Small form factor cameras vs. large cameras – who is the winner?

Camera size has always been debated in the embedded vision world. While many believe smaller is always better, some are not concerned about a compact design. But where does the truth lie? Should you go with small form factor cameras or relatively larger ones? What are the factors you should consider while making a decision on this?

Let’s explore all these in this article.

Why does camera size matter?

Electronic devices and components are getting smaller and smaller while achieving similar (or sometimes better) performance. This has transformed many sectors – the smartphone industry being the most notable of them. In fact, the widespread adoption of smartphones is one of the factors that led to the rapid advancements in embedded vision. Researchers and product developers invested heavily in optimizing camera modules while striving to achieve better performance.

This led them to develop more compact camera modules over time. But why is camera size a big deal after all?

The obvious reason is that it helps make the design of the vision system (whether it’s a consumer-end smartphone or an industrial security and surveillance camera) sleeker and more compact. In addition, it gives more space for other critical components to fit into the overall design. Sometimes, this might allow for integrating additional components that could enhance the performance of the system.

TechNexion small form factor camera – VLS3-AR1335-CB

For example, embedded vision systems rely on onboard processors for their image processing needs. These processors tend to increase in size with an increase in performance. For instance, the NXP i.MX series has a compact design but does not come with extremely high performance. On the other hand, the Jetson series from NVIDIA is larger in size and takes up more space but offers better performance. Hence, a small form factor camera might make integration of high-performance processors easier (There are other factors you need to consider before finalizing the size of your camera module, which we will discuss in a later section).

How small is a ‘small camera’?

Honestly, there is no standard definition as to what a small or small form factor camera is. In addition, ‘small’ can be used in relative terms too. Compared to an endoscopic camera, a smart checkout camera is big.

However, in the embedded vision world, any camera module below the size of 30×30 mm can be considered small. This is because a camera module with these dimensions is usually suitable for applications that have small camera size as a key requirement (a skin scanner for example).

Is small always good?

With the advantages of a small form factor, there is a common notion that a small size is always good. But this is not true. Here’s why.

A larger image sensor will be better at low light performance in comparison with a smaller sensor of the same resolution.

For example, the 13MP AR1335 sensor from Onsemi comes with a sensor size of 1/3.2-inch and a pixel size of 1.1µm. Now, imagine a sensor double its width and height with the same resolution. The pixel size would also be twice the original value (2.2 µm) with 4 times the area. As you might know, the larger the pixel, the more light a pixel can collect, and the higher the sensor’s sensitivity.

So, while a smaller size has its advantages, it need not be always the best option to go for. You need to pick a sensor that suits the application requirements of your vision system.

Factors to consider while choosing a camera size

When we say camera size, it is the sensor that mostly determines the size of a camera module. So, in this section, we will deep dive into all that you need to consider while choosing a sensor size for your embedded vision system.

Sensitivity
Field of view and lens compatibility
Overall size of the device
Architecture of the system
Size of the objects you need to capture
Resolution

Sensitivity

We already discussed this. Large sensors offer better sensitivity and low light performance compared to smaller sensors with similar features. So, if your vision device has to operate under a limited light supply, consider going with a large sensor.

Field of view and lens computability

The maximum field of view you can achieve using a camera system not only depends on the lens but the sensor as well. Larger sensors offer a larger field of view and hence are recommended for applications that need a wide field of view.

In addition, you also need to pick the sensor-lens combo in a way that the lens is compatible with the maximum field of view the sensor can achieve. To avoid underutilization, a lens with an image circle diameter that matches the size of the sensor is selected.

In short, if the field of view you need to achieve is low (and if other factors are favorable), you can go for a small form factor camera.

Overall size of the device

It goes without saying that the size of the embedded vision device matters a lot when it comes to choosing a particular camera size. For example, a robot (in general) can accommodate a large camera whereas a handheld retina scanner would need a small form factor camera.

Similarly, a smart surveillance system can go with a large camera while a smart cart camera has limitations when it comes to size. So, always factor in the size of the product you are building while selecting a sensor size.

Architecture of the system

Just because the vision system is large, that doesn’t always necessarily mean that the camera can be large. The design and architecture of the system also matter.

Let’s understand this through an example.

Consider a goods-to-person robot used in a warehouse for moving goods from one point to the other. It has to use multiple cameras for different purposes – 1 or 2 for depth sensing, a few for capturing images & videos, and maybe 1 or 2 more for barcode reading. This means that the robot can have 8 to 10 cameras (or sometimes more depending on its functionalities) in a single device.

This might limit the maximum size of a camera used in the device. For this reason, you need to consider the architectural design of the vision system as well while choosing a sensor. Another example of such a device is a drone.

Size of the objects you need to capture

Wondering what this has to do with the sensor size? Read on.

How do you ensure your camera system can capture minute details of a scene? One factor is of course resolution. But that’s not the only one. If the size of the object you need to capture is small, you need a sensor with a small pixel size. The smaller the pixel, the smaller you can go in terms of the size of the target object.

For example, VLS3-AR1335-CB, a 13MP AR1335 camera from TechNexion, comes with a pixel size of 1.1µm. This is perfectly suited for capturing tiny objects.

The dimensions of a sensor depend on the individual pixel size and the number of pixel columns & pixel rows. In the case of VLI-AR1335-CB, these are 4208 and 3120 respectively. With these, the sensor size comes to 1/3.2-inch – which is considered medium size for a sensor.

You need to pick a sensor in a way that an individual pixel is small enough to capture tiny objects and details in the scene.

Resolution

Resolution too has a role in determining sensor size. Let us take the same example to understand this.

The pixel size is 1.1µm. If the resolution were 1/4^th of the original value, the area or size of the sensor also reduces to one-fourth. This means that with the increase in resolution, the size of the sensor also increases, provided the pixel size remains the same. So, the right approach would be to choose a sensor size that meets the design criteria of the system while still achieving the resolution and pixel size requirements.

Closing words

Camera selection is a complex process. It involves considering numerous parameters – whether it’s about selecting the right sensor, lens, or processor. From performance to cost and durability, there are plenty of parameters to look at. Moreover, a ‘one size fits all’ approach never works in camera selection. Some applications are so complex that you might have to experiment with multiple camera modules during the prototyping stage before finalizing one.

In essence, instead of looking at each of the factors we discussed in isolation, analyze them holistically before arriving at a decision. Also, there is no clear winner here. It all depends on the specific requirements of your vision system.

How TechNexion can help in the sensor selection process

Size is not the only sensor parameter to be evaluated. There are resolution, shutter type, and sensitivity, just to name a few. Given this, sensor selection can be an exhausting and overwhelming process. TechNexion can ease that for you by leveraging the two decades of experience we have in the embedded space. Our wide variety of camera solutions also enables our customers to choose sensors for almost any of their vision needs.

In addition to offering off-the-shelf products, we also work closely with our customers to ensure that our cameras are well integrated into the end product – even if it involves customization. We can also help you with lens and processor selection, making the entire journey of camera integration faster and more productive for you. Contact us today to elevate your product development journey (more about our embedded vision solutions here).