Vehicle detection is the core foundation of adaptive intelligent traffic signals. Traffic controllers rely on various sensors to acquire real-time vehicle quantity, position and moving status, then dynamically adjust signal timing to ease road congestion. Five mainstream sensing technologies are widely deployed in global urban intersections, including inductive loop, single visible-light camera, millimeter-wave radar, radar-vision fusion Sensor and LiDAR. Each solution differs greatly in detection range, environmental adaptability, installation cost and applicable scenarios.
As the most traditional mature detection scheme, inductive loops need pre-buried metal coils inside road pavement at stop lines. When metal vehicles pass over the coil, the loop’s internal magnetic field changes and triggers signal feedback for traffic controllers.
Advantages: Low component cost, stable performance under normal weather, mature matching with traditional traffic signal machines.
Defects: Road cutting and pavement construction are essential during installation and maintenance, bringing road closure and construction cost. It can hardly identify small-sized two-wheelers like electric bikes, which is a prominent drawback under mixed traffic flow. It only realizes point detection instead of full-lane continuous monitoring.
Common scenarios: Old-fashioned intersection reconstruction projects with limited budgets.
Ordinary traffic monitoring cameras are fixed on signal light poles to capture road images and analyze vehicle information via built-in image algorithms, covering one or multiple adjacent lanes. Its effective stable detection distance is roughly 100 meters, adjusted by lens parameters.
Advantages: No road excavation required for installation, simple later maintenance, capable of outputting a lot of details of the traffic and objects, like vehicle type,LPR and lane occupancy data etc.
Defects: Highly susceptible to ambient light interference. Strong backlight at sunrise/sunset, heavy fog, rain and snow will sharply reduce recognition accuracy and even cause detection failure. Apart from light interference, standalone cameras consume high computing resources for continuous frame analysis. Meanwhile, calculations of target motion parameters such as speed and moving distance is not as accurate as radar.
Common scenarios: Urban arterial roads with fine weather conditions and low complicated light interference.
Millimeter-wave radar transmits high-frequency microwave signals to detect target objects based on echo feedback. Its effective detection distance can reach 200~500 meters far beyond ordinary cameras.
Advantages: Excellent all-weather adaptability, free from fog, rain, snow and day-night light changes; outstanding performance on tracking continuously moving vehicles, long detection range and convenient pole-mounted installation without road damage.
Defects: Poor recognition for static vehicles staying on lanes; lack of detailed visual feature information, hard to classify vehicle types precisely only relying on radar data.
Common scenarios: Highway entrances, long straight suburban roads focusing on dynamic vehicle flow statistics.
This integrated device organically combines millimeter-wave radar and industrial camera into one unit, absorbing the core strengths of two single sensors while covering respective technical defects. Radar undertakes all-weather dynamic target tracking, and the camera supplements visual classification and static vehicle identification.
Advantages: Balanced environmental adaptability and recognition precision, no road construction, realizes full-lane all-object continuous detection, gradually becomes the mainstream option for old intersection intelligent retrofit across cities. In terms of algorithm coordination, the camera can only processes image data within the target areas marked by radar, which drastically cuts overall computing consumption.
Defects: Higher unit cost than single radar or standalone camera.
Common scenarios: Urban core intersections, county-level road intelligent upgrading projects.
LiDAR forms high-precision 3D point cloud data by laser scanning, which can accurately outline vehicle outline, size, height and spatial position. Stable recognition at night yet easily disturbed by dense rain and fog particles in the air. Its valid detection distance is only dozens of meters.
Advantages: Top-tier detection precision, accurate statistics of various vehicle specifications, reliable pedestrian recognition performance.
Defects: Extremely high procurement cost restricts large-scale popularization; short detecting range and vulnerable to severe rainy and foggy climate.
Common scenarios: Municipal traffic big-data investigation, key demonstration intersections for traffic survey rather than daily automatic signal control.
There is no universal perfect sensor for all road scenes in current intelligent traffic industry. Municipal engineering teams generally select single equipment or mixed collocation according to road conditions, investment budget and actual management demands. From our practical project experience at EnerTraffic, radar-vision fusion equipment has been put into service on multiple urban intersection renovation projects to balance detection performance and overall construction cost.
