CHAPTER 6. INFRARED SENSORS

INTRODUCTION

Infrared sensors are non-intrusive devices mounted overhead or at the side of the road to detect approaching or departing traffic. Like other traffic detection technologies, they can collect traffic volume, speed, and vehicle classification data. Some sensors can count both motorized and non-motorized traffic; however, they are more effective for motorized traffic counting.

Infrared sensors are categorized into two types: active and passive. Active infrared sensors function by emitting low-energy laser beams and measuring the time it takes for the beam to reflect back to the device. These sensors combine transmitter and receiver functions, determining the presence of a vehicle based on the reduced signal return time caused by the object’s interference. Active infrared sensors can classify vehicles and measure traffic volumes, speeds, vehicle length, and queue length.

On the other hand, passive infrared sensors are non-intrusive devices that operate only as a receiver without emitting any radiation. They detect changes in infrared radiation emitted by all objects with a temperature above zero (FLIR, 2023). The energy captured by the sensors is converted into electrical signals, which are processed and used to determine the presence of a vehicle. Passive infrared sensors with a single detection zone are capable of measuring traffic parameters such as volume, lane occupancy, and vehicle passage. Thermal sensors are a subset of passive infrared sensors that convert infrared energy (heat) into a visual image (FLIR, 2020). Figure 33 shows pictures of thermal sensors.

STRENGTHS AND WEAKNESSES

Table 6 and Table 7 summarizes the main strengths and weaknesses of active and passive infrared sensors, respectively, for counting motorized and non-motorized traffic.

Table 6. Strengths and Weaknesses of Active Infrared Sensors.

Table 7. Strengths and Weaknesses of Passive Infrared Sensors.

NCHRP Project 03-144 presented validation results for thermal cameras installed at three intersections in Arizona. All accuracy errors (WMAPE) exceeded 10%, with significant undercounting of volumes observed in most lanes. In general, infrared sensors are more effective for motorized traffic counting due to stronger heat signatures and clearer movement patterns. Detecting non-motorized users can be more challenging in crowded environments or under certain weather conditions. Their accuracy for non-motorized traffic counting depends on deployment characteristics, sensor type, and calibration.

Active infrared sensors may undercount traffic due primarily to:

• Occlusion: Like other non-intrusive equipment, infrared sensors can miss vehicles if larger vehicles cover smaller ones. Occlusion is often observed in lanes farthest from the sensor, where the vertical field of view is limited, especially when two or more adjacent lanes allow the same vehicle movement (e.g., two left-turn lanes or two through lanes). The problem is more pronounced when cameras are not installed high enough, particularly when volumes are high. Occlusion can also affect the accuracy of non-motorized count data.
• Incorrect placement and alignment: Incorrect height, position, and alignment of the sensor can result in missed detections.
• Limited field of view: If a sensor’s field of view does not include all target areas (e.g., lanes or crosswalks), some objects may be missed.
• Low-reflectivity objects: Objects with low infrared reflectivity, such as dark vehicles, matte surfaces, and heavy clothing, may not reflect enough infrared radiation to be detected. Similarly, wet pavement can absorb or scatter infrared light, reducing the reflected signal received by the sensor. These factors can make the detection of vehicles or other objects more difficult, resulting in missed counts.
• Fast-moving objects: Objects passing fast through the sensor’s detection area may not be detected.
• Environmental conditions: Heavy rain, dense fog, dust, smoke, and other atmospheric obstructions can scatter or absorb the emitted radiation, reducing detection accuracy.
• Poor calibration: Improper calibration may result in missed detections.

Except for low-reflectivity objects, all the reasons listed above may affect the accuracy of passive infrared sensors. Additionally, these sensors may undercount traffic due to the following:

• Low temperature contrast: If the temperature difference between the object and the background is small, the sensor may fail to detect the object. Rapid changes in temperature may affect the accuracy of thermal sensors.
• Small objects: Small objects, especially those passing fast through the detection zone, may not generate a strong signal to be detected.
• Sunlight glint: Sunlight glare can cause false signals and interfere with sensor accuracy.

• Thermal reflections: In the case of thermal cameras, reflections from nearby heat sources such as warm vehicles or equipment might obscure the actual thermal signature of objects, causing undercounting.

Active infrared sensors may overcount traffic primarily due to the following reasons:

• Reflective surfaces: Shiny or reflective surfaces (e.g., windows and metal surfaces) may reflect infrared radiation, triggering false detections.
• Double-counting of large objects: Different parts of the same object stopped or moving slowly in the sensor’s detection area may be detected and counted as separate objects.
• Environmental factors: Background infrared sources such as sunlight, heat from vehicles, or artificial lighting may cause false detections.
• Vibrations: Vibrations due to wind or nearby traffic may result in false detections.
• Improper calibration: High sensitivity settings can cause the sensor to detect and count small undesired objects (e.g., birds, leaves, or debris).
• Misaligned sensors: Misaligned sensors may detect and count undesired objects (e.g., tree branches).

In addition to these reasons, passive infrared sensors may overcount traffic due to:

• Non-vehicular heat sources: Non-vehicular sources of heat (e.g., sunlight, pedestrians, animals, hot pavements, etc.) may be incorrectly detected as vehicles.
• Heat retention: Thermal sensors can sometimes register heat signatures from recently passed vehicles, especially when temperatures are high, causing overcounting.

RECOMMENDED PRACTICES

Recommended practices and ideal characteristics of infrared sensors and data for traffic monitoring use are described below.

Installation and Placement

• Proper installation according to specifications: Mount sensors according to specifications at a height that allows a clear line of sight over all target lanes without interference (MnDOT and SRF, 1997). Attention should be paid at large intersections or intersections with high volumes because the counting performance of infrared sensors tends to decrease as traffic volumes increase and as large vehicles enter the intersection. Mounting according to manufacturer specifications helps ensure the sensor covers the necessary detection areas, even in large intersections or busy roads with high traffic volumes.
• Optimal tilting: Tilt the sensors at an optimal angle to maximize the detection range and accuracy for all target lanes and avoid coverage of irrelevant zones that could lead to false detections. For active sensors, carefully align the sensor to ensure the emitted beam covers the intended area without significant gaps or overlaps. For passive sensors, position the device to detect infrared radiation effectively from all expected road users in its field of view. A correctly angled sensor can monitor traffic across multiple lanes while minimizing the chances of missing vehicles. For instance, tilting the sensor too far upward or downward may result in partial lane coverage or inaccurate vehicle counting.
• Secure mounting: Mount sensors securely to prevent shaking due to strong winds, heavy vehicles, or nearby machinery.

• Heat source avoidance: Position sensors to avoid potential non-vehicular heat sources such as heating vents, air conditioning exhausts, or industrial exhausts, which could result in false detections. Careful site assessment is needed before installation to identify potential sources of heat interference and ensure the sensor’s accuracy is not compromised by external factors.
• Number of sensors: Consider using multiple sensors with overlapping coverage to monitor large intersections or eliminate blind spots.

Calibration and Maintenance

• Initial calibration: Perform initial calibration according to manufacturer guidelines to account for specific site conditions, including traffic patterns and environmental factors. For active sensors, test the reflection properties of the target objects (i.e., vehicles, pedestrians, and bicycles) to ensure accurate detection. For passive sensors, calibrate sensitivity to detect heat signatures reliably without reacting to background temperature changes. Define specific detection zones for vehicles, pedestrians, and cyclists to ensure accurate counting and classification. Configure settings to reduce false detections caused by environmental noise such as temperature variations.
• Initial data validation: Validate count and other types of data recorded by the sensors by comparing them against benchmark data (e.g., manual counts or manually reduced data from videos). Make adjustments upon validation, as needed.
• Periodic recalibration after installation: Perform periodic calibrations to maintain accuracy, particularly in locations where traffic patterns change over time or where environmental factors, like temperature fluctuations, can affect sensor performance. Implement automated calibration routines, if available, to maintain accuracy over time. Some modern thermal sensors feature automated calibration systems that adjust the sensor’s settings based on real-time conditions, ensuring consistent performance without requiring manual intervention. These automated routines help correct any shifts in the sensor’s operation caused by temperature changes, environmental factors, or other variables that could impact detection.
• Regular inspections: Conduct regular inspections to verify that the sensor is functioning correctly, particularly after strong winds or other weather conditions that could shift the sensor’s alignment. Inspect for physical damage or obstructions (e.g., new signs or vegetation growth).
• Frequent cleaning: Clean sensor lenses or covers to remove dirt, dust, or moisture that could obstruct the sensor’s view.
• Field testing: Conduct regular field tests to verify counting accuracy and signal control effectiveness for both motorized and non-motorized traffic.