Why Use ASMVs in Life Science Applications

Application Specific Machine Vision Systems (ASMV) are turnkey machine vision systems that addresses a specific application throughout one or more industries. A true ASMV system will help in many ways; it reduces the burden of figuring out how to interface all of the machine vision components together, which can be a challenge especially when using interfaces that are less plug and play than others.

ASMV manufacturers solve a variety of problems for their customers, like matching the camera and lens, making sure there is sufficient processing power for multiple cameras, and ensuring the right video processing and bandwidth capability is present in the CPU and GPU setup. The final system has had the variables pre-tested to ensure it focuses on what it needs to accomplish.

Sometimes users' setups are not running the way they should be, and ASMV manufacturers can debug a system to determine what frame grabber card is used, or how much CPU capability the PC has, among other things.

ASMV systems may or may not have the basic software capabilities needed in terms of pre-defined functions to meet a customer's goals. It may be more of a raw development or sdk environment with libraries that customers have to put together to accomplish specific tasks. But at any level, buying something pre-bundled removes many question marks from the equation.

Top considerations

When considering an ASMV system for the life sciences, the first consideration should be if the system can accomplish what is desired. Some ASMV manufacturers offer cameras suitable for applications where high color precision is needed, such as microscopy, staining slides, or some type of digital pathology that looks for cancer cells.

In applications such as these, the vision system looks for precise colors or shapes, and it may be more beneficial to use supply prism-based CCD cameras that provide high spatial precision and true RGB values instead of interpolating color information from sensors that have Bayer filters. Only a handful of companies in the machine vision world produce these types of cameras that have dichroic coatings on the sensor. A customer could specify such a camera when building a system from the component level themselves or locate an ASMV provider and use one of their systems as a starting platform and put cameras such as these inside.

There is no Typical ASMV System

Both color and monochrome cameras are used in ASMV systems. In the industrial inspection market, color may not matter and get in the way while looking for defects. But in many imaging, microscopy, and other applications, color is more important.

Most cameras in ASMV systems are digital. Analog cameras aren't typically sold into ASMV applications at all in the life science industry, with the holdouts being in the defense industry that has long transition times.

When building a system from the component level, many users like the plug and play convenience and the direct connection advantage of USB 3.0 in life science industries. Their applications don't need GigE because they don't require large networked environments, and they don't want to wrestle with frame grabbers that may make their systems bigger, bulkier, or more cumbersome. But in ASMV systems, this is not as big an issue, since ASMV makers handle the design and remove some of the testing complexity.

As the industry transitions more towards CMOS sensors, the life science industry is more of a holdout of CCDs than other industries. Pattern noise has been one knock against CMOS, because every pixel has its own amplifier and the readout is done differently so there can be a visible pattern in the image. In applications where clean images are needed, life science and even in astronomy markets, noise could still become an issue with CMOS when longer exposures are needed and CCDs should be used. CMOS is gradually winning over the holdouts and CCDs are no longer the only sensor that will provide accurate solutions. Whether component based or in an ASMV, the considerations are the same.