The rapid technology advancements in Automobile Industry encourage design engineers for latest innovations in vehicle security, autonomous vehicle and in vehicle infotainment along with BMS. E- con Systems a pioneering company dedicated to revolutionizing the automotive industry with cutting-edge embedded camera solutions for autonomous vehicles. Their innovative technology seamlessly integrates embedded vision, real-time image processing, and edge computing, resulting in unparalleled performance, intelligence, and safety.
By harnessing this advanced solution, their cameras achieve swift object recognition, precise facial detection, and accelerated image processing, enabling faster decision-making, automated analysis, instant anomaly detection, and highly efficient operations crucial for autonomous vehicles’ success.
In conversation with Electronics Era Mr Maharajan Veerabahu, Co-Founder and Vice-President at e-con Systems talks about their cutting-edge embedded camera solutions for autonomous vehicles, their vision for Industry 4.0, embedded technology and their future prospects for automotive Industry.
“In the context of patrol robots for security and surveillance, e-con Systems’ cameras facilitate autonomous navigation, real-time theft and intrusion detection, remote surveillance, and facial recognition.” Mr Maharajan Veerabahu, Co-Founder and Vice-President at e-con Systems
Electronics Era: Explain in brief e-con Systems as a company, its achievements and product portfolio?
Maharajan Veerabahu: e-con Systems has been a pioneer in the embedded vision space; designing, developing and manufacturing custom and off-the-shelf camera solutions since 2003. With a team of 450+ extremely skilled core engineers, our products are currently embedded in over 350 customer products. So far, we have shipped over 2 million cameras to the United States, Europe, Japan, South Korea and many more countries. Our cameras are suitable for applications such as autonomous mobile robots, smart agricultural devices, medical diagnostic systems, smart checkouts/carts, sports broadcasting systems, industrial handhelds, drones, biometric systems, etc.Our wide portfolio of products includes MIPI camera modules, GMSL cameras, USB 3.1 Gen 1 cameras, TOF cameras, stereo cameras etc. e-con offers a wide variety of cameras with low light performance, HDR, global shutter, etc. These cameras range from a resolution of 2MP up to 20MP.
We are also powered by a strong partner ecosystem to offer end-to-end vision solutions, including sensor partners, ISP partners, carrier board partners, etc. What sets e-con Systems apart is our deep expertise in building customized product designs while ensuring rapid prototyping and custom modifications in camera hardware and software, including form factor modifications, ISP tuning, carrier board development, lens calibration, and much more. By empowering machines to see, e-con Systems looks to create a world where humans have enriching life experiences so that they can make the world better.
Electronics Era: Explore the role of deep learning in driving the growth of embedded vision technologies and its impact on various industries?
Maharajan Veerabahu: Deep learning has played a significant role in driving the growth of embedded vision technologies, and its impact on various industries has been substantial. Embedded vision refers to the integration of computer vision capabilities into various devices and systems, enabling them to interpret and understand visual information from the surrounding environment.
Traditionally, before deep learning, embedded vision was dependent on computer vision algorithms alone. Now, with Deep learning, a lot of the embedded vision capabilities have improved substantially. Developers use a combination of CV algorithms and DL models to achieve better accuracy, improve on capability of perception, etc.
Deep learning has impacted embedded vision in many industries like medical devices & diagnostic devices, automation robotics in industries, security and surveillance in the form of face and people detection/recognition, retail and customer analytics, smart traffic and cities.
Electronics Era: Highlight the transformation of industries like manufacturing and robotics through embedded vision applications, including obstacle identification and route planning?
Maharajan Veerabahu: e-con Systems offers a comprehensive product portfolio of embedded cameras tailored to empower robots and autonomous mobile robots (AMRs) across various industries, including Industry 4.0, industrial automation, agriculture, and patrol robots. These camera solutions enhance the vision and capabilities of these robotic systems, enabling them to perform a wide range of functions in diverse environments.
For Industry 4.0 and industrial automation, e-con Systems provides cameras such as the STURDeCAM25, NileCAM25, and STURDeCAM20. These cameras enable depth sensing, object identification, barcode reading, obstacle detection, and other vision-based capabilities crucial for advanced industrial applications. In the agricultural sector, e-con Systems offers a range of cameras suitable for autonomous tractors, agricultural drones, and farming equipment. These cameras provide high dynamic range, global shutter, NIR sensitivity, and 3D capabilities. Key models include the DepthVista, NileCAM81_CUOAGX, and e-CAM82_CUOAGX. Delivery robots in last-mile logistics benefit from e-con Systems’ product portfolio, including USB 3.0 Cameras, MIPI Cameras, GMSL Cameras, and more. These cameras support navigation, obstacle detection, package tracking, and safety features for delivery robots.
In the context of patrol robots for security and surveillance, e-con Systems’ cameras facilitate autonomous navigation, real-time theft and intrusion detection, remote surveillance, and facial recognition. These cameras come in various forms, including GMSL2, ToF, stereo, and global shutter cameras.
Electronics Era: Discuss the shift towards decentralized deep learning with embedded vision and how it benefits real-time processing and decision-making?
Maharajan Veerabahu: Deep learning on the edge has made a lot of applications possible today. An edge based deep learning implementation has a lot of benefits like
- Real time & Latency – The decision making is closer to the action and can be done in real time rather than waiting on a cloud, connectivity and other variables. This is important for applications like autonomous cars or industrial robotics
- Bandwidth – If DL has to be done on the cloud, then the video stream has to be uploaded to the cloud which is a huge data hog on the bandwidth. With edge DL, all processing is done locally without the need to send gigabytes of data to a server. This is particularly useful in smart city, security implementations
- Privacy – Video data is not streamed outside which gives improved privacy and data security
- Off line operation – Off line operation is possible without the need of a network
The whole scenario of Industrial automation, autonomous driving, surveillance, smart city is dependent on quick and real time actions and deep learning on the edge makes that possible.
Electronics Era: Explore how embedded vision is driving the Industry 4.0 revolution by enhancing automation, quality control, and efficiency in manufacturing and industrial processes?
Maharajan Veerabahu: Embedded vision is a transformative technology at the heart of the Industry 4.0 revolution. It empowers robots and industrial systems with advanced visual capabilities, enhancing automation, quality control, and efficiency in manufacturing and industrial processes. Embedded vision is driving Industry 4.0 by enabling robots to excel in various industrial use cases:
- Cutting and Sawing: Robots equipped with embedded vision accurately assess object orientation, ensuring precise cutting and sawing while enhancing safety through proximity detection.
- Material Handling, Picking, Packaging, and Placing: Embedded vision automated material handling tasks, enabling object identification, barcode reading, and obstacle detection, leading to efficient and error-minimized operations.
- Non-Destructive Testing (NDT): Robotic vision is employed for non-destructive testing across industries to assess product quality and reduce defects, with data-driven insights optimizing production and warehouse operations.
- Automated Dimensioning: Automated dimensioning systems utilize cameras to derive precise object dimensions, eliminating manual inaccuracies, and enhancing supply chain and warehouse efficiency.
- Machine Tending: Embedded vision enhances machine tending by enabling robots to load and unload materials with accuracy, traditionally performed by humans, leading to heightened productivity.
- Remote Warehouse Management: Telepresence robots, controlled remotely, use embedded vision for navigation, object detection, and positioning in warehouse management, even without human presence, addressing the need for remote work environments.
Embedded vision serves as the cornerstone of Industry 4.0, empowering robots to execute tasks with precision, safety, and speed. This technology propels automation, quality control, and efficiency, making Industry 4.0 a reality with smarter, autonomous systems.
Electronics Era: According to you, what are the future trends you are looking for in embedded vision technology?
Maharajan Veerabahu: The future looks exciting for embedded vision technology as I feel that we are just getting started and there is scope to implement so many things. Embedded vision is giving eyes and perception to machines and which machine will not want to have this capability. Every industry has the potential to be impacted in a large manner by embedded vision and that will drive the maturity and improvements in embedded tech space as well.
To be particular, I expect the following tech trends in embedded vision
- A more seamless integration of AI and Vision – With more advanced AI models and advancements in processing power at the edge, we can expect more sophistication in what we can do at the edge leading to a more seamless AI and Vision integration solving real life problems
- Vision AI chipsets – There is a huge possibility of low power, application specific chipsets which can do specific vision AI processing really well. This will lead to lower cost and size and thereby enabling more embedded vision applications to become reality.
- Robustness and Reliability – More implementations will lead to the tech becoming stronger and mature.
- Solutions to privacy problems – Anything that comes with a vision system comes with a privacy concern. This will be solved as there is a lot of work going on in this area.
- Better AI and Human coexistence – Humans are learning to live with AI and better regulations and ethical codes will be established as we progress which will lead to greater adoption and acceptance as well.
