汽车 > 无人驾驶博客



Hayden AI: The Future of Law Enforcement

by Roger Lanctot | 7月 10, 2019

By now, we are all familiar with license plate readers being used by public authorities for everything from toll payment to camera-equipped patrol cars tracking down parking violators and stolen vehicles. Some of us remember how smartphone camera video was collected by investigators in the wake of the Boston Marathon bombing and used to identify the perpetrators.

With the growing number of cars equipped with front, rear and surround view cameras, what if live video could be collected from cars in real- or near-real-time in order to solve crime? What if public authorities were willing to pay for access to that camera data in the interest of solving crime?

That is the rationale behind Hayden AI Technologies, an early stage startup based in San Francisco. The concept is a fairly simple one and is initially targeted at smartphones and aftermarket devices - but the opportunity to tap into existing integrated cameras built into cars for parking assist, collision avoidance, lane keeping and other applications is significant.

A growing roster of companies aggregating video data from a wide range of sources has emerged and includes such organizations as Bellingcat (known for its contributions to identifying suspects in the Skripal poisoning incident in the United Kingdom) and Utah-based Banjo, which recently shifted its focus from a media and hedge fund-focused clientele to public safety applications.

No alt text provided for this image

SOURCE: Hayden AI Technologies

The range of applications is immense and growing. A growing number of cities in the U.S. are already offering bounties for citizen enforcers reporting illegally parked or idling vehicles. The range of traffic infractions identifiable via still images or video is broad, but only the tip of the ice berg.

A network of connected vehicle-mounted cameras could be tapped into in real-time by law enforcement in the event of an actual incident in progress, filling in the gaps left by low quality, unreliable, public closed-circuit television cameras. Additionally, vehicle-mounted cameras offer higher resolution and more sophisticated processing.

Further image processing and interpretation will be applied in the cloud by Hayden AI's service - designed as it is to identify events or individuals and communicate directly with local law enforcement. The Hayden AI platform automates the process of video aggregation and analysis acting as a force multiplier for the police.

Imagine such a system integrated with Palantir for optimizing data, image or video collection correlated to predictive crime mapping. That may sound extreme and intrusive, but what if it were to be effective? Would we object? Might it change the concept of Neighborhood Watch?

The opportunity is part of a broader adoption of video generally by law enforcement in the wake of multiple shootings of armed and unarmed black "suspects" in the U.S. Police officers in the U.S. and around the world are increasingly being required to wear body cameras. One of the more clever implementations related to vehicles used by law enforcement is an integration shown by Australian police in a Tesla Motors Model X allwoing body camera video to be displayed in the dashboard of the vehicle:

https://interestingengineering.com/australian-police-buy-tesla-model-x-for-research

Hayden AI offers a means for consumers to monetize their vehicle-mounted cameras, which are normally only relevant to safe operation of the vehicle. One of the key challenges to increasing the penetration of safety systems into the new car market is the cost of those systems and the need to convince consumers/drivers to keep their blind-spot detection and lane keeping systems always on.

Hayden AI provides that "always on" justification and offers a means of mitigating the cost of the system on a new car. The market for parking violations is a multi-billion dollar opportunity, but it is only one source of bounty from real-time reporting of moving and non-moving violations - or even suspects on the run if facial recognition can be legally applied.

Aggregating vehicle collected data, with user consent, for a wide range of other applications is also implicated including traffic, weather, road surface condition, visibility, and traffic incidents. Every application has a price tag and Hayden AI is preparing to serve those opportunities.

It's early days for Hayden. More information is available at www.hayden.ai.

Previous Post: OnStar Rolls Back Free Trials | Next Post: 5G Faith & the Automotive Industry
Leave a comment