Field of View in Machine Vision

Field of view (FOV) describes the physical area captured by a camera at a given working distance. It is a key design parameter in machine vision because it influences the selection of the lens, image sensor, working distance, and required image resolution. The field of view must be large enough to capture the entire inspection area while providing sufficient detail for reliable detection, measurement, or analysis.

Field of view is an application requirement rather than an inherent camera specification. It is defined by the size of the object or inspection area that must be captured. The required field of view is then achieved by selecting an appropriate combination of lens focal length, image sensor size, and working distance. It is defined by factors such as the size of the object or inspection area, expected variations in object position, and whether the entire area must be captured in a single image or can be distributed across multiple cameras.

How Field of View Is Measured

Field of view in machine vision is expressed as a physical dimension, typically in millimeters, representing the width and height of the area captured by the camera at the specified working distance.

Three values are commonly referenced:

Horizontal FOV (H-FOV): The width of the imaged area. This is the primary dimension used in most calculations because most objects and conveyors are measured laterally, and the horizontal sensor dimension typically determines the critical field.

Vertical FOV (V-FOV): The height of the imaged area. Relevant when the object has significant height and the full vertical extent must be captured in a single frame.

Diagonal FOV: The corner-to-corner measurement of the imaged area. Less commonly used in machine vision calculations but sometimes referenced in lens specifications.

For most industrial inspection tasks, the horizontal field of view is the starting point. The vertical field of view follows from the sensor's aspect ratio. A camera with a 4:3 sensor capturing a 200 mm horizontal field of view will simultaneously capture a 150 mm vertical field of view at the same working distance.

Determining the Required Field of View

The required field of view for an application is not simply the size of the object being inspected. It should include a margin for positioning variation, trigger timing tolerance, and any mechanical offset between where the object is detected and where it appears in the frame.

A practical starting point:

Required FOV = Object size + (2 × positioning tolerance)

If a component on a conveyor measures 80 mm wide and its lateral position can vary by up to 10 mm from the nominal center, the required horizontal field of view is at least 100 mm. In practice, an additional margin of 10-20% is commonly applied to account for edge effects and lighting fall-off at the frame boundaries.

Once the required field of view is established, it becomes the primary input to lens selection. The Imaging Source's Lens Calculator calculates the required focal length directly from the field of view, working distance, and camera sensor dimensions.

Field of View and Resolution: The Critical Relationship

Field of view alone does not determine whether a machine vision system can detect a given defect or measure a given dimension. What matters is how many pixels are available across that field of view: the spatial resolution of the imaging system.

Pixels per millimeter across the field of view is the key parameter:

Pixels per mm = Sensor pixel count (horizontal) ÷ Horizontal FOV (mm)

A 2 MP camera with a 1920-pixel horizontal sensor imaging a 200 mm wide field provides 9.6 pixels per mm. A defect measuring 0.5 mm would span approximately 4.8 pixels. Whether that is sufficient depends on the detection algorithm and the required detection confidence, but as a rule of thumb, a detectable feature should span at least 2-3 pixels to be reliably identified.

This relationship defines the minimum camera resolution for a given field of view and defect size:

Required pixel count = FOV (mm) × required pixels per mm

If the application requires detecting features as small as 0.2 mm across a 150 mm field of view, and the minimum feature size should span 3 pixels, the required horizontal resolution is 150 × (3 ÷ 0.2) = 2,250 pixels. A 2 MP sensor with 1920 horizontal pixels is marginal; a 5 MP sensor with 2448 horizontal pixels would be appropriate.

This calculation is the bridge between the inspection task and the camera specification. It is the reason field of view must be established before resolution can be selected.

Field of View vs. Depth of Field

Field of view and depth of field describe different aspects of an imaging system. Field of view defines the extent of the scene captured in the image, while depth of field defines the range of object distances that remain in acceptable focus.

Field of view is a lateral measurement: how wide and tall an area the camera captures at the working distance. Depth of field is an axial measurement: the range of distances from the camera over which objects appear acceptably sharp.

An application can have a wide field of view with shallow depth of field, or a narrow field of view with a greater depth of field, depending on the lens and aperture selected. For inspecting flat objects on a conveyor, depth of field is a secondary concern. For inspecting three-dimensional components with height variation, both parameters become important and may require tradeoffs when selecting the lens and imaging configuration.

Multi-Camera Field of View

When a single camera cannot cover the required inspection area at the needed resolution, multiple cameras are used in parallel, each covering a portion of the total field of view. In these configurations, an overlap zone between adjacent cameras is typically included to ensure full coverage and to allow for image stitching if a composite view is required.

The overlap zone is commonly set at 10-20% of each camera's individual field of view. This accommodates minor mechanical misalignment and provides redundant pixel coverage at the seam, which is typically the most difficult area to inspect reliably in a tiled arrangement.

For hardware-synchronized multi-camera setups, the trigger and timing considerations covered in Hardware Triggering and Strobe Synchronization are directly relevant. All cameras should fire simultaneously to prevent relative motion between frames when the object or conveyor is moving.