Video Analytics Market Guide and Favorities 2014By: John Honovich, Published on Mar 03, 2014
For more than a decade, video analytics has been the great hope for the surveillance industry. Indeed, it remains the pick as the Next Big Thing.
Yet the past 2 years has been rough, with ObjectVideo's lawsuit campaign against almost every major industry manufacturer and no investment / new startups in the space, culminating with one of the biggest names, VideoIQ, being acquired at less than their total funding.
IPVM State of Video Analytic Survey
IPVM has conducted a survey of 120+ integrators from over 20 countries, answering:
- How often video analytics are being deployed
- What is driving video analytic adoption
- Who are the favorite video analytic manufacturers
Integrators are not deploying a lot of analytics:
Indeed, the rate is actually slightly worse than it was in 2011:
With quite a number moving from sometimes to rarely.
On the other hand, this should not be surprising, as the supply of video analytics and the development of video analytics has been equally flat or declining.
Video analytics had problems in 2011, and without solutions to those problems, it is hard to make progress, despite naive people who use to claim it was just an 'educational' issue.
Why or Why Not Analytics?
Let's first start with the integrators who say they never deploy analytics.
Problematic performance was the #1 cited reasons:
- "Analytics are expensive, troublesome, never guaranteed to work, most of our customers are happy to just have video available. That doesn't mean i don't try to sell it. I do but often stopped all to soon, as "we don't need all that"."
- "Analytics still needs a lot of 'man' acceptance, they are not on parr with other detection forms.(PIR, traditional perimeter detection systems) they should be used in conjunction with other technologies to provide reliable results."
- "All of our past deployments have failed miserably to meet the requirements even for simple tasks. We prefer to accompany the surveillance system with supportive detection (radars, laser scanners, outdoor detectors, etc)"
- "For the expense, it is still hit or miss."
- "False Alarm / low accuracy is the major deterring factor. Cost prohibitive will be the second factor."
Quite a number, also mentioned no demand:
- "My customers have no need for analytics"
- "Lack of use for our primary customer base."
- "We do not have a need for video analytics with our customer base."
- "We have not come across a customer that had a need that we felt required video analytics to solve yet."
- "Never asked for by customers"
- "We have no requests from Clients."
While it is believable that there is minimal demand, we certainly believe there is a fit in any and every market for video analytics that work well - even residential applications would benefit from knowing / tagging potential intruders as well as improving search.
Contrast to those who often deploy video analytics.
Monitoring and perimeter protection were the key themes here:
- "We have a monitoring station doing video alarm verification so we only use the analytics for alarm trips human or vehicle."
- "When paired with offsite monitoring, it is a cost effective solution for perimeter protection."
- "We install video analytics to save on costs and provide reliable perimeter detection systems. We use thermal cameras as sensor to provide reliable detection system in all weather conditions"
- "We need to use advanced video analytics. If not, we will get too many unwanted/false alarms into our operations center. We need to be fairly confident that the incoming alarm is for real."
Favorite Video Analytic Manufacturers
Results were similar to 2011, with none 'winning' and VideoIQ being the favorite of those who had a selection. Two big surprises were the emergence of Axis and Mobotix, who back in 2011, had no votes and no real offerings.
Given that 2/3rds of integrators rarely or never deploy analytics, it comes as no surprise that nearly half of all respondents did not offer any selection for favorite, with many writing in "none", or "n/a".
Contrast to cameras and VMSes, where essentially everyone had at least one favorite.
VideoIQ pulled in ~15% of the vote, similar to their performance in 2011, and still more than twice of the next manufacturer.
Integrators emphasized VideoIQ's reputation and field performance:
- "We selected them since several of the members in our group have used this product for similar applications with great success."
- "Based on reputation and seeing it in action ourselves."
- "The only one that works"
- "VideoIQ is accurate with options to suit our needs."
- "I am the most familiar with them."
- "VideoIQ , most positive experience for outdoor use ."
- "VideoIQ. Works well and Avigilon acquisition built further credibility."
This is good for Avigilon, as it gives them an impressive all-around Top 5 favorites for IP cameras (2nd), VMS (4th) and video analytics (1st). No one else had such showings in all 3 categories and only 1 had it in even 2 of them.
While a distant second after VideoIQ, Axis established itself as the 2nd most favorite analytic.
- "Axis tripwire is good for simple applications."
- "Axis cross line works great some stores and other places that want certain places tracked."
- "Mostly Axis as the cost of their tripwire analytics is low, $25, all others cost a lot more, we have wanted to deploy AgentVi but its kind of pricy"
- "Small list of ACAP processing for AXIS cameras"
- "I have the most confidence in Axis analytics because they have been in the business of IP cameras from the beginning."
Tripwire does work fairly well (see our test results) and given that it costs essentially nothing and comes from the most widely used IP camera manufacturer, there is logic behind this.
On the other hand, Axis does a relatively terrible job marketing / promoting any of their analytics (including VMD2.1 which is free and works very well). And even though they support 3rd party analytic apps (via ACAP), those receive almost no promotion and, worse, without any real clarity about how well any of them might work.
One interesting sign is that Axis has just recently started to add higher end analytics into their cameras, but for now, only on their 1080p PTZ.
If Axis gets serious about their own analytics, this showing demonstrates that they could easily become the industry's favorite analytic provider, simply leveraging their position in cameras. And with Avigilon loading up with VideoIQ, they should have more motivation to do so.
Sightlogix had a solid showing as they continue to represent a niche in the market for higher end, 'military' style analytics. Integrators noted:
- "Edge processing by a company that specializes in outdoor analytics."
- "Sightlogix - more complex and admittedly only used in higher end"
- "Sightlogix seems to be a reliable, but expensive solution"
- "Reliability for our environments"
- "I like their products, they put a ton of horsepower at the edge from what I can tell; and they also set "relatively" realistic expectations in front of customers (issues with water, foliage, etc)."
The main challenge, is that even at decreased prices with a new generation, they still remain very expensive for the mass market.
AgentVI still received some votes, though down from 2011, with integrators noting:
- "AgentVI, we have test their solutions several times."
- "We have wanted to deploy AgentVi but its kind of pricy"
- "Time in service, advanced algorithm offerings, and personal experience."
AgentVI's hybrid model was more interesting / appealing in the past decade when processing power was limited. Now, the challenge is that it is 'pricy' and requires adding it on to cameras and/or servers.
Mobotix only released analytics in the last year or so but they received a number of selections, with integrators noting:
- "Their MxAnalytic and McLeo analytics have thus far worked as promised."
- "We use Mobotix video analytics. It works great."
- Mobotix. Built into every camera" and another, "MOBOTIX due to the integration into the camera"
Its free and does work fairly well (see our test results). The only downside is that its requires one to use Mobotix.
Iomniscient is the only analytics company left making bold claims (save for BRS Labs who is, well, in their own wild world). It seems to be paying off, at least somewhat, as 3 people cited Iomniscient being their favorite (BRS Labs had none).
Where's ioimage / DVTel?
Since acquiring ioimage in 2010, DVTel has essentially hidden their analytics (on purpose or not). They only received one vote despite the fact that the company still has a reputation for high performance. New products are planned for this year and with the VideoIQ deal impacting channel availability, DVTel has an opportunity to regain some ground in analytics.
Future - IP Camera Manufacturers Rising
The market for video analytics is still wide open and the shifts from 2011 show that incumbent IP camera manufacturers can likely dominate video analytics by simply being aggressive about moving into the space with product that perform adequately - a feat far easier today than a decade ago during the initial entry of video analytics into surveillance.