If you’ve been following machine vision (MV) in fleet management, you may be wondering, “what is MV technology?” It’s not surprising since this technology seems like it has jumped out of science fiction and into many vehicles on the road today. You may also be wondering what role it plays in fleet management and how it can help your fleet.
Reducing distracted driving is a big issue for fleet managers. Not only is it dangerous and a significant cause of accidents, but it also puts your drivers’ safety and your bottom line at risk. Safe driving requires constant attention, which is why distracted driving is so dangerous. MV technology offers an exciting future where drivers can better focus on driving and spend less time off the road due to accidents and repairs.
Many fleets have already installed machine vision cameras that alert drivers and fleet managers about risky behavior and help them correct it before an accident occurs. These fleets have seen tangible results in fewer accidents and, in many cases, lower insurance costs.
With this relatively new technology becoming more prevalent in many industries, including fleets and logistics, it’s a good time to explore what it is and what it does. Below is a quick introduction to MV that will help you grasp the basics and why it is so essential to fleet safety.
One of the easiest ways to understand MV is the think of it as the "eyes" of a machine or system. IT uses input captured by digital cameras to identify action. Businesses use these systems in numerous ways to improve efficiency, operations, and quality.
At its most basic level, MV vision technology analyzes video and image data to make predictions. MV can detect items such as coffee cups and cell phones in pictures and videos. It can also detect when drivers are engaging in risky behaviors.
MV automatically extracts information from digital images and uses this information for various applications in various industries. It isn’t one technology but rather a combination of hardware and several technologies that automate the extraction of information. The information can be simple or complex, based on the system, and can be used for applications such as automatic inspections, security monitoring, agriculture management, and vehicle guidance.
MV systems need to “learn” what to look for and what to expect in images. A machine vision camera is used to feed thousands of pictures or videos through the system, thus teaching it what to look for and what to expect. These images are manually tagged so that the machine vision software knows what objects they contain.
Advanced MV systems can be taught to better understand what is and is not in the images they “see.” The final step of this process often uses artificial intelligence (AI) to calculate risk or make predictions.
An MV system is only as accurate as the data it is fed. To develop systems that can handle all of the divergent possibilities in real life, the software needs to be fed millions of unique images that depict all imaginable situations. For instance, machine vision software for fleet management will have been taught by many years of driving data from many different vehicles, road types, and weather conditions.
Several components make up an MV system. This includes a machine vision camera to capture data. The machine vision camera is connected to a computer, and from there, the software is used to crunch the data and generate analyses or make predictions. Each of the components below plays a distinct role in the system.
The optical component can be a lens or a machine vision camera that integrates the lens and other components such as the image sensor. The optical system or lens used depends on the specific application of the system.
The sensor captures light from the optical system and changes it into a digital image.
This element of an MV system takes the digital image and uses software to evaluate it based on a set of parameters or conditions that identify what needs to be measured or observed and establish a result.
This element provides a usable output in a standardized format so that the data can be used.
The majority of miles that commercial fleets travel in the US are uneventful – safe and incident-free. Drivers may feel comfortable eating, drinking, smoking, or taking a short call on their mobile phone while driving because they’ve done it many times with no issues. But it takes just one unlucky moment of distraction to cause an accident, and that one time can have far-reaching consequences.
That's where MV cameras and MV systems come in. The ability to recognize serious moments of risk is critical to reducing accidents. These systems sift through hours of video-generated footage to identify short but key moments of distracted driving.
A machine vision camera installed in fleet vehicles can help prevent injuries, fatalities, and costly accidents. Drivers and fleet managers can get alerts when drivers engage in risky or dangerous behaviors. Machine vision software can send instant alerts when a driver exhibits signs of fatigue or distraction. It will also help identify when drivers neglect to put on their seat belts or smoke while on the job (driving).
Driving for long distances and long hours is complex, dangerous, and mentally taxing. Fleet drivers have to deal with varying road conditions, distracted pedestrians, cyclists, traffic, other drivers, wildlife appearing on the road, and sudden changes in weather. An MV system helps drivers manage those complexities better.
Smart cameras are a relatively new innovation in MV systems, which generally have relied on PC-based image processing in the past. There are several key differences between smart camera systems and PC-based systems. Early smart cameras had limited capacity for interpreting images; however, both the technology and processing power have advanced, and smart cameras can now handle a wider range of images. Smart cameras can be seamlessly integrated into MV systems.
Multiple smart video cameras can be used to set up a visual sensor network when positioned at specific locations to capture images of a specific area from several angles. These images are fused together and are more useful than each individual image captured. Sensor networks can monitor environmental conditions and track objects in motion.
There are myriad uses for machine vision for industrial automation of quality control, guidance, inspection, identification, and more. Below are some common examples of machine vision capabilities and industrial tasks:
Correcting defects: Machine vision can help identify problems so that they can be corrected.
Farming: Detects location fruit for robotic harvesting machines, monitors crops, and detects plant diseases.
Inventory control and management: Reads barcodes on components and products, important applications for manufacturing, and inventory control. MV vision ensures the correct components move along the assembly line and are essential for bin-picking done by robots in warehouses.
Tracking and traceability: In heavily regulated industries, such as pharmaceuticals, MV makes tracking ingredients and monitoring expiration dates easier.
Measurements and calibration: Machine vision automates the measuring of things like the gap in a spark plug, or a gauge that needs to be calibrated.
A number of industries take advantage of these capabilities, such as:
A machine vision introduction wouldn’t be complete without discussing Artificial Intelligence (AI). MV is the eyes, and AI acts as the brain of an MV system, giving it the ability to interpret and decide. For instance, machine vision AI dashcams “see” a driver light a cigarette or pick up a mobile phone and identifies them as risks.
Road-facing and in-cab cameras identify and alert drivers and fleet managers of risky driver behavior and help drivers correct their behavior before an accident occurs. An AI driving coach guides drivers in real-time with unobtrusive visual alerts. They can also detect driver fatigue and distracted driving to help keep drivers safe on the road.
Dash camera systems without MV and AI are not as effective as their smart tech counterparts. They capture video footage of what is happening while the drivers are on the road and make it available for future viewing by the fleet manager. However, it takes time to review all the footage and then take action after the fact.
Dash camera systems with MV and AI recognize what is happening in real-time without a human having to review any footage. They not only capture video footage but also provide data that can be used for vehicle guidance. Real-time alerts allow drivers to change course to prevent collisions and give fleet managers a clearer idea of what has transpired during a driving event, when reviewed historically.
By sending this key information right to the driver in real-time and to fleet managers afterward, AI-enabled dash cameras provide a much more efficient and effective method of improving and maintaining fleet safety.