LiDARs (Light Detection and Ranging) are sensors that use laser light to measure distances. They create detailed 3D maps of environments by detecting the time it takes for the light to return after hitting an object, enabling precise object detection and classification.

LiDAR uses laser light to measure distances and create high-definition 3D models of the environment, making it effective for precise spatial measurements and less affected by lighting conditions. In contrast, cameras capture images based on visible light, allowing them to recognize colors and shapes but often struggling in low-light or extreme weather conditions.

LiDARs ensure full anonymity by not capturing identifiable data, avoiding unreliable post-processing anonymization, thereby ensuring GDPR compliance.

They enable seamless data fusion from multiple sensors, creating a larger, accurate virtual sensor. Object positioning in a global frame is straightforward and cost-effective.They cover larger areas more cost-efficiently, especially with multiple sensors.

Yes!
LiDAR sensors provide detailed information about all objects in an environment, allowing multiple applications without needing new sensors. For example, you can start with virtual ground loops at an intersection and later use the data for traffic flow tracking, accident alarms, pedestrian light control, car park occupancy, and enforcement of traffic violations with cameras. The data from a few sensors is rich and comprehensive, enabling various applications. Additional sensors can be seamlessly integrated to extend coverage areas. This showcases the power of reliable and consistent 3D object detection.

Point Clouds:The primary output of LiDAR is a point cloud, which consists of millions or billions of individual points in 3D space. Each point represents a location in the environment and is characterized by its XYZ coordinates (latitude, longitude, and altitude) along with additional attributes such as intensity (reflectivity) and sometimes color information.

Depth Information: LiDAR provides precise depth information for each point in the point cloud, allowing for accurate 3D mapping and reconstruction of the environment.

Reflectivity and Intensity: LiDAR sensors often measure the intensity of the returned laser pulse, which can provide information about the reflectivity of surface properties of objects in the environment. This data can be useful for object classification and segmentation.

GDPR compliant:If deployed carefully, it can be ensured that LiDAR sensors do not capture data at a resolution that allows for the unique identification of a person, making them GDPR compliant out of the box.

Perception solutions are systems designed to translate visual 3D data from sensors such as LiDARs, RADARs and cameras into context-aware actionable intelligence. Perception solutions are commonly used in intelligent transportation systems including autonomous vehicles, robotics, security surveillance systems, crowd and queue management, just to name a few.

senseEDGE is a foundational 3D perception solution that works in the real world in any environment. It works with a wide range of types and numbers of sensors, mounted in any kind of positions, in any kind of environmental condition or density of objects - all out of the box.

Under the hood senseEDGE is a collection of highly innovative interlocking algorithms and systems with a core based around proprietary Deep Learning technology. Our technology has been specifically designed for diverse general purpose applications such as ITS, smart spaces and security with an emphasis on scalability to multiple sensors.

By always considering the detection of objects in full context of time and space and the relative capabilities of the individual sensors providing the data, it achieves entirely new levels of detection accuracy and robustness as well as ease of deployment. senseEDGE never removes or filters information provided by sensors, it only adds and contextualizes it with the goal to detect objects earlier, more reliably and with more robustness to challenging conditions. senseEDGE does not require to be specifically set up or tuned to work well in your specific deployment, it works out of the box and continuously tunes itself to the environment and the specific sensor(s) over time in the background. If the deployment conditions, sensor performance or the environment change significantly it adapts immediately with no user intervention to continue to provide the best performance possible.

senseEDGE does not require objects to be physically separated from other objects or their environment. For example it is aware of and tracks the presence of a person sitting on a wheelchair separate from the wheelchair itself. If at some point that person were to get up, it would continue to be tracked just as well as it would individually track two pedestrians leaning on each other and holding hands while walking down the street.

senseEDGE performs well in harsh conditions and applications with lots of expected sensor noise such as heavy rain and fog. This is due to the fact that every measurement point from a sensor, no matter how noisy or unreliable it might be, is being leveraged to extract information from it. Together with the fact that our algorithms have a deep and robust understanding of the types of objects they might encounter and how they would present themselves even in adverse conditions new levels of detection robustness can be achieved.

In doing so senseEDGE shatters the ever present, severe limitations of the many legacy technologies commonly applied to LiDAR perception till this day and elevates the customer value of a LiDAR installation to a new level. We at CRON pride ourselves to have developed a solution that others have deemed impossible.

Our solution has been developed and validated using data from a wide variety of LiDAR makes, models and technologies. Data has been collected in multiple countries spanning 3 continents, time of day, weather and application.
senseEDGE has been exposed to diverse indoor and outdoor environments, from office space, to shops and malls, garages, to busy urban infrastructure to rural roads and even farms and wilderness. It knows how to detect, classify and track a TukTuk weaving through a busy intersection in Central New Delhi just as well as an American 14-wheeler rumbling down an interstate in Oregon.

Running in minutes, not hours:senseEDGE is the only truly plug-and-play solution for LiDAR object detection in the market that uses application and sensor-adaptive perception technology. Our customers report that their setup time is orders of magnitude quicker and easier than other solutions, irrespective of the sensor type being used. The moment the chosen sensors are connected and configured, senseEDGE will start detecting and tracking objects for you, no tweaking or tuning necessary!

Amazing detection quality:senseEDGE’s detection and classification quality and fidelity are a game changer in the industry. We detect and classify a constantly growing range of object classes, with accurate and rich intelligence about each object, even in less than ideal, complex, cluttered and constantly changing environments.

Scale for quality, area or both (“1+1 = 3”): senseEDGE can work very well with a single sensor, but it also allows for the use of more than one LiDAR sensor in a deployment. You can easily mix and match different sensor types to get the best results. In fact, senseEDGE can ‘scale’ additional sensors disproportionately, improving coverage area, redundancy and detection quality at the same time - what we call the “1+1 is 3” effect.

Easy configuration: We do not require a complicated sensor placement with special heights and angles in a deployment. As long as your relevant objects are at least partially visible in the data captured by one or more sensors, senseEDGE will start detecting and tracking them including good estimates of size and location.

Set it and forget it:senseEDGE does not need special attention post deployment. For as long as you don’t want to change the setup, you won’t have to ever touch it again.

Unfazed by adversity and change:senseEDGE keeps working even when conditions change to the worse. Rain, fog, snow, but also big or sudden changes to the structure of the environment or degradations of the sensor due to soiling or partial failures do not faze the system. It continually adapts to the situation and continues to provide the best possible detection results.

The exact number depends on the specifics of the deployment and hardware used, but it can be in the 1000s - 10,000s per deployed unit if necessary.

senseEDGE is uniquely capable of detecting and tracking objects even when they are physically very close. Some examples of supported use cases are: people on wheelchairs, people pushing objects such as cycles or baggage, trucks towing cars, people leaning on cars or dense crowds with people hugging or holding hands.

Deployed correctly for your application, senseEDGE can achieve the specific KPIs your application requires.

Quantitative numbers depend on the specifics of the KPIs/metrics considered, the quality and placement of sensors being used and the complexity of the task. As senseEDGE is being deployed in many diverse applications using diverse sensor qualities, quantities and placements, blank statements that we could make about accuracy such as “up to 99.9% accurate”, while technically true, are misleading and therefore of limited value.

senseEDGE provides best in class precision, recall and classification accuracies for the object classes we support. Our own experience and customer feedback show us that senseEDGE performs better than any other solution we or they have tried to date and their KPI requirements are being met or significantly exceeded.

No, senseEDGE is truly plug-and-play and will work out of the box for all common and most uncommon applications.

We do provide services to add support for unique requirements, such as new or different object classes, very unusual environments or deployment conditions. Please approach our sales team if you think such adaptations are needed. We are constantly growing the range of capabilities of senseEDGE based on customer demand.

Once initially set up, senseEDGE can be moved and redeployed often in seconds with the click of a single button. This allows it to be easily used in semi-permanent installations where the LiDAR sensors might be installed on a mobile unit that is being moved around from time to time.

With select customers CRON AI also allows the integration of senseEDGE into fully mobile, eg. robotic applications. This is a special case that would need bespoke adaptation to your particular use-case.

Yes, we provide a configuration utility and visualiser, which can be accessed through a web browser. This provides the means to set up, configure and monitor the system.

Once the hardware is installed and connected, it will be necessary to configure the sensors, set up the ground plane for each and align the sensors to each other where more than one is used. Additional tools allow zones and regions of interest to be created as required.

Typically this takes just a few minutes for an installation; easy to use graphical tools and automated alignment algorithms are provided.

No, sensors can be safely connected through a switch or router, but the latency and bandwidth of the connection must be suitable to achieve good results.

Yes, we fully support PTP when senseEDGE is used with sensors which support it. PTP is recommended when using more than one sensor.

Yes, senseEDGE can support and synchronize multiple sensors both to increase the area of coverage and to improve point quality and address limited field of view within a smaller area.

Yes, a VPN configuration can be loaded through our visualiser to make remote access easier.

Yes, both polygon and segmented line zones can be created through the visualiser; corresponding occupancy counts (both instantaneous and cumulative) and tracking IDs are included in the output messages.

Yes, regions of interest can be created through the visualiser to include and exclude specific areas from coverage. Regions can also be created for each sensor to e.g. exclude points reflecting from a specific surface which might otherwise lead to false detections.

Yes, both point cloud and object data can be recorded through the visualiser for subsequent playback. Additionally tools are provided to record and download packet captures from individual sensors.

We currently offer object data in JSON or Flatbuffers (binary) format. Full details are available on request.

Object data is transmitted over a websocket hosted on senseEDGE. This makes it easy to connect a client in order to consume the data, and integrations are available with a number of external dashboard and integration systems. Please contact us for details.