One of the most common questions we get from OEMs is deceptively simple: "Should I use radar, a camera or lidar?" The honest answer is that these are not interchangeable parts competing for the same slot. They measure different things well, and a well-engineered system usually uses more than one.
What each sensor actually measures
A camera captures reflected visible light and infers scene semantics — lane lines, traffic signs, obstacles by shape. Lidar fires laser pulses and builds a precise 3D point cloud. Millimetre-wave radar emits a radio wave at 24, 77 or 80 GHz and measures range, velocity and angle from the returned echo, directly and simultaneously.
The critical distinction is velocity. Radar measures radial speed from the Doppler shift with every detection — no frame differencing, no tracking guesswork. A camera must see two frames and compute motion; lidar must do the same. When closing speed matters (forward collision warning, drone approach), radar answers in one measurement.
Side-by-side comparison
| Attribute | mmWave radar (77 GHz) | Camera | Lidar |
|---|---|---|---|
| Range (typical industrial module) | 0.2–280 m | Depends on lens/light, ~50–200 m | 50–200 m |
| Rain / fog / dust / darkness | Unaffected | Degrades sharply | Degrades in fog and dust |
| Direct velocity measurement | Yes (Doppler) | No (inferred) | No (inferred) |
| Angular resolution | Moderate (improves with 4D imaging) | High | Very high |
| Object classification | Limited (RCS-based) | Excellent (semantic) | Good (shape-based) |
| Cost at volume | Low to moderate | Low | High |
| Compute requirement | Low | High (vision pipeline) | Moderate to high |
| Power draw | <1–3 W typical | 1–5 W | 5–20 W |
Where radar wins outright
1. Agriculture and off-highway
A combine harvester works in dust and crop residue, often before sunrise. A camera's coverage collapses in these conditions. Our L5261 radar is IP69K-rated and detects people and obstacles at 0.2–80 m in exactly this environment.
2. Drone obstacle avoidance and terrain following
Drones fly into tree lines, power lines and birds in changing light. Radar holds 0.1–120 m detection at ±3 cm with a 120°×60° field of view in the UAV77-O, independent of sunlight. For higher scene detail, our XD70 4D imaging radar adds elevation and 20,000 points per second.
3. Forward collision and adaptive cruise
When a truck is closing on a vehicle ahead, the radar gives range and closing speed in one measurement at up to 210 m (L7706) or 250 m (adaptive cruise platform). That is why every automotive ADS/ACC stack includes radar.
Where cameras and lidar are better
If you need to classify — read a sign, tell a pedestrian from a cardboard box, recognise a traffic-light state — a camera is the right tool. Radar returns a radar cross-section, not a semantic label. If you need a dense, high-resolution 3D map for localisation, lidar currently offers finer angular resolution than a conventional radar, at a higher cost and with weather sensitivity.
The fusion strategy most teams converge on
- Radar as the always-on safety layer. It never goes blind; it triggers warnings and handles weather.
- Camera for classification and semantics. It confirms what the radar flagged.
- Lidar only where the budget and the use case justify it — typically mapping and high-level autonomy, not basic ADAS.
Launston's role in this stack is the radar layer: compact, low-power modules that deliver dependable range and velocity so your camera pipeline can focus on what it does best.
Frequently asked questions
Can radar replace a camera entirely?
For detection ranging and speed, often yes. For classification — reading signs, recognising lights, identifying object types — no. Most production systems keep both.
Is lidar more accurate than radar?
Lidar has finer angular resolution in clear air. Radar has better weather tolerance and direct velocity measurement, at a fraction of the cost and power. "More accurate" depends on which quantity and which conditions.
Does radar work at night?
Yes — mmWave radar is an active sensor and is unaffected by visible light. This is the main reason it is specified for autonomous and agricultural platforms that operate around the clock.
Which sensor should I start with on a tight budget?
Start with radar if your priority is reliable detection and speed in real-world conditions; add a camera later for classification. Talk to us about your range and environment and we will suggest the exact module.