OpenALPR Acquired By Mysterious Novume

By Dan Gelinas, Published Mar 13, 2019, 01:50pm EDT

Startup OpenALPR has been acquired by Novume, a company virtually unknown in the industry.

While there are many LPR providers (see our directory), OpenALPR is best known for its open source option + cloud LPR service.

In this note, based on our talks with Novume CEO Robert Berman and OpenALPR Founder and Developer Matt Hill, we look at the following:

  • Terms of the deal
  • Pricing
  • Go-to-market strategy
  • Who Novume is and what they are trying to do
  • Novume stock
  • Novume revenue
  • OpenALPR revenue
  • Future intentions

Terms ** *** ****

********-***** *************** *** **** ** March ******* ~$** ******* ********* ** a ******** ****** **** the ***, **** *** ***** release ****** ****** ******** "additional ******* *******":

** ******* *** *********** and ******* ********** ******* capital, ****** ****** $** million ** ****** ******* promissory ***** ** * consortium ** ******* *******.

*** **** ******** * combination ** ****, *****, and *****, ********* ** sources ***** ** *** company.

************, **** **** **** would ********** **** **** to ****** ** *** company's *** ***** ******* Officer ***** ** ***** help ****** ********** ** a ******-****** ********** *******, the ******* ********* ** IPVM ****:

*** ** **, *** Matt *** * **** understanding ** *** ** make ***** ************ ****. We **** ** *** technology ** **** *** products ******. **** **** help **** **** ********** to * ****** ****** technology *******.

OpenALPR *******

** ******** ********'* ******* is ** *** *** to *** ****** ***** millions, ****** *** ******* declined ** *******. ******* close ** *** ******* said *** ******* *** growing *** **********, ***** is ********** ***** ***** small **** (***** ** employees). *** ************ *** mention "********’* ******* ***** recognition (“***”) ******** *** increased ** **** *,*** cameras". *******, *** *** of ***** ******* *** on ***** **** ** unknown *** **** ***** from **** ** $** per ***** ** * one **** $*** *******.

************ ***** ********** * **** ****.

*******

***** ****** **** ***** were ****** ***** ** adjust ********* *****, **** would *** ********* ** increases ** *********, *** provide * *********.

******** ********* *** * few ********* *****—**-******* ****,***** ******,***** ***, ********** ***** ******. ****** ***** **** a *******-******* **** ******* to $**/******/***** ** $*** per ***** *** *** cloud *** ********** ********.

Go-to-market ********

******** *** ************* **** available ** *** ***** directly *** *********** *******. ****** **** ***** were ** ***** ** change *** ***** ** this ****.

About ******

****** ** * ******** traded ******* ******* **** has **** ********* / rolling ** ******* ******-****** focused ********* **** *** past *** *****. *** company **** **** **** 120 ********* *** ********* own * ***** *********, including:

  • ********* ******* ******, * ******** ** automated ******* ****** *********** solutions, ******* *********** ******** collections, *** ****** ********** equipment ********* *** ****** safety ********.
  • ****** ****** ******, * ******** *** integrator ** ****** ****** vehicle ********** ********* ** municipalities, *** ** ********, and ***** ******* *** public ********.
  • *** *** *********, * ********** **** focused ******** ***********, ******* planning, *** ****** ********** services *** ******* ******** for ********** *********.
  • ********* *********, * ****** ********** consulting ****.
  • ****** ********* ********, * ********* ******** company **** ******** *********** support ** *** *** and ********* ********.
  • ****** ******** *************, * ******** ****** that ****** ********* ************* and *********** *** *********** facilities

******** **** **** **** portfolio *** ********'* **** says ** **** ******* opportunity:

*** ** ***** ************ is *********. **** ***** a ****** ** ********* products. **** **** ***** **** ********** ****’** ****** *** to. ** ******** **** over ****. ****** ***** partnerships. * *** *** potential *** **** *******. Plus ******'* ******* *** was ***** *********** *******, so ***** *********** **** help ** ***** ** as ****.

Risky ********

****** **** **** *****. The ******* *** * market ************** ** **** ~$14 ******* *** *** its ***** ****** ** less **** $* *** share, ******* *********** ***** after *** *********** ************.

***** ****** ************** ****** in **** **** ** ~$100 ******* ***, ****** is **** **** **** 80% ***** ****.

Financials **** ******

******* ******* ** **** Novume *** *** ****** its ** **** **********. However, ********* ** ******, ***** ******* *** on **** *** ~$** million *** **** **** a *** **** ** ~10% ** *** ******* below *****:

*** ***** ***** ******** the ******* *** ****** and *********' ******** ***** the ********* ** ******* these ********** ***********.

Future **********

****** **** **** ***** forward, *** ******* ***** focus ** ******* *** business *** ******* ** the ******** ********. **** said **** *** ***** Science ******* **** ***** undertake * ****** ** pursuits **** *** *** of ********** ** ******* the ******** ******** *** would **** ****** ** some *** ********:

**** **** ** ****** out **** * *** version ** *** ******** in *** ******, ****** we ***'* **** * definite ********* ** ****. That **** **** *********** improvements.

*******

** ** **** *** us ** ***** * clear *******. ***** ******** sits ** * ******* trend *** ***** * different ******** ** **** LPR *********, ****** ** embarking ** * *********** and ***** ******** ****** all ** ***** *********.

Comments (22)

Undisclosed #1

03/13/19 06:20pm

Congrats to Matt on this. While not the largest acquisition in terms of size, this is clearly one of the more successful acquisitions just in terms of size of the company and market penetration relative to the acquisition price.

Agree: 2
Disagree
Informative
Unhelpful
Funny

Undisclosed Manufacturer #2

03/13/19 06:34pm

Congrats Matt Hill and Steve Lewis, proof that engineering and business leaders can be a great team!

Agree: 4
Disagree
Informative
Unhelpful
Funny

Steve Lewis

03/13/19 06:42pm

Thank you UM#2.

Agree: 6
Disagree
Informative
Unhelpful
Funny
Avatar

Matthew Del Salto

BLACKBOX USA
03/14/19 02:36pm

Excellent work Matt and the OpenALPR team. I hope the cloud API stays in place it is crucial for some of the integrations we do.

Agree
Disagree
Informative: 1
Unhelpful
Funny

Undisclosed #3

03/17/19 12:21am

"LPR is AI, and Matt has a deep understanding of how to make these technologies work."

i don't know Matt.

But I know that LPR is not AI... so I'm not sure how that comment means anything.

Agree: 4
Disagree
Informative: 1
Unhelpful: 1
Funny

Undisclosed #4

IPVMU Certified | In reply to Undisclosed #3 | 03/17/19 02:33am

OpenALPR Acquired By Mysterious Novume

i am disappointed to learn that the acquiring company’s name is simply “Novume”, not “Mysterious Novume”, which imo is far cooler :)

Agree
Disagree
Informative
Unhelpful
Funny: 4

Undisclosed Manufacturer #5

In reply to Undisclosed #3 | 03/17/19 03:37am

I would have to disagree on a philosophical level.  While traditional LPR is just OCR of an image, that isn’t really AI.  

However, if the camera or server continues to “learn” what is a plate, what plates do expect without programming, then that would be AI.  

Outside the US it appears plates are pretty uniform.   In the US, there are plates that have symbols, small letters and such.

Agree
Disagree
Informative: 2
Unhelpful
Funny

Undisclosed #3

In reply to Undisclosed Manufacturer #5 | 03/17/19 03:49am

"However, if the camera or server continues to “learn” what is a plate"

You are right that I was referring to LPR being simple OCR - and not AI....

Can you explain how AI could be used to 'learn' what a plate is - when inputting the standardized characteristics of a plate alread 'teaches' the software all it really needs to know?

Agree
Disagree
Informative: 1
Unhelpful
Funny

Undisclosed #4

IPVMU Certified | In reply to Undisclosed #3 | 03/17/19 05:56am

Can you explain how AI could be used to 'learn' what a plate is - when inputting the standardized characteristics of a plate alread 'teaches' the software all it really needs to know?

neural networks (machine learning/human brain) are (or should be) better than hard-coded ocr type programs.

if, (brains in this case) weren’t better, captcha’s would not have worked for so long. and if it wasn’t for machine learning getting better at cracking them, they’d be easier today.

you don’t (typically) tell AI what a character (glyph) looks like ahead of time. you show it a plate, it takes a guess, you say right, or show it the answer when it’s wrong.  

based on how it did, it then makes minor changes to various weighted probabilities, which, after bazillions of test plates, leaves you with a neural network.  it then uses this neural network in production to make highly informed “guesses” based on its accumulated biases.

the neural network can continue to learn from production examples, although the number of new events is typically far fewer than the initial examples.  but it can help fine tune a neural net for regional differences.

Agree: 1
Disagree
Informative: 1
Unhelpful
Funny

Undisclosed #3

In reply to Undisclosed #4 | 03/17/19 06:47am

good explanation of what a neural network is and how it works... I think everyone needs to understand the fundamentals of how they work and what they can do.

but ALPR was one of the first large-scale specific use cases for VCA - primarily because the use of license plates is pretty standardized.

i.e. shape and size of both plates and characters are already known - even though each state may differ in some manner.  it was a good fit for the technology of the day - and it worked well then (with certain lighting, AOV and resolution considerations) and continues to work well today.

so if ALPR pre-dates commercially available neural network options - and seems to work pretty well already, then how does AI/neural networks provide additional value?

I'm not saying AI has no value.... I'm simply wondering how AI improves ALPR - which already works pretty well.

 

Agree
Disagree
Informative: 1
Unhelpful
Funny

Undisclosed #4

IPVMU Certified | In reply to Undisclosed #3 | 03/17/19 07:05am

maybe it’s not that much better, since with good cameras, good views and daytime IR etc, you can get good results.

though maybe it lets you get away with less high-end gear to achieve the same.  just speculating here.

Agree
Disagree
Informative: 1
Unhelpful
Funny

Undisclosed Manufacturer #5

In reply to Undisclosed #3 | 03/17/19 03:00pm

Maybe in other countries plates are basically the same.  In the US there are 50 states and about a thousand variations of plates, each having the ability to access any state.   New plates appear annually.   I can see where AI or “constant learning” would be useful and if the additional information of “plate design” could be added, or which state it was issued by.

Agree: 1
Disagree
Informative: 2
Unhelpful
Funny

Undisclosed #3

In reply to Undisclosed Manufacturer #5 | 03/17/19 04:30pm

"In the US there are 50 states and about a thousand variations of plates, each having the ability to access any state. New plates appear annually."

while each state in the US has its own plates (and some states have various different types/themes, etc) - every version available is still standardized by that state.

meaning ALPR can 'know' all the existing state license plates - and all variations of each - simply by entering this info into a database.

i.e. ALPR is 'taught' by data entry and has no need for AI - since each state uses identifiable standardization of plate size and character font/size/spacing, etc.  

"toll by plate' is a cashless tolling system that has been in place for many years on the turnpikes/thruways of numerous US states.  

I can assure you that out of state drivers are paying their fare share (intentional typo pun).  😏

  

Agree
Disagree
Informative: 1
Unhelpful
Funny

Undisclosed #6

In reply to Undisclosed #3 | 03/18/19 03:29pm

AI is a broad term.  It would have been more accurate to state... LPR and vehicle recognition is an application of AI of which "machine learning (ML)" or "deep learning" is a subset.  OpenALPR uses machine learning to recognize license plates of dozens of countries/regions and vehicle make, model, color, body type, etc.  Tens of thousands of real world images are labelled by humans and that image data and metadata is fed into a convolution neural network so that the algorithm 'learns' from all that input.  This is the typical process for many content video analytics. Cameras are generally not learning or processing the data on the edge, rather that is done at the server / cloud.

Agree
Disagree
Informative: 1
Unhelpful
Funny

Undisclosed End User #7

03/19/19 04:30pm

Congrats & Much success to the OpenALPR team & its leaders. From a satisfied and happy customer. Superior customer service, keep the good work up. Look forward to the Spring release.

Agree
Disagree
Informative
Unhelpful
Funny

Knight Hamalian

04/02/19 10:53pm

I was wondering how a company called NoVuMe would be successful selling ALPR...   Then I find this press release:  Novume Solutions to change name to Rekor Systems, Inc.  I'm looking forward to seeing how they do with the OpenALPR acquisition.

Novume Announces Increased Focus on Technology Products and Services and Intention to Change Name and Segment Businesses

Agree
Disagree
Informative: 1
Unhelpful
Funny

Undisclosed #6

In reply to Knight Hamalian | 04/03/19 12:42am

Novume (pronounced "no voom") is just the holding company for other subsidiaries (likr Brekford Traffic now Rekor). I think the OpenALPR brand is worth preserving and Novume would be wise to utilize "powered by OpenALPR" in future products.

 

Agree
Disagree
Informative
Unhelpful
Funny
Avatar

Matthew Del Salto

BLACKBOX USA
05/29/19 05:07pm

Looks like they just doubled the price on the saas cloud alpr. Really ashame since it just started taking off for us. 

Agree
Disagree
Informative
Unhelpful
Funny

Undisclosed #6

05/29/19 05:36pm

The prices (effective July 2019) on the Cloud API changed as follows:

Free tier =  2000 recognitions per month, now 1,000 per month - 100% decrease in the number of recognitions.

Basic tier = Was $49 per month for 50K recognitions per month now $99 for 25K per month - 100% increase in cost, 100% decrease in recognitions.

Pro tier = was $195 per month for 250K recognitions per month now $395 for 125K per month -  100% increase in cost, 100% decrease in recognitions

Enterprise tier = was $995 per month for 2M recognitions per month now $1995 for 1M per month -  100% increase in cost, 100% decrease in the number of recognitions.

Not a smart move to make people pay more and give them less.  The cost to do this has not changed significantly in the cloud to justify the increase.

Sighthound prices had basically mirrored OpenALPR's original plans: https://www.sighthound.com/products/cloud  and for now remain unchanged.

 

 

Agree
Disagree
Informative: 1
Unhelpful
Funny

Dirk Hedlund

10/09/19 06:53pm

Rekor is in the process of cannon-balling OpenALPR for integrators.

We inquired with their inside sales department if there was a channel partner program to sell OpenALPR, we were firmly told they are selling direct to end-users and it would be a conflict of interest if we were to buy just their software as we would be 'competing' against them.

OpenALPR will be made unavailable (sooner rather than later) for custom solutions and gobbled up by Rekor in an attempt to monopolize what was an open and powerful piece of software.

Agree
Disagree
Informative: 1
Unhelpful
Funny

Steve Lewis

In reply to Dirk Hedlund | 10/09/19 07:28pm

This is very discouraging to hear especially because it is contrary to the strategy on which we built the company and one of the main reasons we were able to grow so fast in a short period of time.

Agree: 1
Disagree
Informative
Unhelpful
Funny

Dirk Hedlund

In reply to Steve Lewis | 10/09/19 07:58pm

Thanks for chiming in on this Steve, very disappointing to say the least. We'll just have to wait and see how things 'evolve'. I hope my assumptions are incorrect.

Agree
Disagree
Informative
Unhelpful
Funny
Read this IPVM report for free.

This article is part of IPVM's 6,958 reports, 927 tests and is only available to members. To get a one-time preview of our work, enter your work email to access the full article.

Already a member? Login here | Join now
Loading Related Reports