Sony FCB camera block
In the maze of dense shelves in warehousing and logistics, the bustling crowds in retail stores, and the strong light interference during bridge inspections, traditional LiDAR often fails due to missed detection of low-reflectivity targets, dynamic occlusion interference, or overload of ambient light.
The SONY AS-DT1 lidar, with its dToF+SPAD technical architecture and multi-dimensional anti-interference mechanism, demonstrates outstanding adaptability in complex scenarios and redefines the boundaries of high-precision measurement. This article will deeply analyze how it breaks through core challenges such as dynamic occlusion, low-light environments, and multi-path reflection, and becomes a versatile expert in the era of Industry 4.0.
I. Precise Perception in dynamic Occlusion environments
In the operation scenarios of warehouse robots, dynamic obstructions between shelves (such as moving cargo boxes and personnel passing through) can cause traditional LiDAR to have blind spots in distance measurement or data distortion.
The SONY AS-DT1 achieves penetrating perception through the following technologies:
Detection of suspended objects
SPAD single photon detection: It captures weak reflected light (such as a nylon rope suspended and wrapped) through the avalanche multiplier effect. Even if an object only blocks 5% of the laser path, its distance can still be measured.
Three-dimensional point cloud layering: The dense ranging of 576 points per frame (24×24 array) can distinguish the shelf partition from the suspended objects above, preventing the robotic arm from mistakenly grasping.
Ii. Resistance to interference from personnel movement
1.Dynamic target filtering algorithm: In retail stores, the instantaneous interference caused by pedestrian movement is eliminated through the inter-frame difference method to lock the stable distance between the goods on the shelves and the fixed facilities.
Multipath reflection suppression: dToF technology directly measures the round-trip time of photons, avoiding the error accumulation caused by multipath reflection in traditional triangulation methods, ensuring that the repeatability accuracy of shelf height measurement is less than 1 cm.
2. Stable output under extreme lighting conditions
From the 100,000 lux strong light outdoors in summer to the low-illumination environment in the warehouse, the AS-DT1 achieves reliable all-weather operation through dual-mode light emphasis nodes and SPAD sensitivity optimization:
Strong light suppression technology
940 nm band anti-glare: Select the near-infrared and invisible light band to avoid strong interference in the solar spectrum (such as water surface reflection and glass curtain wall glare), and maintain an accuracy of ±5 cm when measuring 20 meters outdoors.
Automatic Gain Control (AGC) : Dynamically adjust the laser emission power based on the intensity of ambient light to prevent sensor saturation failure in strong light environments.
2. Weak light enhancement scheme
SPAD single-photon sensitivity: In an underground parking lot without auxiliary lighting (< 1 lux), the reflected signals from metal shelves can still be detected, with a ranging error of < 8 cm (at a distance of 10 meters).
Iii. Intelligent Analysis of Multi-object Mixed Scenes
Facing high-density target scenarios such AS retail shelves (dense goods) and logistics sorting areas (package stacking), SONY AS-DT1 achieves precise decoupling through spatial segmentation algorithms and multi-level data fusion:
Micron-level distance resolution
0.25mm point spacing: A single frame of 576 ranging points can cover an area of 0.576 square meters, distinguishing adjacent shelf partitions (with a spacing of 3 cm) from the edge of the cargo box, thus avoiding the risk of collision in the path planning of warehouse robots.
2. Target classification and priority marking
Ai-assisted point cloud segmentation: Built-in deep learning chips cluster point clouds, automatically identify categories such as “fixed shelves”, “mobile personnel”, and “suspended packages”, and give priority to processing dynamic obstacle data.
ROI area focus: During bridge inspection, it is possible to set up a key focus area for bridge pier cracks (such as at a distance of 10-15 meters), and increase the distance measurement frequency of this area to 60 Hz to ensure the real-time performance of deformation detection.
Iv. Industry Application Breakthrough Cases
1. Stereoscopic warehouse navigation
Scene pain point: The shelf height exceeds 8 meters, and traditional LiDAR cannot cover the top layer due to insufficient ranging. Dense cargo containers cause multipath reflection interference.
AS-DT1 solution: Indoor 40-meter ranging + anti-multipath algorithm, achieving three-dimensional modeling of shelves and dynamic optimization of AGV paths, significantly improving the efficiency of inventory taking.
2. Highway inspection drones
Scene pain point: Under strong backlight, traditional LiDAR distance measurement decays rapidly, and the missed detection rate of crack detection exceeds 30%.
AS-DT1 scheme: Outdoor 20-meter strong light ranging +SPAD low-light enhancement, combined with the AI crack recognition algorithm, the detection accuracy reaches ±5 cm, and the missed detection rate is effectively suppressed.
The SONY AS-DT1 is not only a sensor, but also an infrastructure-level product for perceiving complex scenes. Its technological breakthrough marks the evolution of lidar from a “tool attribute” to an “intelligent perception hub”, opening a new chapter of full-scenario digitalization for the industrial Internet of Things.