Case study: An Ninh Số deploys 3D AI volume measurement cameras for logistics

The AI 3D industrial volume measurement camera is a system integrating Time-of-Flight (ToF), LiDAR, and Stereo Vision technologies, enabling automatic measurement of cargo dimensions in logistics lines with accuracy up to ±2 mm. In Vietnam, large distribution centers are transitioning from manual measurement to this solution to reduce errors and increase throughput. Việt Đức Trí Group, through its member brand An Ninh Số, has just completed deploying this system for a 5,000 m² logistics center in Bình Dương.

Customer: Bình Dương Logistics Center (name withheld upon request) | Industry: Logistics & Supply Chain | Scale: 5,000 m², processing 5,000 packages/day | Location: VSIP II Industrial Park, Bình Dương

Customer Challenges

• 12% manual measurement errors: Warehouse staff measure each package’s dimensions using a tape measure, with an average deviation rate of 12% compared to actual dimensions, causing cost losses in transportation and storage.
• Slow processing time: Each package takes 3-5 minutes to measure and manually enter data, resulting in a total time of 250-400 minutes/day for 5,000 packages.
• Lack of digitized data: Cargo dimensions are only recorded on paper, not synchronized with the WMS system, making it difficult to optimize pallet space and plan transportation.
• High labor costs: 3 full-time employees work on the measurement and inspection stage, costing approximately 25-30 million VND/month.

An Ninh Số Solution Deployment

An Ninh Số deploys a 3D AI volumetric measurement camera system integrating three core technologies: ToF for precise distance measurement, LiDAR for 3D spatial scanning, and Stereo Vision for object edge recognition. The system is installed on the main conveyor belt, automatically measuring the dimensions of each package within 0.5 seconds.

1. Survey & Design (2 weeks): Measure installation space, conveyor speed (0.8 m/s), maximum package dimensions (1.2m x 1.2m x 1.5m). Design camera mounting frame with 3 sensors: Intel RealSense D455 (Stereo Vision), VL53L5CX (ToF), and Ouster OS0-64 (LiDAR).
2. Hardware Integration (1 week): Install cameras on aluminum profile frame, connect to conveyor PLC via Modbus TCP. Central processing unit uses NVIDIA Jetson Orin NX with 16 GB RAM.
3. Software Development (3 weeks): Write data fusion algorithm from 3 sensors, calibrate error to ±1.5 mm. Integrate API with customer’s WMS system via RESTful, automatically record dimensions into database.
4. Testing & Calibration (1 week): Run test on 500 sample packages, cross-check results with manual measurement. Average error 1.8 mm, meeting technical requirement of ±2 mm.
5. Handover & Training (3 days): Train 2 system operators, hand over technical documentation. Provide 24/7 remote support for the first 30 days.

Achieved Results

• Error rate reduced from 12% to 0.5% — measurement accuracy reaches 99.5% compared to actual dimensions.
• Processing time reduced by 90% — from 3-5 minutes/package down to 0.5 seconds/package, saving 45 minutes/day.
• Labor cost savings of 25 million VND/month — reduced from 3 employees to 1 supervisor.
• Expected ROI of 18 months — with a total investment of 380 million VND (hardware + software + installation), saving 300 million VND/year in operating costs.

“Previously, my employees spent the whole morning measuring and entering package dimensions. Now, the system does it automatically in minutes, with data synced directly into the WMS. Errors are nearly zero. An Ninh Số helped us digitize the most critical step in logistics.”

— Mr. Nguyen Van Hung, Operations Director, Binh Duong Logistics Center

Lessons Learned and Application for Other Businesses

This project demonstrates three key success factors: choosing the right sensor fusion technology, deep integration with the existing WMS system, and systematic operational staff training. Logistics businesses with a throughput of over 2,000 packages/day should consider investing in a similar solution to achieve ROI within 24 months.

• Cannot rely on just one type of sensor: ToF is good for distance measurement but poor at edge detection; Stereo Vision detects well but is noisy in low-light environments. Fusing 3 technologies is optimal.
• WMS integration is mandatory: Without data synchronization, measurement alone reduces value by 50%. RESTful API facilitates easy connection with most popular WMS systems.
• Low operating costs: The system consumes 45W, with electricity costs < 500,000 VND/month. Periodic maintenance every 6 months costs 5-7 million VND per service.

Frequently Asked Questions

What is the maximum size a 3D AI volume measurement camera can measure?

The system deployed by An Ninh Số can measure packages with maximum dimensions of 1.2m x 1.2m x 1.5m (length x width x height). For larger packages, additional cameras or longer-range LiDAR sensors are required. Accuracy is ±2 mm under standard warehouse lighting conditions (300-500 lux).

How much does it cost to deploy an industrial volume measurement camera system?

The average cost for one measurement station (including 3 sensors, a Jetson processor, mounting frame, software, and installation) ranges from 350-450 million VND, depending on the complexity of the existing infrastructure. With a throughput of 5,000 packages/day, ROI is typically achieved within 18-24 months due to labor savings and error reduction.

Can the system integrate with my current WMS?

Yes. An Ninh Số provides standard RESTful APIs that can integrate with most popular WMS systems such as SAP EWM, Oracle WMS, or local solutions. The integration process takes 1-2 weeks, including input/output data testing.

Do I need to modify my existing conveyor belt when installing the volume measurement camera?

No. The system is designed as an independent mounting frame installed above the conveyor belt without interfering with the existing structure. The conveyor belt only needs a stable speed below 1.2 m/s and a flat surface to prevent packages from tilting while passing through the measurement zone.

Does An Ninh Số support system maintenance and upgrades?

An Ninh Số provides a periodic maintenance package every 6 months, including sensor cleaning, accuracy calibration, and firmware updates. The maintenance cost is 5-7 million VND per session. Additionally, customers can upgrade to higher-resolution sensors or add more cameras as throughput increases.

Do you want to deploy a similar AI 3D volume measurement camera solution for your warehouse? Contact Việt Đức Trí Group immediately via hotline 0935 295 337 or email [email protected] for free consultation and survey.