Property crimes, thefts and robberies, home intrusions have always made their way to the newspapers, and do not seem to decline now. According to National Crime Records Bureau (NCRB), 244,000+ cases of robbery, theft, burglary, dacoity, etc. happened in residential buildings in 2017. This was an increase by over 10% from 2016 when the numbers were around 220,000. The financial loss due to these break-ins and robberies are astonishing. The value of property stolen from domestic buildings in 2017 was more than Rs. 2000 crores, 40% more from Rs. 1,475 crores stolen the previous year. Its high time now that we start imagining about our own methods to keep our home safer. Here’s where home security systems and video surveillance can help.
It’s really hard to believe that you might not have heard about Artificial Intelligence and Machine Learning in today’s times, and what marvels the developers are accomplishing with the help of AI. AI always seems to be one step ahead of humans and yet, its works are totally dictated by few hundred lines of codes. So why not bring the miracles of AI into the world of security and surveillance? Well, this is not really a brand-new concept. Ventures like SimpliSafe, Frontpoint and Vivint have been working on this unrealistic collaboration of AI and Video Surveillance.
Talking about India, companies like Vidooly have progressed in the discussed domain, still you won’t see a lot of houses and buildings, even the luxurious ones, equipped with AI-based security systems. I have come up with my own reasons to that:
· High cost of AI-based technologies
In countries like India, people are easily fascinated by new wonders. Big companies have always been able to manipulate the crowd and extract huge chunks of profits. AI technologies share the same story. Any service can see its cost to be quadrupled if assisted by an AI.
· Ignorance to probability of intrusion by majority of households
It has been witnessed by us in our respective vicinities. They acknowledge the fact that a neighbour had a break-in, even several break-ins. They know that the robbers have still not been caught, or more of them have joined the gang. But they still do not take action to make their home a bit safer.
· No trust or use of scientific technologies
Most of the Indians do not tend to rely on advanced technologies for things like security and safety. Even if they do, they don’t really use it. Take car seatbelts for an example. You would see majority of car drivers don’t put their seatbelts on. So, when they are not using a safety system handed over to them for free, why would they invest their “hard-earned” money in technologies like AI.
Owing to these factors, I came up with a venture of our own- CenSuS. CenSuS stands for Centralized Surveillance Systems. My idea is not different from the ones that SimpliSafe or Vivint use, i.e. using AI for better video analytics and surveillance. CenSuS will provide with AI-equipped surveillance cameras and locking systems that provide real-time analysis, threat and intrusion detection and remote locking/unlocking facilities. The AI used in the system, called SurvAI, will be a highly trained algorithm that can track and analyse following domains:
· Intrusion and trespassing
· Lockpicking attempts
· Scans for unknown individuals within building range
The cameras might excel in the following domains:
· Real time footage, broadcasted on owner’s devices
· Equipped with SurvAI, it can precisely scan objects and humans
· Equipped with infrared lights to continue surveillance at night
· Special Image processing algorithms help produce clear images even in rains
MARKET DEMAND
The home security market is growing day by day, with new research being performed by experts to be implemented in real time. In order to keep up with market demands, I plan to collaborate with public universities to help advance the technologies. Though the demand is not “huge” in India, I predict that smart home security systems’ demands would increase tenfold by the end of next decade. People are realizing the large cost of employing manpower to protect their neighbourhood and are switching to smart systems.
COMPETITORS AND THEIR PRODUCTS
Companies like Vidooly and Weebly are already out their testing their products in the Indian market. Although they are using AI for video analytics, they lack in providing real time footage broadcasting. With CenSuS, you can watch your home from any where in the world, using the CenSuS iOS and Android app. Apart from that, it provides accurate AI Image recognition to monitor intruder activities, which analyse the level of threat. A mild threat sends an alert to the owner’s devices, as well as to the mobile phones of a local friend that the owner designates through the app. A severe threat alerts the owner, the local friend and the local Police station, and triggers the alarm systems installed in the building. Lastly, the video footages and surveillance reports are stored in an on-site storage disk, as well as on cloud storages.
TARGET CUSTOMERS
I believe we should build products for the world. Therefore, I would provide subscription packs for the customer’s needs. I aim to expand the userbase from small households to large companies and industries. Based on the subscription and need, CenSuS provides the number of devices, cameras and security servers to the respective locations, and makes sure no one is left out in acquiring our services.
REQUIRED EXPERTISE AND KNOWLEDGE
To build such solutions, I will require extensive training in building Artificial Intelligence and Machine Learning projects, as well as good knowledge of image recognition, image processing and video analytics. These skills are required to train the algorithms for more accurate video analytics. Apart from that, surveillance cameras, modems, and server hardware are required to be installed in the building or residence. Programming languages that will be used would primarily be Python, Swift and Android. I would use Amazon AWS for cloud services.
WORKING OF OUR PRODUCTS
After being installed in the facility, the cameras constantly monitor the area and look out for suspicious activities and temporarily store the images of individual within the prescribed range. The alarm system goes off based on the threat level, and sends constant reports and alerts through mails and security alert notifications. The cloud server helps in simultaneous uploading of footages, which users can access through their mobile phones any time, any where. CenSuS would use 4K 60 fps cameras to decrease the probability of inaccuracy and missing out on significant information. The central control server is also equipped with remote technologies, so users can check and control the locks from anywhere.
ESTIMATED BUDGET
Owing to the cutting-edge technologies, the budget is bound to be hefty. I project an estimated investment of around Rs. 10–20 lakh to become a global standard and trustworthy venture and acquire resources to develop the best security systems.
PRODUCT PRICING
As we discussed above, the pricing is based on subscription plans, that go as follows:
Small Household
Small houses and apartments, mostly with one or two house entries.
Rs. 1–3.5 lakh
Big Households
For larger estates and mansions
Rs. 4–10 lakh
Corporate Building
For apartment societies, office building and small institutions
Rs. 10–12 lakh
Large Scope Pack
For large universities, plants and industries.
Rs. 20 — 35 lakh
The prices might seem a bit high. But I will provide with the cutting edge technology. You would pay for the ultimate security system with lowest probability of going wrong.
MARKETING AND ADVERTISEMENTS
The marketing plan for CenSuS is a rather straightforward one. If I get an investor to back me up, I plan to invest 8–10% of the total budget to advertisements. These include pay-per-click and SEO marketing. Alongside, YouTube ads and sponsored posts on Instagram, Facebook, Twitter and Reddit would be an important thing to see. Lastly, direct communication with housing and real estate companies, and more mouth-to-mouth advertisement methods would be a few thing I will rely on.
RISK ANALYSIS — WHAT CAN POSSIBLY GO WRONG
This is one of the most crucial things to analyse. Now, when you build an AI based technology, the biggest worry is that, “What if it does work how I want it to?” or “What if it starts acting weirdly, and not what it was programmed for?” Apart from these, the biggest concern would be cybersecurity. Hackers might see these systems as a real playground to test their not-so-ethical hacking skills.
MILESTONES AND TIMELINE
By my approximation, the project/venture is likely to take an year or two before becoming a fully functional and secure surveillance and reconnaissance system.
Starting in 2021, the start-up would most likely be shown to the investors by the end of 2023. Planning goes for 3 months. The programming languages would take a hefty time of 6–7 months. Building the first prototype would take around 5 months. Then integrating everything and developing the final fully functional system would take more than 5 months.
Although its still just an idea, it is not hard to imagine such a system becoming a reality in the coming future. AI and Video Analytics are both running at full pace towards advancements. So, if not me, someone else will probably turn this idea into a venture for real. There are millions of ideas out there, you just need to pick one to make this world a better place to live.
