera which would handle the computer vision algorithm for detecting rock bolt failures. This report is divided in to introduction – which would introduce the problem statement, background – which would briefly explain the existing solution, procedure – the working principle of the proposed solution and a conclusion
Rock bolt has been used by mining industry to stabilize rock excavation used for tunnels construction in mines. Rock bolts are used to distribute the load of earth over a mesh network of rock bolts. Rock bolt are specially designed to clamp together the rock mass before it can move enough to fail by loosening. It can also be used to prevent rock fall.
In tunnel excavation, rock bolts are specially designed to show the movement of rock masses exerting on rock bolts. From the challenge description, the reinforcing effect is achieved through some form of friction or grout-based coupling between the rock and the bolt, established at the time of bolt installation, refer to Figure. This schematic shows the case of a grouted bolt type with a single anchor; other configurations and bolt types also exist. As mining progresses, the development of this effect at the operating depths of today’s mines requires bolts capable of sustaining substantial design loads as well as design deformations. Thus for example, a 3m long rock bolt may be required to deform in excess of 100mm and be subject to both tensile and shear loads approaching 300kN, including failure. Failure here is defined in its most basic sense: bolt breakage anywhere along its active length, including the threaded section at the bolt plate and nut.
Figure 1 – layout of rock bolt
Some companies offer rock bolts that have indicators at the end of bolt plates which could be used by the operators to measure the length of bolt, Bolt Plate Load P, Bolt Plate Displacement D, or something similar as shown above.
The proposed solution is to use Linux-powered computer incorporating high resolution camera using computer vision to detect and monitor the bolt failure by measuring the length of bolt by monitoring the length of reinforced bar from the end of its faceplate. Most operators and companies color code the reinforced bar to tell the level of deformation by direct observation.
With the advent of small scaled microcomputer such as raspberry pi, it is easy to create a small scale system capable of processing high quantity data. As shown below, the proposed idea is to connect raspberry pi with a high definition camera and placing the system in an area where bolts location would be easily visible. The raspberry pi will be programmed to take the initial screenshot of the bolts with reinforced bar protruding from the faceplate when the bolts were placed on the rock mass. The system would use this initial image to compare the difference of the length of the reinforced bar in the new image. This would allow the operator to measure the stress and strain of earth over the rock bolt and pin point which rock bolt requires replacement. Moreover, the system would be able to send the information via Zigbee Protocol using XBEE shield which can be bought off the shelves. One possible system design is shown in figure 2 below.
Figure 2 – Propose Linux-powered rock bolt detector
The details of each component are listed in the description below. It should be noted that alternative to these components also exist, however, these components were selected based on its easy accessibility and low cost.
As described by Raspberry Pi foundation, “The Raspberry Pi is a credit-card sized computer that plugs into your TV and a keyboard. It is a capable little computer which can be used in electronics projects, and for many of the things that your desktop PC does, like spreadsheets, word-processing and games. It also plays high-definition video. We want to see it being used by kids all over the world to learn how computers work, how to manipulate the electronic world around them, and how to program. (Raspberry Pi Foundation, 2014)”
Figure 3 – Raspberry pi – Linux based computer
Moreover, “Raspberry Pi delivers greater flexibility for professional embedded applications. This new Raspberry Pi board has been developed specifically for engineers in the professional market to create their own embedded system. The Compute Module fits into a standard DDR2 SODIMM socket and provides the same basic features of a standard Raspberry Pi, including a Broadcom BCM2835 processor and 512 Mbyte of RAM. It replaces the SD card with an onboard 4 Gbyte eMMC Flash device, and integrates everything on to a compact 67.6 x 30 mm board. The Compute Module is initially available as part of a development kit, bundled together with the IO Development Board, which brings out all of the IO connectivity to form a prototyping platform for electronics design engineers. (Raspberry Pi Foundation)”
As per the specification, it is clear that Raspberry Pi could be used to compute sophisticated calculation required by computer vision algorithm to work effectively.
High Resolution Camera:
The purpose of using high resolution camera is to allow more pixels to gives off information regarding an image. In case of low resolution camera, the number of pixels need to describe the image are small which make the system unintelligible to identify the smallest difference in the image. Since the system is being designed to detect and identify the smallest difference of the rock bolt therefore high definition camera with resolution of 720p or 1080p would prove to be very efficient for the implementation of this system.
Figure 4 – TeckNet® C016 USB HD 720P Webcam, 5 Megapixels, 5G Lenses, USB Microphone & 6 LED
One possible low cost and portable HD camera is TeckNet C016. This can be bought from Amazon or any local retail store. This system has the ability to reach resolution of 720p and would cost around $20.
This is the very popular 2.4GHz XBee XBP24-AWI-001 module from Digi. The Pro series have the same pinout and command set of the basic series with an increase output power of 60mW! These modules take the 802.15.4 stack (the basis for Zigbee) and wrap it into a simple to use serial command set. These modules allow a very reliable and simple communication between microcontrollers, computers, systems, really anything with a serial port! Point to point and multi-point networks are supported.
Figure 5 – XBEE Pro Shield
XBee is becoming an industrial standard used for communicating between devices. The transmitter and receivers of XBee are easy to connect and the protocols are defined by the hardware itself. The operator would be able to monitor the data over the internet if they choose to do so. However, if the operator chooses to send the data on a private channel then that can be done too.
The procedure of implementing the rock bolt detection process is to first take a snapshot of the area with newly installed rock bolt. The rock bolt we propose will have color indicator at the end of the reinforced bar which would indicate the level of deformation. This color indication will be detected by the Linux-powered system that we have proposed. Once the snapshot of the area is taken, the system would then calculate and count the color distance based on the initial observation. The system would use this initial image as a control to detect any changes based on measuring the initial and current images. If a change in color is detected then the system would send an automated alert signal to the operator via XBee module which would also carry information regarding the location of the rock bolt where deformation has been detected. This will pinpoint the problem to the operator for further analysis.
Figure 6 – Proposed rock bolt with color indicator reinforced bar
Figure 7 shows a proposed method of detecting the rock bolts. As seen from the figure, the system will detect the changes and based on computer vision algorithm ported in raspberry pi, the system will be able to identify the changes based on the analysis.
Figure 7 – Proposed system in action
The following flowchart will detail how the code should work. Programmer is free to change the outcome of the program since coding is left up to the programmer.
Figure 8 – Flowchart of the system program
Cost of Solution:
Hardware Price (USD)
Raspberry pi 39.95
TeckNet® C016 USB HD 720P Webcam 16.45
XBee 1mW Wire Antenna 24.95
System Housing 5
Note: The cost does not include labour
The system proposed above is modular in its technicality. This system will remove the need of human operator from constantly monitoring the status of rock bolt which is a time consuming process. The propose solution is capable of identifying multiple status of multiple rock bolt to a single operator via XBee wireless devices which utilizes Zigbee protocols to send the data to the main computer for the operator to monitor and analyze the deformation process.
The advantage of the propose system is to save time and money while eliminating errors made by human observers. This system could prove to be highly efficient and mobile enough to be place at any location of the mine given that the system’s camera has the view of the rock bolts. The system’s modularity would allow the programmer to improve the code which would allow easy upgrades.
Original paper published in ResearchGate [link]