Six 3D Vision Solutions from MRDVS

Introduction: Is your warehouse still relying on binocular vision for obstacle avoidance? Still counting inventory by hand? This article shows how MRDVS uses 3D vision solutions to connect every link of the warehousing and logistics chain — from inbound to outbound — enabling visualized goods, intelligent equipment, and digital management.

Scenario	Solution	Core Product
Inbound	Large-item volume measurement	M10 dToF Camera
Storage	Storage slot status detection	S10 Pro dToF Camera
Depalletizing	Soft-package recognition & depalletizing guidance	M10 dToF Camera
Transport	AGV obstacle avoidance + navigation	S10 / S10 Ultra dToF Camera
Forklift Digitalization	High-precision positioning + pallet recognition	V2 Pro / M4-Mega
Safety	Human safety protection	S10 Pro

Let’s break down MYVI’s vision solutions across these six scenarios one by one ↓

Solution 1: Multi-Camera Stitched Volume Measurement — Digitalization from the Moment of Inbound

Pain Points

Traditional manual measurement of large items is time-consuming, labor-intensive, and error-prone. Laser profilers require uniform-speed scanning and cannot measure static objects, while light-curtain devices have limited accuracy on irregularly shaped goods.

MRDVS’ Solution

An array of multiple M10 cameras captures real-time 3D point clouds and RGB texture data for goods, automatically calculating volume and syncing with the WMS. It covers a measurement range of 0.3–8 m and supports precise measurement of complex goods, including large irregular items and highly reflective surfaces:

• Centimeter-level volume calculation: dToF technology directly measures laser time-of-flight, delivering far greater precision than manual or light-curtain estimation.
• Simultaneous multi-target measurement: a single frame captures 3D data for multiple objects at once, supporting batch inbound processing.
• Adaptable to complex shapes: irregular items, highly reflective surfaces, and film-wrapped goods can all be measured precisely.

Core Advantages

Operates reliably indoors and outdoors, even in strong light of up to 100kLux, enabling static, fast, and batch precision measurement — truly achieving “the moment goods enter the frame, they enter the digital inventory.”

Solution 2: Storage Slot Status Recognition — Keeping Every Rack Position in Sync

Pain Points

Warehouses frequently experience temporary manual stacking, pallet misalignment, and mixed storage of goods. A single-point LiDAR may overlook gaps and cause misjudgment, while pure RGB cameras lack depth information and cannot plan stacking tasks.

MRDVS’ Solution

Built on the S10 Pro RGB-D depth camera, this solution fuses 3D point clouds with RGB information to deliver full-dimensional storage-slot detection, accurately determining occupancy status and placement compliance:

• Slot occupancy detection: depth sensing determines whether a slot holds goods, unaffected by lighting changes.
• Compliance verification: identifies non-compliant placement such as pallet misalignment, tilting, or boundary overruns.
• Real-time data output: syncs with the WMS via a standard interface to drive dynamic slot-status refreshes.

Core Advantages

An integrated hardware-software design with onboard compute eliminates the need for an external industrial PC. Each storage slot can be deployed independently, greatly reducing gap-related misjudgments and lighting interference.

Solution 3: 3D Vision Soft-Package Depalletizing — Guiding Robotic Arms to Grip Precisely

Pain Points

Soft-package and sack stacks in industries such as plastic pellets, fertilizer, and food processing are prone to deformation. Traditional robotic arms struggle to handle them, and manual depalletizing is slow and prone to damaging the packages.

MRDVS’ Solution

Based on the M10-series dToF 3D camera paired with a compute module, RGB-D depth sensing is fused with AI depalletizing algorithms to accurately recognize irregular objects such as soft packages and sacks:

• Adaptive stack-shape recognition: no pre-training required; supports complex structures such as six-pack, five-pack, and interlocking stacking patterns.
• Precise grasp guidance: combines depth point clouds with RGB texture to output accurate grasp poses, reducing package breakage.
• Resistant to ambient lighting interference: maintains stable recognition performance under both strong and weak lighting conditions.

Core Advantages

The deeply optimized M10 + compute-module combination shortens deployment from weeks to days, is compatible with mainstream robot communication protocols, and lets a single system handle multiple stack shapes.

Solution 4: 3D Vision Obstacle Avoidance — AGVs’ Safety Eyes

Pain Points

2D LiDAR can only scan a single plane, largely failing to detect low-lying or overhanging obstacles; pure RGB solutions fail in low light, directly leading to collisions in complex warehouse environments.

MRDVS’ Solution

Based on the S-series dToF RGB-D camera, 3D point clouds are fused with RGB texture, and built-in obstacle-avoidance algorithms deliver full-space obstacle detection:

• Full-space obstacle awareness: covers a range of 0.3–8 m, precisely detecting low-lying and overhanging obstacles.
• Dynamic avoidance strategy: flexibly configure warning, deceleration, and stop zones to adapt to complex warehouse traffic patterns.
• Multi-unit coordinated deployment: minimal interference between units, suitable for safe and efficient operation of large-scale AGV fleets.

Core Advantages

A built-in high-performance SoC and avoidance algorithm output detection results directly, with no external industrial PC required — achieving proactive avoidance that “understands before it decides,” truly serving as the warehouse’s “safety eyes.”

Solution 5: Fused Forklift Positioning & Navigation — Precise Positioning, Traceable Paths

Pain Points

Manually operated forklifts are a “black box” in the warehouse — managers struggle to know their location or route. Traditional positioning solutions require laying down large numbers of base stations, resulting in long deployment cycles and high cost.

MRDVS’ Solution

Based on the V2 Pro positioning and navigation module, which fuses 2D LiDAR, top-view vision, and IMU inertial navigation, multi-sensor SLAM fusion delivers high-precision positioning and navigation:

• Centimeter-level precise positioning: ±3 cm repeat positioning accuracy, with real-time output of position and heading angle.
• Rapid natural-feature mapping: one-click mapping by scanning ceiling and wall features, with no need to lay down base stations.
• IMU-assisted continuous positioning: inertial navigation automatically takes over when vision is obstructed, ensuring uninterrupted positioning in complex environments.

Core Advantages

Built-in 6 TOPS of compute power and a wireless module eliminate the need for an external industrial PC. Compared with traditional solutions, deployment cost is lower and the cycle is shorter — turning forklifts from a “black box” into fully trackable and navigable assets.

Solution 6: Human Safety Protection — An Electronic Fence for the Warehouse

Pain Points

In restricted areas such as depalletizing stations and robotic-arm work zones, accidental entry by personnel can lead to serious safety incidents. Traditional safety light curtains have blind spots and cannot distinguish people from objects.

MRDVS’ Solution

Based on the M4 Pro iToF 3D camera and human-recognition software, AI algorithms monitor personnel movement within 3D space in real time, precisely identifying intrusions into restricted areas:

• Real-time 3D spatial monitoring: high-resolution depth plus RGB precisely captures personnel movement, with multi-camera coverage eliminating blind spots.
• Precise AI boundary-crossing detection: deep-learning human-recognition algorithms respond in under 100 ms.
• Distinguishes people from objects: RGB-D fused sensing differentiates people from goods, greatly reducing false alarms.

Core Advantages

The M4 Pro’s built-in high-performance compute eliminates the need for an external industrial PC. Compared with traditional light curtains, it achieves 360-degree monitoring with no blind spots, building a reliable “electronic fence” for warehouse areas.

Product Overview: Vision Hardware Matched to Different Scenarios

S-Series dToF RGB-D Cameras

Covering models such as S10 / S10 Ultra / S11, suited to medium- and long-range scenarios such as obstacle avoidance, navigation, and volume measurement, with strong resistance to interference.

M-Series High-Resolution RGB-D Cameras

Covering models such as M4 / M4 Pro / M4-G, suited to close-range, high-precision scenarios such as depalletizing, pallet recognition, and human safety protection.

V2 Pro Positioning & Navigation Module

A multi-sensor SLAM navigation controller fusing vision, LiDAR, and IMU, suited to positioning and navigation for manually operated forklifts and AGVs.

Warehousing and logistics automation depends heavily on professional vision solutions. Whichever system’s machines see more clearly and understand more deeply will have the safer, more efficient, and more intelligent automation system.

With its full-stack capability spanning 3D vision, AI algorithms, and integrated hardware-software design, MYVI provides warehousing and logistics enterprises with complete capabilities — from single-point solutions to fully integrated platforms.