This solution is to count vehicles by the direction or turns they take even when from a shared lane. The number in the arrow denotes the count of the vehicles in the direction of the arrow. The count is made after the vehicle is traversed the path. This solution deals with occlusion, and multiple methods to accurately detect the entry and exit points of the vehicle to make the turn count accurate. This helps traffic engineers understand the traffic demand and freight ratio on individual corridors, which can be used to design better intersection signal timing plans and apply other traffic congestion mitigation strategies when necessary.

This solution computes the instantaneous velocity of vehicles from any surveillance camera night or day without the need for camera specification, its height, or road marking. This is delivered in real-time using Nvidia Nano or from the cloud. This solution can be used in college campuses, industrial and residential complexes, school crossing, or just about any intersection such as in this video. Data captured can be used to filter most frequent offenders, and to take remedial steps if found to be during certain days of the week or hours of the day.

This solution can detect red light violations by any vehicle or two-wheeler. The solution looks at the status of the traffic light and monitors the approaching vehicles. If the vehicle were to cross the intersection with the traffic light on red an alert identifying the vehicle is made. This solution can also calculate the real-time velocity of all the vehicles (refer INSTANTANEOUS VELOCITY CALCULATION video). The solution can also support Automatic license plate recognition depending on the city/country.

This solution can take any video feed and map that to an aerial view and project it using google earth. The tracking ID on video is the same as the one on google earth. This model can deliver robust and accurate tracking for perimeter, security in stadiums, airports, commercial locations. They can be used in a retail setting to count and track movement with duration. This solution can be used in elder care homes or nursing rooms to track the movement of patients in private rooms and common areas to detect falls and alert the right healthcare team while offering privacy without the use of real images. This solution works in real-time using Nvidia Nano.

This solution can be used at any indoor or outdoor parking lot. It provides metering only using vision. Look at the 30th second when a car enters the parking stall. The parking time starts automatically. We have set the max parking time as 90 mins. around the second 45, one of the car previously parked reaches the max time limit and the violation is highlighted. All this data can also be shared as alerts. Using this solution, the traditional approach of using chalks to mark the tire and with that the time parked is no longer needed. The parking attended can be notified to approach the car at the right time. This solution can also support car count, empty parking spot count, and suspicious activity making is a cost-effective, safe, and environmentally friendly solution.

This solution is designed for indoor and outdoor parking lots. Using video from CCTV the software can locate any parking spot that is visible to the camera. The count can be posted on signs and using additional software the location of the spot can also be shared. This approach using the camera eliminates the need to put sensors under the ground and having to maintain them. The added benefit of using a network of cameras include safety and security for owners, parking meter (refer PARKING ASSISTANT WITH METER video) for operators, and alert suspicious activity providing peace of mind to patrons.

This solution lets you draw a threshold, in this case, the pink polygon, and count people crossing the virtual threshold. The polygon can be of any shape or size, and additional filters such as pose or activity can be included to further. narrow down the count. All this can work from any video source. We can also trace the movement on an aerial view by projecting it on google earth. We can also compute the instantaneous velocity of each runner. Please browse other. videos with this feature. This solution can help with public safety in open and closed areas by providing real-time count and tracking.

This solution uses computer vision from a dash-mounted camera to detect other vehicles on the road to compute relative velocity, distance, and real speed to calculate the risk of collision from tailgating. The solution is configured for safe distance based on the type of vehicle. This way the following distance for a fully loaded trailer is more than a family sedan. ADAS systems that use Gyro and accelerator cannot detect this driving behavior unless using Radar or Lidars. The camera can be placed on the dashboard or mounted on a windscreen for good viewing.

This solution uses computer vision to detect distracted driving without the need for secondary sensors such as cellphone-based gyro, ODB devices commonly used. The model detects the driver is not keeping up with the traffic to warn the driver. The camera can be placed on the dashboard or mounted on a windscreen for good viewing.

Like ADAS for motorcycles. This solution uses GoPro video mounted on a motorcycle to detect proximity to moving objects on the road, compute the relative velocity with them to warn riders. The solution does not use lane or road markings making it versatile in all types of road surface conditions including offroad. The solution is efficient to run on Nvidia Nano and can be used as a training aid in addition to ADAS for motorcycles. The camera can be mounted on the fairing or handlebars.

Like ADAS for motorcycles. This solution uses GoPro video mounted on a motorcycle to detect proximity to moving objects on the road, compute the relative velocity with them to warn riders. The solution does not use lane or road markings making it versatile in all types of road surface conditions including offroad. The solution is efficient to run on Nvidia Nano and can be used as a training aid in addition to ADAS for motorcycles. The camera can be mounted on the fairing or handlebars.

Like ADAS for motorcycles. This solution uses GoPro video mounted on a motorcycle to detect proximity to moving objects on the road, compute the relative velocity with them to warn riders. The solution does not use lane or road markings making it versatile in all types of road surface conditions including offroad. The solution is efficient to run on Nvidia Nano and can be used as a training aid in addition to ADAS for motorcycles. The camera can be mounted on the fairing or handlebars.

Video object mapping to aerial view – Objects in any video can be projected on to aerial view on google earth. This can be done using any surveillance camera or a network of surveillance cameras. For this video, we have one video stream processed in real-time using Nvidia Nano. The tracking ID on video is the same as the one on google earth. This video can be enhanced using our tracker for occlusion management.

This solution can be used to detect falls from any video source including for people with walkers. This solution can be used with CCTV units in nursing homes, hospitals including private rooms, and be set up to only provide alerts and thereby protecting the privacy of the individual. The solution does not require the user to have any wearable and works purely using vision. We have several of these videos on our YouTube channel so please feel free to check it out.

This video demonstrates the advanced occlusion management using our solution. Occlusion occurs when two or more objects come too close and seemingly appear to merge. This can also happen when objects change direction or occluded behind traffic lights. Object tracking often wrongly tracks the occluded objects. The segment labeled “Before” is using traditional means, and “after” uses our robust tracker. The detector software in both ‘before’ and ‘after’ are the same but our tracker used in ‘after’ can manage occlusion better.

This video demonstrates the advanced occlusion management using our solution. Occlusion occurs when two or more objects come too close and seemingly appear to merge. This can also happen when objects change direction or occluded behind traffic lights. Object tracking often wrongly tracks the occluded objects. The segment labeled “Before” is using traditional means, and “after” uses our robust tracker. The detector software in both ‘before’ and ‘after’ are the same but our tracker used in ‘after’ can manage occlusion better.