Changing the Thinking on Artificial Perception
We pair a well-characterized, high-performing vision sensor with our proprietary analysis engine, comprising algorithms developed in-house based on our broad experience in vision systems, machine learning and deep learning.
Inspired by how the brain processes information, Kelzal’s system collects pertinent data to identify people, objects or obstacles differently than conventional cameras. As a result, traditional image processing techniques and established algorithms are not suitable for this application. We’ve taken a fresh approach to the data in order to develop unique algorithms which take advantage of the high temporal resolution of the sensor as well as its broad dynamic range. Use of those characteristics results in earlier recognition at longer range and under diverse light conditions.
Employing a passive sensor means that the mobile robot is not burdened with heavy, power-consuming EM transmission equipment. With no need for time-consuming frame grabs and compute-intensive image analysis, data gathering is faster and more precise. Coupled with the high temporal resolution of the sensor and high speed algorithms, the result is less time from acquisition to evasive maneuver at every step of the process.
Kelzal’s system brings together proven advanced technology sensor hardware with unique proprietary data analysis and detection for a system ideal for modern mobile robots.
- Low-power—better battery endurance
- High detect and recognition rates—enhanced safety
- Accommodates diverse light conditions—Systems can operate in a wider range of environments and more hours of the day
- Rapid detection—more time to initiate effective actions including raising alerts
- Compact and Light weight—installs anywhere