Computer Vision in CRE: 5 Current Applications
How Can Computer Vision Be Used in Commercial Real Estate?
The short answer to this question is: the same way that you can use your eyes. Checking for damage to structures, determining the most trafficked areas of a property, monitoring for security threats — these are all tasks that CV models can excel at. The main benefit of CV isn’t that it can do entirely new tasks, but that it can do the same tasks humans have been doing with more consistency, more precision, and in less time. Instead of walking around a property looking for issues, you can program a drone to fly around and report back on its findings, freeing up your time to work on other tasks.
Currently, computer vision is primarily being used in commercial real estate for:
- Building inspections
- Energy management
- Leasing and property management
- Marketing and advertising
- Security and surveillance
Computer Vision for Building Inspections
Comprehensive building inspections are essential for property maintenance and for determining whether a property is suitable for investment. Traditionally, investors conduct inspections by walking around the property and reviewing past repair and renovation documentation. This presents two problems: for one, walkthrough inspections of large buildings are time consuming and can often delay closings. Secondly, two different inspectors (and sometimes even two visits from the same inspector) can turn up wildly divergent results, indicating that the process is not standardized and, consequently, not entirely reliable.
Computer vision can help on both of these fronts. Currently available CV solutions can analyze aerial photography to evaluate roof condition, precarious trees, and fire hazards, while drones can be utilized to collect interior and exterior images for assessment by the CV model. Since only minimal human involvement is required, computer vision can drastically expedite the inspection process.
Additionally, given that the CV model will always follow exactly the same process, the results are standardized, which makes CV ideal when comparing multiple properties for purchase or when making insurance decisions.
For landlords and property managers, computer vision models can be used to monitor areas that are susceptible to damage or may need maintenance. Managers can set up security cameras aimed at an area of interest, such as a utility room, and the CV model can notify them when it detects the need for a repair.
Computer Vision for Energy Management
As awareness of the climate crisis grows, commercial real estate owners are facing increased pressure to go green from tenants, lawmakers, and investors. The push is by no means unwarranted: the U.S. Department of Energy found that commercial buildings account for 18.7% of energy usage, 88 percent of potable water consumption, and 40% of carbon dioxide emissions in the United States.
These figures are not only concerning from an environmental perspective — from a purely business standpoint, this exorbitant resource utilization is expensive. Tenants are also increasingly interested in green buildings in order to align their physical operations with their CSR objectives, which means that investments in sustainable practices are not only environmentally responsible, but economically sensible.
Computer vision is one of the many tools that CRE owners can employ to meet sustainability goals and keep utility costs low. CV models can monitor occupancy and footfall to determine where to direct resources both in real time and based on historical averages. For example, if a CV model detects that an area of the building is unoccupied, it can wind down the HVAC system and then automatically ramp it back up when occupancy increases. Combined with historical occupancy and weather datasets, AI models can predict usage patterns to help utilize resources efficiently.
Computer Vision for Leasing and Property Management
All commercial properties, be it shopping malls, office buildings, or apartment complexes, need to understand the wants and needs of tenants to provide a return on investment. But it’s impossible to truly understand and act on those needs without any hard data — at that point, it’s largely just guessing.
Computer vision, however, can provide that much needed data through guest/door counts, guest journey tracing, and heatmapping asset utilization.
- Guest/door counts measure how many people enter and exit a location and can be used to negotiate with tenants. Vacancies near more popular entrances can fetch higher prices.
- Guest journey tracing maps the trajectory that guests take through a location. Like guest/door counts, it can help tenants and owners make better leasing decisions.
- Heatmapping asset utilization is a hybrid of the two other metrics. Heatmaps visually depict the occupancy of specific areas of a location over a period of time. By looking at a heatmap, owners and tenants can identify the most popular spots in a building and set informed prices.
Computer Vision for CRE Marketing and Advertising
Attractive listings and search optimization are important parts of getting a property in front of potential investors. Now, computer vision is being used to create 3D walkthroughs of buildings and make properties searchable by visual features.
During the early stages of COVID, 3D walkthroughs became an integral part of many real estate searches. For larger buildings, manually creating a walkthrough can be a taxing endeavor, but computer vision can help interpret the images and string them together to create an appealing virtual tour experience. These same capabilities can also be used to automatically generate floor plans.
Although it’s not involved in the commercial real estate market, Zillow has already begun integrating AI-generated 3D walkthroughs and floorplans into its listings.
CV models can also augment listings by making visual features searchable. For example, a CV model could look through a series of photos and extract the tags “modern,” “cozy,” and “brick wall.” A potential buyer can then search listings for those tags.
Computer Vision for Security and Surveillance
Alongside routine maintenance, security and surveillance are essential for ensuring a safe and desirable property. Although automated surveillance systems are already prevalent (practically every alarm system has motion detection capabilities), computer vision can help improve on existing systems’ weak spots.
False alarms are a major problem. A 2007 study from the Arizona State University Center for Problem-Oriented Policing found that police responded to approximately 36 million alarm calls in 2002, 94-98% of which were false. These false alarms cause a waste of time and resources.
Computer vision may be able to help cut down on the number of false alarms. Current motion detectors are notorious for having a very high sensitivity for motion but a very low specificity for human motion. CV models trained to recognize humans may be more specific, reducing the amount of wasted emergency resources.
Key Takeaways: Computer Vision in Real Estate
To stay competitive, investors, landlords, and tenants must ensure they have the tools they need to succeed. Computer vision offers new ways to optimize and streamline building maintenance and inspections, property valuation, energy management, advertising and marketing, leasing and property management, and security and surveillance.
Ready to get started with door counting? Learn more about how curbFlow can help you optimize your property.