Friday, July 27, 2012

A good month for the motors

One of the biggest markets for vision systems is in the automotive industry, where systems are not only deployed to perform numerous quality inspection tasks on the production line, but increasingly, within the automobiles themselves to provide additional levels of security for both drivers and pedestrians.

So I was pleased to hear that this month has proved a particularly splendid one for the industry. According to a monthly sales forecast developed by J.D. Power and LMC Automotive, July’s new-vehicle retail sales are expected to post the second strongest year-over-year growth rates during the past 12 months.

"Retail sales got off to a fast start in July, and while they've slowed down a bit as the month has progressed, through the first 16 selling days, they're still up 15.1 percent, compared to July 2011," said John Humphrey, senior vice president of global automotive operations at J.D. Power and Associates.

All major segments are expected to show year-over-year sales gains in July, with the exception of the midsize crossover utility vehicle segment. That aside, the sub-compact conventional, midsize conventional and compact conventional segments are projected to show year-over-year increases of 28 per cent or more.

The report goes onto say that through the first half of this year, North American light-vehicle production volume increased 22 per cent, compared with the same period in 2011. More than 1.4 million additional vehicles have been built in the first six months this year, relative to the first half of 2011, with inventory replenishment and stronger demand in the first quarter being the main factor for the higher production volume.

Honda and Toyota's production in the first half this year is up 75 per cent and 65 per cent, respectively, as both manufacturers continue to recover from the impact of the Japanese tsunami. What is more, US manufacturing growth is outperforming the rest of North America, with a 26 per cent year-to-date increase. Production in Mexico has increased 13 per cent and Canadian manufacturing up 19 per cent.

While this is clearly good news for the auto makers themselves, it bodes well for many of us involved in the vision systems business who supply vision systems to them.

You can find more details of the report here.

Thursday, July 26, 2012

Walk this way

As the editor of Vision Systems Design, I get to fly around a lot during the course of my work. But during my travels, there is little time to observe the behavior of the individuals in the various countries that I visit.

However, that's definitely something in which Dr. Rajshree Mootanah, the Director of the Medical Engineering Research Group at Anglia Ruskin University (Chelmsford, Essex) in the UK is interested.

You see, the learned doctor is currently involved in a project to measure the gait of individuals with the aim of using the data he acquires as a measure by which the joint functions of those who have just undergone hip or knee surgery can be assessed.

Now a lot of work in this area has already been carried out by researchers at the Hospital for Special Surgery in New York, one of the leading hospitals for orthopedics in the US. But the trouble is that the database of normal gaits from that hospital was captured, naturally enough, from New Yorkers.

Dr. Mootanah believes that the people in the county of Essex in the UK are likely to have a different gait to New Yorkers, and that his research project to establish a local database will allow more accurate testing and analysis of UK patients.


"The only database we have is of the New York population and we believe there may be slight but still significant differences to the way our local population walks due to the different racial make-up of the two groups," Dr. Mootanah says.

For that reason, his team is now on the lookout for volunteers, aged 18 or over, who are able to walk without impediment. The volunteers will have the force of their steps measured by special pressure plates embedded in the floor while their gait will be recorded by a 3-D motion capture system.

The results from Mootanah's research will certainly be interesting and may be more useful that he realizes. According to the New York Times, five studies presented at the Alzheimer’s Association International Conference in Vancouver this month provided striking evidence that when a person's walk gets slower or becomes more variable or less controlled, his cognitive function is also suffering.

So not only could the database created by Mootanah be useful as a means to evaluate patients who have undergone surgery, it might also provide a valuable tool to other researchers who might also use it to evaluate the cognitive functions of individuals.

Reference: Footprints to Cognitive Decline and Alzheimer's Are seen in Gait, The New York Times, July 16, 2012.

Friday, July 20, 2012

3-D vision beats the clock

Whether you are short or tall, skinny or overweight, at one time in your life you have most likely used a pedestrian crosswalk to help you cross a busy street.

Using a crosswalk is undoubtedly a much safer bet than simply choosing your own spot to cross the road and risk being hit by a moving vehicle. Nevertheless, using such crosswalks can be rather an intimidating affair.

That's especially true of new those new fangled crossings commonly seen in Florida that incorporate a countdown timer to help pedestrians know just how long they have got to cross the road.

While they undoubtedly put the spring back in the step of many pedestrians who can then visualize just how long they have got before a hoard of Fords start hurtling towards them, they don't do much for the blood pressure of disabled or elderly people who may not be able to increase their velocity to beat the countdown.

Now, thanks to the help of a 3-D vision system, engineers at Migma (Walpole, MA) have come up with an interesting solution to the problem that has already been tested out at certain crosswalks with great effectiveness.

The system itself makes use of a stereo vision-based infra-red camera that can detect pedestrians during the day and at night. The output from the camera is hooked up to a computer that runs pedestrian detection algorithms that extract 3-D features of the human figures on the crossing from any other images that are present.

Detecting the pedestrians by using such a computer-based vision system enables the timing of the lights on the crosswalk to be controlled by their presence, a much more sensible approach than giving them an ultimatum which they may not be able to meet.


But is this really such a terrific idea? After all, how would the system be able to accommodate the behavior of those rowdy ne'er-do-wells who were intent on slowing traffic to a standstill by continuously walking back and forth across the crossing, perhaps in a protest to government cutbacks?

I think I have an idea. Perhaps what the system needs is an intelligent neural network-based back end that can "learn" to identify the behavior of such folks and then alert the authorities to take the appropriate action if its spots them acting in an untoward fashion.

Mind you, by the time you add up the cost of doing that, such a system might be just too gosh-darned expensive to make it worth installing. Back to the drawing board!

Wednesday, July 18, 2012

Motoring on Linux

Years ago, an old friend of mine remarked that if his car was controlled by the Windows operating system, it would take him ages to get anywhere.

He reached this conclusion after spending many mornings downloading updates to his Windows-based system and then rebooting it before he could start work. From this, he surmised that if such Windows software was used to control the electronic systems in his car, he would spend an equal amount of time in the driver's seat waiting for updates before he could even put his key in the ignition.

My friend, of course, was a complete and utter technical Luddite -- one of those chaps that would rather have lived in the age of steam where at least he would have some vague notion about how large amounts of very hot water could be used to propel vehicles along a track.

Fortunately, however, the goods folks at the Massachusetts Institute of Technology (MIT; Cambridge, MA, USA) are not like my friend at all. They are always coming up with new and interesting ways to enhance the performance and safety of vehicles using computer systems and software.

Just this week, for example, two of the folks there -- Sterling Anderson and Karl Iagnemma -- announced that they had developed a semi-autonomous safety system that can help drivers of vehicles avoid colliding into objects in their path.

The system uses an onboard camera and laser rangefinder to identify hazards in a vehicle's environment. Then, system software analyzes the data and identifies safe zones that the vehicle can travel in. The system allows a driver to control the vehicle, but takes control of the wheel when the driver is about to exit a safe zone, thereby avoiding objects in the vehicle's path. 

Anderson, who has been testing the system in Michigan since last September, has observed an interesting phenomenon after several folks tried out the system on a course.

Notably, those who trusted the system tended to perform better than those who didn't. For instance, when asked to hold the wheel straight, even in the face of a possible collision with an object, drivers who trusted the system to take control and avoid the object drove more quickly and confidently through the course than those who were wary of the system.

So far, the team has run more than 1,200 trials of the system, with few collisions. Most of these occurred when glitches in the vehicle's camera failed to identify an obstacle.

As for my Luddite friend, I don't know what he'd make of it all. I know one thing though. He'd be tickled pink that the researchers chose to perform experimental testing of the software they developed using a PC running the Linux operating system!

Editor's note: Interested readers can find a technical article entitled "Constraint-Based Planning and Control for Safe, Semi-Autonomous Operation of Vehicles" by MIT's Sterling Anderson, Sisir Karumanchi and Karl Iagnemma here.

Friday, July 13, 2012

Rock and roll vision


Our erstwhile European Editor is always sniffing around the Interweb to see if he can discover any snippets of information that might be useful to the readers of Vision Systems Design. I'm rather glad he is, because that's, in part, what I pay him rather exorbitant sums of money (by my standards, not his, of course!) to do.

On many occasions during his surfing activities, he stumbles across news articles written by folks that have attempted to popularize the work of academics working in the field. Unfortunately, in doing so, many of the writers of such pieces generalize the work of the researchers such that the point of the work becomes almost incomprehensible.

Fortunately, it's often the case that the folks that write such pieces take the time to provide hypertext links to direct the reader to a specific technical paper that the engineers have published in learned journals.

Sadly, though, this hasn't proved much use to our European Editor, who has discovered that -- upon reaching the sites of such learned journals -- he is required to pay a certain sum of money to read any further about the system or software that has been designed and developed.

Somewhat frustrated by this turn of events, the conniving old European Editor has figured out a way around the problem. That's right. Once realizing that a piece of work of interest has been developed that he thinks that you, our reader, might be interested in, he then performs a quick search for the author of the piece on the Interweb.

Once he has located the author's home page, of course, he inevitably discovers that the researcher -- proud to have had his work accepted for publication by the learned journal -- has posted a copy of it as a PDF on his own personal web site! That’s the place, of course, where the wily old hack discovers what is really under the hood of the technology and how relevant the work might be.

But the whole affair worries me just a little. Should such information really be available for free?

In one of my last blogs, I revealed how a chap who goes by the name of Pablo Caicedo has (apparently illegally) uploaded a 442-page volume entitled "Image Processing and Mathematical Morphology: Fundamentals and Application" by Professor Frank Shih from the New Jersey Institute of Technology onto a web site called scribd.com. Needless to say, the learned professor was less than impressed when we pointed this out to him.

But are the academic researchers in the vision industry not guilty of the same form of infringement too, when they choose to publish PDF of the articles owned by the academic publishers on their own websites?

It looks to me as if the paid for publishing industry is going down the chute, following in the footsteps of the recording and movie businesses. Perhaps that's why so many academics are now going on tour and on television like their rock and roll counterparts.

Wednesday, July 11, 2012

Cameras in cars help fight planned accidents

It's not very often that I take the trouble to read any press releases that are issued by insurance companies. My days, as you can imagine, are filled with attempting to interpret in an intelligent fashion much of the information that is (or to be more accurate, isn't) issued by the good folks in the vision systems design industry.

But this week, I did spot one story that really caught my attention. A story issued by none other than InsuranceQuotesUSA.com, part of the Insurance Quotes USA network, a self-proclaimed all-in-one stop for tips, information, and quotes on car insurance.

You see, this outfit is now recommending that all drivers purchase and install "dashboard cameras" to fight against what is known in the trade as "planned" automobile accidents. Having been recently involved in a minor incident with my own trusty Hyundai, I was intrigued to find out more about what constitutes such a planned accident and how the cameras could help.

Apparently, what I discovered was that these planned accidents can occur in a couple of ways, resulting in the victims who have suffered from them having their car insurance premiums jacked up.

The first planned accident scenario -- called Intentional Backing Up -- involves two vehicles stopped on an uphill road. The criminal in the front vehicle puts his vehicle into neutral and his car falls back onto the victim behind him. The criminal driver then claims that the driver from the vehicle behind has failed to brake and stop properly.

The second accident scenario -- called Intentional Falling -- involves a pedestrian and a vehicle. When the victim's vehicle is stopped in front of a crosswalk, the pedestrian falls over intentionally and then acts as if the victim’s vehicle has failed to brake and stop properly.


Victims are often given the "opportunity" to resolve the accident without going through the insurance agency and claim process by offering cash compensation up front.

By having a dashboard camera in the car, however, such accidents can be recorded and victims can produce the evidence of the incident to the police. With that in mind, InsuranceQuotesUSA now recommends that drivers equip their vehicles with a 30 fps camera with a 5Mpixel color sensor and 16GB of memory that can capture 1-2 hours of data that is time and date stamped.

Clearly, this is great news for camera manufacturers across the globe. I'm sure they are more than grateful to the company for pointing out why it is so important for every driver to hit the local store and make such a purchase as quickly as possible.

Personally, however, I'm hitting up the Gunstar web site to see if I can get my hands on a second hand rocket propelled grenade launcher that I can affix to the hood of my car. Rather than a camera that they can't see, I believe that this will act as a much stronger deterrent to those in the criminal fraternity intent on carrying out such dastardly crimes.

Friday, July 6, 2012

UAV crowd-sourcing project fails to take off

It must have seemed like a rather good idea to the folks at DARPA to hold a competition to discover who might be capable of designing, building and manufacturing an advanced small unmanned air vehicle (UAV) that would be capable of performing a simulated military perch-and-stare reconnaissance mission.

Indeed, as the so-called UAVForge project took shape, they must have been delighted and encouraged to see more than 140 teams and 3,500 individuals from 153 countries crawl out of the woodwork to attempt to develop systems that would meet the rigorous demands laid down by the agency. But that was hardly surprising, since a whopping $100,000 prize was up for grabs for the team that could demonstrate that their system met the agency's goals.

With so many teams competing and so much cash at stake, it seemed inevitable that one of the teams would win the competition. Sadly, however, none of the nine finalist teams managed to do so. When they demonstrated their air vehicles at an event at Fort Stewart, Georgia, not one of the teams proved that they had what it took to fly home with the big bucks.

The fly-off scenario, conducted on a training site at Fort Stewart was a simulated military perch-and-stare reconnaissance mission that required that the teams' UAVs performed a vertical take-off, navigated to an area beyond the line of sight from the take-off location, land on a structure, capture video and then return to the starting point.

While some teams were able to reach the observation area, none were able to land on the structure and complete the mission. Since no team completed the fly-off event, the $100,000 prize was not awarded, and a design will not be manufactured for further testing in a military exercise as originally envisaged by the folks at DARPA.

If the failure of the so-called UAVForge project has proved one thing, it is that developing such a small unmanned air vehicle (UAV) is clearly beyond the role of so-called citizen scientists.

Might I dare to suggest, then, that if such persistent, beyond-line-of-sight, perch and stare surveillance systems are still of importance to DARPA, they may be better off calling upon one of the usual large military contractors to help them out.

That, however, is certain to cost a lot more than $100,000 in prize money, I'll wager.

Interested in reading more about UAVs? Then why not check out these recent news stories from Vision Systems Design?

1. UAV captures 3-D images of buildings
 Engineers at the University of Granada (Granada, Spain) are using UAVs to help them produce 3-D models of historical buildings.

2. UAVs help utilities bring back the power
Researchers at New Mexico State University (NMSU; Las Cruces, NM, USA) and the Electric Power Research Institute (EPRI; Palo Alto, CA, USA) recently completed tests that concluded that unmanned aircraft can be safely and effectively used to assess power grid damage following a storm or natural disaster.

3. Small UAV uses hyperspectral imager

Headwall Photonics' (Fitchburg, MA, USA) Micro-Hyperspec imaging sensor is being successfully deployed onboard small commercial unmanned aerial vehicles (UAVs) to help agriculturalists monitor vegetation over wide areas.

4. Robot vision helps guide UAV for crop spraying

Australian researchers are developing a flying robot as small as a dinner plate and a fleet of eco-friendly robotic farmhands that could help cut down the amount of herbicide sprayed on crops.

5. Cameras take flight on UAVs to help soldiers spot suspicious activity
An unmanned aerial vehicle (UAV) developed by a team of engineers from Middlesex University (London, UK) could help soldiers to spot hidden dangers during military operations.

Monday, July 2, 2012

Computers identify tech trends

One of the biggest tasks of any senior manager is to identify new trends in technology and take advantage of that knowledge to develop new products before his competition.

Traditionally, most folks have performed this task in a variety of ways -- by attending technical conferences and seminars, visiting trade shows and yes, even reading technical trade magazines such as Vision Systems Design.

But the trouble with all these approaches is that they require a great deal of time consuming and exhaustive human effort. What makes things worse is that since no-one can claim to be a font of all knowledge, oftentimes the technologies that may appear new to one individual might actually be years old.

Now, however, it looks as though it might be possible to employ the use of computer systems to help alleviate the drudgery of such research. At least, that's what the folks at Rensselaer Polytechnic Institute (Troy, NY, USA)  believe they might be able to do.

That's right. The scientists there have begun work on a new Intelligence Advanced Research Projects Activity (IARPA) project to develop computer systems that can help quickly identify emerging ideas and capabilities in technology.

The research is part of the IARPA Foresight and Understanding from Scientific Exposition (FUSE) program under a team led by BAE Systems that includes Brandeis University, New York University, 1790 Analytics, and Rensselaer.

The computer and web scientists at Rensselaer -- led by Professor Deborah McGuinness --- will work with the FUSE team to develop computer programs that will analyze millions of pages of text looking for the emergence of new technological and scientific trends in multiple languages.

"No one can keep up with the massive amount of data currently out there even in one language, let alone in many different languages," McGuinness said in a recent statement.

"(The project) will allow us to look at a far greater number of documents in less time to understand the significant trends that are out there. Once identified, these trends can then be better studied by human analysts."

While this work is admirable, I'd like to suggest that the good folks at the Intelligence Advanced Research Projects Activity (IARPA) might also like to sponsor another group of software developers to create a computer program that could analyze the technical specifications, price and performance of products and correlate those characteristics with how successful they have been in the market.

I think that this would provide a terrifically important tool to many individuals -- especially those in the vision industry -- who might then be able to make more informed decisions about what products to launch into the market.