Zircon Applies Video Analytics for RSSB Funded Feasibility Study

Zircon Applies Video Analytics for RSSB Funded Feasibility Study

Zircon Applies Video Analytics for RSSB Remote Condition Monitoring Feasibility Study

Languages: Python
Zircon were awarded funding from the RSSB, who play a major role in improving and regulating the safety of the railway industry, to undertake a feasibility study as part of the RCM program. RSSB had identified ten unresolved challenges that prevent the full implementation of remote condition monitoring on the railway.

Zircon was responsible for investigating a system that would utilise the forward facing CCTV cameras on trains to detect unauthorised human presence within the boundary fence, whilst filtering out authorised human presence. Additionally, our client expressed interest in utilising the train CCTV feeds to identify any significant movement of the objects within the vicinity of the track that may begin to block the passage of rolling stock as well as check the visibility of essential rail side infrastructures.

As a feasibility study, we were called on to examine a high number of variables and potential design problems. For example, the system will be expected to function accurately in a variety of lighting and weather conditions, distinguish the difference between rail workers in high-vis clothing and members of the public who have strayed inside the boundary of the track, and do all this from a moving platform whilst providing accurate position information.

With regards to detecting human incursion, the focus of the study was on our ability to detect human presence in a multitude of different lighting conditions, similar to those experienced on the daily journey of trains, and the ability to identify and differentiate authorised personnel from unauthorised human presence. During testing we found that the probability of detecting people over a variety of contrast and lighting conditions was surprisingly high, and following the inclusion of pre-configurable Hi-Vis parameters the identification of authorised presence was achievable.

Unlike detecting human presence, the ability to detect the movement of objects was slightly more complex and unfamiliar. As there is no way to define what the system should be searching for the techniques used in previous object detection projects we had done would not necessarily be suitable. Our solution for this problem was to generate an ‘interest map’ of the journey landscape in order to compare CCTV footage from train journeys in order to monitor and identify changes in landscape. Our tests found that generating the map itself was easy, however monitoring changes against the map was less so but still possible.

Currently the entirety of the study has been completed off-train without representative CCTV footage, and is awaiting the opportunity for a train mounted trial in order to try and identify real life problems. It is our hope that once this system has been proven, the technology could be utilised to forewarn of further issues which have an impact on the levels of safety and performance of the railway.

Similar Projects

Lets Talk Software

Looking for a team to support your next Rail software venture?  Zircon is there to help you ensure project success, contact the team today.

Zircon Assists Synoptix Ltd with Object Detection Feasibility Study

Zircon Assists Synoptix Ltd with Object Detection Feasibility Study

Zircon assists Synoptix Ltd with a feasibility study to detect objects within video footage and images.

Operating Systems: Windows
Languages: Python
Other: Blob Detection,  Casscade Classifier,  Colour Comparison,  Contour Matching,  Edge Detection,
Histogram Of Oriented Gradients,  Image Comparison,  Sliding Windows,  Subcomponent Classifier,  SURF Descriptors

 

In the past much of our work has been on projects related to the transport industry. What’s more, we usually have a pretty clear idea what the solution is going to look like. But that’s not always the case. We recently completed a project related to defence and security (we can’t tell you much more than that as…well, we just can’t!). And although the requirements were clearly defined the end goal was merely to test the feasibility of an idea and outline the most promising ways forward.
Background To The Project

Our immediate client was Synoptix Ltd, a provider of bespoke, high level Systems and Safety Engineering solutions for a wide range of different industrial sectors. They were working on a project for the Defence Science and Technology Laboratory (DSTL), the centre of scientific excellence for the UK Ministry of Defence (MoD). Its stated aim is “to maximise the impact of science and technology for the defence and security of the UK.”

Synoptix was collaborating with DSTL in a project that involved the development of object detection software. As Synoptix had no in-house software capability they were looking to partner with a suitable third party.

The Head of Sales for Synoptix was familiar with Zircon and knew we had recently been working on a project with similarities to DSTL’s requirements but within the rail industry environment. So we were asked to submit a proposal, which was accepted.

“I’d definitely endorse them with regard to their quality of communication, the quality of documentation they produced.”

Ben Durant

Operations Manager, Synoptix

The Challenge

The general idea, in layman’s terms, was to get a better understanding of how far one could go in developing an effective object detection solution using Commercial Off The Shelf (COTS) software and libraries. The task was to see how difficult it would be to get a conceptual solution through to Technology Readiness Level (TRL) 3 via the development of algorithms on an open-source image processing platform.

More specifically, the software had to be capable of detecting the presence of specific pre-determined objects within a digital image (either a still photograph or the individual frame of a video). The program had to be capable of picking out the shape of an object from whatever else was going on around it, in a variety of different conditions and with a range of different image formats. What’s more, with video footage it had to be capable of tracking the shape through multiple frames. Being even more specific the system was required to:

  • Recognise objects by referring to still images stored in the system
  • Categorise objects detected
  • Sub-categories objects where possible
  • Perform recognition on video stream shot at any viewing angle; any compression; any frame rate; any illumination; with any level of clutter; with any confusers or obscuration; and at any viewing range, from 5 to 100 metres
  • Operate in visible spectrum (RGB and grey scale)
  • Track detected objects through scene
  • Create a timestamped list of when objects appear in video

So it was a very tall order!

However, the end deliverable was not a fully functioning, tested and effective system. It was to see how far one could go down that road using COTS software and libraries; then present those findings, alongside Synoptix, to the DSTL.

How Things Progressed

The primary COTS component we proposed was OpenCV, which is a commonly used open source computer vision library. This provided much of the actual algorithmic functionality utilised by the object detection software we developed. Because the project was essentially an R&D exercise we selected Python 2.7.5 as the development language as we find code can be written very quickly in Python. This also meant that the software could easily be ported to C++, and better performance achieved, should a release version be required.

Ben Durant, Operations Manager at Synoptix, explains that it proved a very hard task. “It was very difficult, much more so than the project they had undertaken on the railways. The system had to work under so many different conditions and situations that they were confronted with a variety of awkward challenges.”

“They tackled these in a variety of different ways. They tried doing co-ordinate point detection, but that didn’t work very well. They tried turning those shapes into shadows, that didn’t work very well either. Then they ended up using a couple of filters for image processing, and these were reasonably good at detecting the shapes we were asked to look for. They got it to work at the testing stage, but it was a long way off a finished product.”

“I found Zircon very easy to work with, They were very cooperative, very helpful throughout that process”

Ben Durant

Operations Manager, Synoptix

Conclusion

Ben makes the point that “The key deliverable was not a product that was ready to be deployed. It was the production of a joint report, prepared in partnership between us and them, to outline the lessons learned during the exercise.” He adds that “They were very good at reporting what was going on and keeping us in the loop. This was essential, because the exercise was essentially an information gathering one, so good communication was very important.”

Explaining this point in more detail he comments that “I worked closely with the main developer at Zircon and we had to thoroughly analyse what was going on and where the technical difficulties lay. We had to make a lot of decisions on what findings were most meaningful and how best to frame them for presentation to the MOD.”

He concludes that “I found Zircon very easy to work with. The process involved a lot of discussion and negotiation. They were very cooperative, very helpful throughout that process. I’d definitely endorse them with regard to their quality of communication, the quality of documentation they
produced. In the end we successfully delivered a presentation to the MOD, albeit with limited functionality, that met their specification. It proved to be a very valuable piece of research.”

His final comment was “If we get some further work from the MOD on this I’ll be teaming up with Zircon, no question.”

Similar Projects

No Results Found

The page you requested could not be found. Try refining your search, or use the navigation above to locate the post.

Lets Talk Software

Looking for a team to support your next software venture?  Zircon is there to help you ensure project success, contact the team today.

Feasibility Study into Monitoring Overhead Lines Aims to Avoid Expensive Equipment Failure

Feasibility Study into Monitoring Overhead Lines Aims to Avoid Expensive Equipment Failure

Feasibility Study into Monitoring Overhead Lines Aims to Avoid Expensive Equipment Failure

Languages:

Python

Technologies:

OpenCV

Configuration Management:

git

Other:

BS EN 50128

 

Zircons client provides manufacturers of passenger trains across the world with high quality CCTV systems. As part of a project where Zircon were assisting the client with the development of their next generation CCTV system, camera’s will be mounted to the roof of trains in order to provide real-time footage of the pantograph and overhead lines.

The part of the pantograph that makes direct contact with the overhead line is covered with a carbon shoe that conducts electricity whilst working as a form of lubricant. Due to the friction from continuous contact with the overhead lines this carbon covering wears down over time. In order to wear the surface down in an even fashion the line tracks across the pantograph from left to right in a zig zag motion.

The consequences of damage to the overhead line equipment are both expensive and far-reaching. In order to provide the end users of this new CCTV system with additional value, it was our clients desire to utilise this footage to monitor the pantograph/overhead line interface to quickly identify areas where the path of the cable is incorrect or arching is present.

In order to monitor the path of the overhead line across the pantograph Zircon carried out a feasibility study to identify the position of the point of intersection between the two surfaces in footage from the client’s camera. Upon locating the intersection, a record is made of its position, velocity and acceleration as the train travels along the track. If the software identifies that the position of the intersection has remained constant for longer than a predefined acceptable period or the velocity and/or acceleration exceeds the defined thresholds, the GPS position is recorded and an alert is raised.

Identifying occasions of arching was a simpler task. The sudden burst of light generated by the spark causes the camera to experience a sudden brightness overload, especially in darker environments, which can easily be identified by the software. As with the tracking of the overhead lines if the software detects arching the GPS position is recorded and an alert is registered. All of the alerts registered during a journey are processed upon the completion of the train’s journey and the relevant line maintainers are informed.

This system was successfully delivered to the client as part of the CCTV development project, and Zircon has subsequently provided the client with support services as they introduced and modified the system for new clients.

Similar Projects

Lets Talk Software

Looking for a team to support your next Rail software venture?  Zircon is there to help you ensure project success, contact the team today.