My first lane detection algo

I’m all for practical learning by building things. There’s nothing like getting stuck into a project and seeing results. Whilst a little progress is a good motivator, it also shows you how much it is that you don’t know.

I was pleased with the results of my first project doing the Udacity Self Driving Car Engineer Nanodegree. Yet what was more pleasing was being shown how much experimentation is really required. That is that is so much to learn.

This first module was about understanding some of the principles of computer vision that apply. We first started with Canny Edge Detection

and then Hough transform to detect lines within a region of interest.

The first project was to apply this learning, first to set of static images and then to a couple of videos captured whilst driving on a highway.

How cool is the final result .

I submitted my Jupyter Notebook for review. To pass you need to ensure that you meet specification. I passed here is the review feedback.

Some of my reflection thoughts on what can be improved in a future iteration of the project:

  • look for the road horizon starting from bottom centre of the image working up – asphalt has a fairly unique colour
  • break ROI into left and right lanes earlier – seems that at least with driving on a highway without lane changing that we can assume with confidence where they should start at the base of the image
  • segment each ROI into chained vertical blocks of a smaller width
  • when drawing connect intersection of cv2.fitLine lines
  • increase number of segments if lanes are curving left or right
  • label the lines with colour and type – continuous, dashed etc
  • feed the previous result into the evaluation of the next image
  • determine when an image has no lanes that could be considered reasonable
  • lane changing and entering a lane from a curb needs more though
  • if using smaller more specific left and right lane ROIs should allow for following a vehicle
  • not sure how rain affects this – might have to do a test and capture video in a tropical down pour this storm season
  • this approach wouldn’t work in snow. would require a different approach

Its still early yet in the nano degree but I’m hooked already. Happy coding and driving.

Advertisements

Thoughts, learnings, observations and next steps from my August Bay Area trip

I’ve finally done my first US trip. There was a lot that fascinated me, it was so familiar, I was not an outsider, and I was a little surprised how easily I just fitted in. I seemed to run into intelligent and smart people everywhere, that spoke my language/dialect and liked talking about the same things as me.

My first superficial observations of the bay area, was that it appeared old and not well maintained. Like it had peaked in the past. In parts it was very dry and eucalyptus trees were everywhere – I nearly thought at times, I was in a very dry Adelaide summer. In San Francisco itself, on Market St, where my first hotel was, the homeless and other type of street people were a concern. You soon learnt, that as long as you were street aware, they were not going to cause you trouble. There seemed a disconnect though, with my perceived notion, that techies liked newer more modern things.

I was looking to break out of my techie rut (and comfort zone) and seek stimulus to motivate myself. To that end I had the following questions in my mind:
1/ Would I want to live here for an extended period of time?
2/ Is this a good place to pivot my cycling app or create a new startup?
3/ Perceived effort to find co-founders and/or build out an initial team? and
4/ If not in the Bay Area, where else across the globe should I possibly look at?
5/ whats next?

After I’ve summarised my relevant activities for the two weeks, I’ll answer the above the questions in more detail.

Summary of Activities

As it was US summer time, there weren’t as many meet ups on, as I had originally hoped. I did come a little unstuck in the first week with the distances needed to be travelled and accommodation costs put a hole in my budget somewhat.

Summary of first week activities:
1/ familiarisation, getting a US sim and being a tourist – doing a bay tour
2/ attended a Quantified Self meetup http://www.meetup.com/quantifiedself/events/187124622/ – met people from Intel research, Pebble, MisFit
3/ had lunch at Twitter HQ – met an online friend from the UK for the first time
4/ was chauffeured around Palo Alto, Stanford, Sand Hill Rd by a Californian based friend
5/ Attended the Open Forum: The Intel Trinity: Noyce, More & Grove – a lot there were Intel Alumni
6/ visited Runway http://runway.is/ and caught up with an associate from the Gold Coast – who added me onto the secret Aussie Mafia Facebook group
7/ Caught up with another associate now working for IBM after the startup he was working for was acquired by them

Accommodation was really expensive in SF, I flew to Las Vegas Thursday night (was cheaper then staying in SF) and returned to Mountain View Sunday afternoon (was supposed to be morning but flights delayed 2 or so hours due to fog in SF – seems an ongoing interruption).

Summary of second week activities:
1/ Caught the BART/Caltrain from SFO to Mountain View – met and enjoyed a great conversation with two Canadians that had just gained employment at Facebook as Data Scientists
2/ drove for the first time in a LHD vehicle on US roads for a day around Mountain View
– unfortunately the Computer History Museum was closed on Mondays & Tuesdays
3/ Went and saw HQs for
– Facebook (big campus but nothing to see),
– Google (a few things to see, was larger then I imagined)
– Apple (quite large and bought souvenirs from the company store)
– E-bay (smallish)
– Intel (Stood out as people were wearing suit and tie) – spent time in their museum
4/ Stopped by Red Rock Coffee Co, Mountain View http://bit.ly/1t1HWCp
5/ Travelled via Caltrain from Mountain View back to San Francisco – a lot jump off at Palo Alto in the morning
6/ Visited http://startuphouse.com/ (they have two floors of dorm/bunk bed accommodation http://www.startupbed.com/) went across the street and spoke to http://www.techshop.ws/tssf.html
7/ Caught up with an online friend (from Adelaide, that I had not met before) who was interviewing with Google
8/ Took a Tesla for a test drive with a friend near Sausalito
9/ Attended a 500 startup conference – weapons of mass distribution http://wmd.co/ presentation slides http://www.slideshare.net/500startups and YoutUbe videos http://www.slideshare.net/500startups tagged #500Distro. (These guys just sell online – never once talked about old fashioned shoe leather salesmen sales.)

I didn’t quite end up with a full diary on some days. Many persons, I thought I might catch up with based in San Francisco seemed to be travelling and as I stated earlier there weren’t as many meetups as I thought and some were just too far away from where I was staying. You really do need a car and I now have a better sense, of the time required to travel and to get from Point A to Point B around the bay area.

Returned home via a few days in Hawaii. I luckily dodged two hurricanes.

Would I want to live here for an extended period of time? 
The price of accommodation in San Francisco is very expensive. Unless sharing with another (who had a US credit history) or others, I’m not certain I’d be able to afford it initially. I was informed you need to be earning >$200K USD per annum to really live in SF. It is of course a colder climate there to.

Culturally wise I enjoyed meeting and talking with techies and entrepreneurs. All roads appear to lead to the Bay Area presently. I’d fit in well but I would miss Australia. I’d also have to question whether over time, if the San Francisco Hipster culture would wane on me.

Am not certain at my age to, if I could really bunk with 20 year olds, in dorm style accommodation in the SoMA area. I’d much prefer Mountain View or Palo Alto – the cafes (and warmer weather) along University Drive had appeal. I did not however get a sense of the costs, besides that it to, was expensive.

One friend (working for IBM) doesn’t rent, as he is flying around the US and sometimes back to Aus but instead uses friends rooms (if they are available) and/or AirBnB (has some regulars) whilst in SF. He flies out to Reno or elsewhere where he can get a hotel for $30 per night. I was a little surprised to hear this, but he seems comfortable now living out of a suitcase for a year or so.

Another associate, is planting himself in San Diego and will fly in as required to SF. The fog though may impact his timely arrival if just for a day trip.

So really to answer this question, if I was not funded in USD, it would appear just to be too expensive to live there. As to, if it should be in San Francisco or closer to Silicon Valley, that would somewhat depend on the startup itself. (In Whats Next? i’ll state what I observed from Michael S. Malone)

I have heard of a proposal put forward to the Queensland Government for funding of a Startup House for Queenslanders in SF that was initially received positively by the minister here. Something for me to follow up on.

Is this a good place to pivot my cycling app or create a new startup?
The simple and quick answer of course is yes, if accommodation was not a concern.

However I’m not certain I could fund the cost to find product/market fit if I was to do it all in the Bay Area. Plus I wouldn’t be in a position to fund others initially. I’d need to find other local co-founders who are finically independent or work with another complementary startup e.g. working on a tracking device.

A standard Aussie US visa can be used for up to three months at a time, if you do not earn income in USD whilst there. A business visa for Aussies is six months with similar restrictions re earnings. Apparently, others upon setting up a US company then arrange for a H1B (I think??) visa to sponsor themselves. Suspect the Aussie Mafia on FB can assist here. US health insurance is something I don’t understand properly yet.

I did observe, that any discussion around my cycling app (or variations) were generally well received in that people thought there was something there. On reflection, I did not temper this, though with other ideas, to A/B test to see if it was just common to talk about ideas or not. However I did seem to set off a spark within a few.

There was still great momentum around ideas related to health and wellness with the Quantified Self movement. They now will be focusing on access to the data (not just the files but from an everyday users perspective). The Apple Health Kit and Google alternate were not really mentioned. I suspect these will generate new startups.

People were willing to bounce ideas around to – chew the fat if you like. Of course some are looking for angles for you to leverage their startup. This is something that I don’t get enough off around here in the Gold Coast or in Brisbane.

Something that was niggling away in the back of my mind, was getting a job in the Bay Area. Using that to learn more so and work after hours on networking on the startup. However, it somewhat goes against my grain of working for a company that I haven’t equity in and its been a while since I’ve been an employee. Whether I could gain work experience with a VC is a thought but I’ve not explored this properly.

I do not perceive it would be an issue to perform structured interviews in the bay area of 30+ persons/organisations related to hypothesis around a business model pivot of the cycling app. I suspect it would be a very positive and a beneficial way to also learn & network and/or would assist in the next question.

Perceived effort to find co-founders and/or build out an initial team?
There was an enough anecdotal evidence for me to be confident that I could form a team if I had funding. Bicycles, hipsters/techies, entrepreneurs seemed to be attracted to the notions I was putting forward. Have even managed to get the cycling data from the guy who runs Startup House|Beds in SoMA.

Finding a co-founder in the US would take more effort. However, if I was to network more so there and perform the structured interviews, my best guess is I’d have a good shot at finding one or more.

There is a strong cycling culture around San Francisco (at least for commute to work) and I noticed also the Google Bicycles at Google HQ. Some I spoke to said around Berkley it was more what I’m used to with MAMILs on good carbon road bikes.

I bought a Palo Alto Bicycles jersey from the bike shop on University Drive. I suspect Palo Alto would be a good area to.

If not in the Bay Area, where else across the globe should I possibly look at?
I asked this question of a few. With my friend from the UK, I was expecting to hear Shoreditch mentioned. Instead he said Berlin. When I mentioned it to some others they to, had heard positive things about Berlin. No one really talked about Singapore.

One person from Coca-Cola Ventures, that I met at the 500 startups event, asked about Sydney & Melbourne. I said that all roads seem to lead back to the Bay Area. She mentioned that her team was trying to understand why the Atlanta startup scene was not as entrepreneurial as they’d expected. She was at the 500 startups event to gain an understanding of the type of organisations attending i.e. she was exploring and learning like myself.

Alas, as Michael S Malone stated at the Churchill Club event – “The next Silicon Valley is Silicon Valley”.

whats next?

The notion of disconnect that I had started to sense via the internet between the hipster style startup entrepreneurs and the more seasoned software developers (where scalable software is their craft) was not really observed in the events I attended. If I had spent more time or been there, when some more relevant meet ups were on, I might have been able to investigate this further. My anecdotal evidence suggests that outside of hardware, the hipster style app startups are currently, a class restricted at least initially, to what those seasoned developers have made available. I need to still spend more time on this in the future.

The one notion that I did observe from the Michael S Malone’s talk is that the Silicon Valley veterans are questioning really whats going in San Francisco. Accommodation costs and other costs of living seemed exorbitant. Whilst in previous waves there had been a slow and steady movement towards San Francisco. It had jumped a number of areas directly to San Francisco with the likes of Zygna and Twitter. They also seem to not think there is much more to find around notions related to social interactions. Because of this his thoughts were that a new cluster will form to the west of the current Silicon Valley area where real-estate costs are still moderate (as my iPhone died – battery drains quicker there, I didn’t write it down). I also left with the impression that hardware will again play a stronger role in the next wave – Internet of Things??. He spent a bit of time talking about the number of transistors produced yearly in the world – was it Gordon Moore who said more transistors produced each year, then rain drops in California? (the number now exceeds rain drop in the world – how they can qualify this I’m not sure).

So for me I think the next steps are:
– create a pivot business model for the cycling app
– revise a questionnaire to address the hypothesis in the business model
– book another trip to the Bay Area – try to arrange interviews before I go
– spend more time outside San Francisco (gain greater understanding of cost of living)
– gain greater understanding of visas