.webp)
Autofocus MIPI Camera Module for Embedded Vision
An autofocus MIPI camera module is a board-level camera module that moves its lens to keep a subject sharp as working distance changes, and delivers the result over a MIPI CSI-2 link to a host processor. The autofocus part is straightforward to specify and harder to integrate than most datasheets suggest, because on a MIPI interface the focus loop does not close on the module by itself. Something on the host side has to decide when the image is sharp. This page covers where that decision happens, how VCM, contrast detection, and phase detection differ in practice, and which of the seven Bolt models suits a given working distance. Every model here is available as an OEM camera module or embedded camera module for board-level integration.
What an Autofocus MIPI Camera Module Is
A MIPI camera module connects to a processor's own camera interface rather than to a general-purpose bus. A MIPI CSI-2 camera module sends pixel data over one or more high-speed differential lanes and is configured over a separate I2C control channel. That control channel is what makes autofocus possible, because the same bus that sets exposure and gain also drives the focus actuator.
A CSI camera module has no enumeration step and no generic driver. The host needs a driver for the specific sensor and a description of how the module is wired to the processor. This is more integration work than a USB module requires, and the return is a direct path from sensor to processor with no intermediate bus, no protocol overhead, and no host CPU cost for moving frames.
A board-level camera module built this way is a bare printed circuit board with a lens holder, not an enclosed camera. It is designed to be mounted inside a product rather than to stand on its own, which is why the focus mechanism has to survive the same vibration and temperature range as the rest of the assembly.
Where the Autofocus Loop Actually Closes
This is the part that surprises engineers moving from USB to MIPI. On a USB module, the camera presents itself as a finished video device, and autofocus usually runs onboard. The module captures a frame, measures sharpness, moves the lens, measures again, and converges, all without the host knowing. The host sees focus as a setting, not as a task.
A MIPI autofocus camera generally cannot do that on its own, because the module has no processor running an image pipeline. Frames go straight to the host. So, on a contrast detection autofocus camera, the host has to compute a sharpness metric on incoming frames, decide which direction to move the lens, and write a new position to the focus actuator over I2C. The loop runs in host software or in the processor's image signal processor, not on the module.
The practical consequences are worth planning for rather than discovering. Focus convergence costs frames, so a continuous autofocus camera module consumes host cycles for as long as it is hunting. The algorithm has to be tuned for the lens and the scene, because a sharpness metric that works on textured surfaces behaves badly on flat ones. And an autofocus MIPI camera module needs that software written and validated as part of the integration, not treated as a feature that arrives working.
Many embedded platforms provide this in their camera framework, and some sensor drivers include a focus control interface that a standard application can drive. Confirming that support exists for the chosen processor before committing to a design is the single most useful thing to do early.
VCM, Contrast Detection and Phase Detection
Almost every autofocus camera module in this class moves the lens with a voice coil motor. A voice coil motor camera module suspends the lens on springs inside a magnetic field and passes current through a coil to displace it, typically over a stroke of a few hundred micrometers. Current position maps to focus distance. It is compact, has no gears, and draws power only while moving or holding.
A VCM autofocus camera module in its simplest form runs open loop. The driver sets a current, and the lens settles where the spring force balances it. That is repeatable enough for most work, but it shows hysteresis, meaning the position reached depends slightly on whether the lens approached from nearer or further. Designs that need tighter repeatability add a Hall effect position sensor and close the loop on actual lens position rather than on current.
Contrast detection is the focus method that pairs with a plain VCM. The system measures high-frequency content in the image, moves the lens, and measures again. It cannot tell from a single frame which direction to move or how far, so it searches. That search is why contrast autofocus visibly hunts, and why convergence is measured in tens of frames rather than in milliseconds. It is accurate when it settles, and it needs no special sensor hardware.
Phase detection changes the economics. A PDAF camera module uses dedicated pixels on the sensor that see through opposite halves of the lens aperture. Comparing what those pixel pairs report gives both the direction and the magnitude of defocus from a single frame. A phase detection autofocus camera therefore moves once and arrives, instead of searching. In this range, the Sony IMX258 camera module, Bolt-258CRA, is the model with phase detection. It is the right choice when focus has to track a subject that keeps moving, because a contrast system spends its life hunting in that situation,
Optical Image Stabilization
An OIS camera module solves a different problem from autofocus, and the two are often confused because they use similar actuators. Autofocus moves the lens along the optical axis to correct distance. Optical image stabilization moves the lens or the sensor perpendicular to that axis to correct angular movement of the whole camera.
An optical image stabilization camera uses a gyroscope to measure how fast the module is rotating, then shifts the optical path to cancel that rotation during the exposure. It corrects vibration and unsteady mounting. It does not correct subject motion because the subject is moving in the scene rather than the camera moving relative to the scene. The only fix for subject motion is a shorter exposure.
OIS earns its cost where the camera is mounted on something that vibrates and where exposures are long enough for that vibration to blur the frame. Handheld equipment, vehicle-mounted assemblies, and machinery with running motors nearby are the usual cases. In this range, the Bolt-1335CRO combines OIS with autofocus, so both the distance and the vibration problems are handled in one module.
Bolt Autofocus MIPI Camera Module Portfolio
Seven Bolt models offer autofocus over MIPI CSI-2, from a 2MP camera module to a 20MP camera module. The 4K camera module option is the Bolt-830CRH, an 8MP camera module on the onsemi AR0830, and the 13MP camera module tier carries three different focus approaches, which makes it the most useful place to compare them. Each is supplied as an OEM camera module for board-level integration.
Model | Sensor | Resolution | Autofocus Type | Buy Now |
|---|---|---|---|---|
Bolt-2020CRH | onsemi AR2020 | 20MP | VCM autofocus | |
Bolt-1335CRA | onsemi AR1335 | 13MP | VCM autofocus | |
Bolt-1335CRO | onsemi AR1335 | 13MP | VCM autofocus with OIS | |
Bolt-258CRA | Sony IMX258 | 13MP | PDAF MIPI | |
Bolt-830CRH | onsemi AR0830 | 8MP / 4K | VCM autofocus | |
Bolt-5640CRA | OmniVision OV5640 | 5MP | VCM autofocus | |
Bolt-234CGH | onsemi AR0234 | 2MP | VCM autofocus |
Reading the range by task rather than by resolution: the AR2020 camera module at 20MP suits inspection work where the detail requirement is the constraint, the AR1335 camera module at 13MP covers general purpose capture with a choice of plain autofocus or autofocus with stabilization, the IMX258 camera module is the one to pick when focus has to track movement, and the OV5640 camera module is the compact, lower cost option where 13MP is more than the application needs, a 5MP camera module for space constrained assemblies.
The Bolt-234CGH is the outlier and deserves a note. It pairs a global shutter sensor with autofocus, a combination that exists for a narrow but real case: subjects that move fast enough to need simultaneous exposure of the whole frame, at a working distance that is not fixed. Most global shutter applications have a fixed working distance and do not need focus at all.
Autofocus or Fixed Focus
A fixed focus camera module has no moving parts. The lens is set at manufacture to a hyperfocal distance that keeps everything from some near limit out to infinity acceptably sharp. If the application's working distance always falls inside that range, fixed focus is the better engineering choice by a clear margin. It costs less, draws less power, has nothing to wear out, survives vibration better, and needs no focus software at all.
An auto-focus camera module earns its place in three situations. The working distance changes during operation, as it does when a subject approaches or an arm extends. The subject is close enough that depth of field is genuinely shallow, which is why close-range reading and inspection almost always need focus while a wide scene at several meters does not. Or the same product has to work across several deployments with different geometries, and a single fixed setting cannot cover all of them.
The honest summary is that autofocus is a solution to variable distance, not an upgrade. Specifying it where the distance never changes adds an actuator, a software loop, a power draw, and a failure mode, and returns nothing.
Where Autofocus MIPI Camera Modules Are Used
Close-range reading and inspection is the largest group. Anything that images labels, codes, printed text, or small parts works at distances where depth of field is measured in centimeters, and the object is rarely presented at the same distance twice. An embedded vision camera module here spends most of its time refocusing, which is also why phase detection is worth the premium in high throughput settings.
Mobile and articulated equipment is the second. When a camera is carried on an arm, a rail, or a vehicle, the distance to whatever it is looking at changes continuously. These are also the installations most likely to vibrate, which is where the combination of autofocus and OIS on a single module removes two problems at once.
The third is equipment intended for several different installations. A product that will be mounted at whatever distance each site allows cannot be built around a fixed lens setting. Autofocus turns a mechanical constraint at installation time into a software behavior at run time, which usually pays for itself in reduced field service alone.
Quick Samples Dispatch
Find the Right Camera for Your Project Today.
Samples procured typically ship within 5-7 business days of order confirmation, with tracked international shipping. Compressed development timeline? You can be capturing frames from a Vadzo camera the same week you shortlist it.
RoHS 3 · REACH · NDAA (Selected Models)
No MOQ (Minimum Order Quantity)
Ships Worldwide
Frequently Asked Questions
1.What is an autofocus MIPI camera module?
An autofocus MIPI camera module is a board-level camera that adjusts lens position to keep a subject sharp and sends video to a host processor over a MIPI CSI-2 link. The lens is moved by an actuator that the host controls over the same I2C channel used for exposure and gain. Unlike a USB module, a MIPI camera module usually does not close the focus loop on its own, so the sharpness measurement and the decision to move the lens happen on the host side.
2.How does a MIPI CSI-2 camera module connect to a processor?
A MIPI CSI-2 camera module connects directly to the processor's camera interface over high-speed differential lanes, with a separate I2C channel for control. There is no enumeration and no generic driver, so the host needs a driver for the specific sensor and a description of how the module is wired. The integration effort is higher than for USB, and in exchange there is no protocol overhead and no host CPU cost for moving frames.
3.What is a VCM autofocus camera module?
A VCM autofocus camera module moves its lens with a voice coil motor, which suspends the lens on springs inside a magnetic field and displaces it by passing current through a coil. Drive current maps to lens position and therefore to focus distance. An open-loop VCM shows some hysteresis, meaning the position reached depends slightly on the direction of approach, and designs needing tighter repeatability add a Hall effect sensor to close the loop on measured position.
4.What is the difference between PDAF and contrast detection autofocus?
Contrast detection measures image sharpness, moves the lens, and measures again, so it has to search and cannot tell direction or distance from one frame. Phase detection uses dedicated sensor pixels that see through opposite halves of the aperture, and comparing them gives both direction and magnitude of defocus immediately. A PDAF camera module moves once and arrives, while a contrast system hunts, which is why phase detection matters most when the subject keeps moving. In this range, the Bolt-258CRA is the PDAF model.
5.When should a fixed focus camera module be used instead?
Use fixed focus whenever the working distance stays within one depth-of-field range. A fixed focus camera module costs less, draws less power, has no moving parts to wear out, tolerates vibration better, and needs no focus software. Autofocus is a solution to variable working distance rather than a general upgrade, and specifying it where the distance never changes adds an actuator, a control loop, and a failure mode for no benefit.
6.Does optical image stabilization replace autofocus?
No. They correct different things. Autofocus moves the lens along the optical axis to correct for distance to the subject. An optical image stabilization camera moves the optical path perpendicular to that axis to cancel angular movement of the camera itself, measured by a gyroscope. An OIS camera module reduces blur from vibration and unsteady mounting, and does nothing about a subject moving within the scene, which only a shorter exposure will fix. The Bolt-1335CRO provides both.
7.How do I choose the right YUV camera module?
Start with the image requirement: resolution, shutter type, dynamic range, and low-light performance. Choose a global shutter YUV camera if the subject or camera moves, and a high-resolution camera module only if the application genuinely uses the extra detail. Then match the interface to the architecture, USB 3.2 Gen 1 for host-connected designs or MIPI CSI-2 for board-level integration, and confirm the available bandwidth supports uncompressed YUV at the target frame rate. For an OEM camera module build, the mechanical format, lens mount, and operating conditions should be fixed at the same time.