You are currently viewing LIDAR Versus Computer Vision

LIDAR Versus Computer Vision


LIDAR technology has been applied mainly in the automotive industry, especially in the development of self-driving cars. Here, LIDAR radars utilize invisible infrared laser light to determine the distance of surrounding vehicles and their velocity for accurate decision making. On the other hand, computer vision is a discipline that involves acquisition, analysis, and interpretation of digital and video images to produce reliable information for decision making. While these technologies can be integrated for maximum operational efficiency, they have stark differences.

LIDAR

As mentioned above, this type of technology sends out infrared laser light to the target object. The time it takes for this light to be reflected in the source will provide an accurate indicator of the object’s distance and velocity. The method of determination of object distance itself is super-fast because of the lightning speed of light, the reason LIDAR boasts of reliability and maximum safety in self-driving cars.

Whereas LIDAR technology prides itself on inaccurate range information, the image resolution of objects is relatively low due to its object recognition limitation. Moreover, LIDAR works effectively in different lighting conditions, whether day or night time.

Computer Vision

In computer vision, automakers collect, process, analyze, and interpret digital images and videos of target objects to produce reliable information for informed decision making. Computer vision tends to work just like human visual cortex does to automate tasks. Object recognition in computer vision is excellent than in LIDAR technology. Moreover, object range information isn’t as accurate in computer vision than in LIDAR.

Comparisons between LIDAR and computer vision about:


  • Object recognition capability

Self-driving cars that rely on computer vision systems possess excellent object recognition capabilities because they detect objects in different colours and better understand the scene. Conversely, LIDAR radars only focus on determining object velocity and distance alone. Thus cars, where LIDAR sensors are installed, cannot recognize nearby cars and other objects.

  • Range information accuracy

LIDAR technology prides itself in its capability to offer accurate and reliable range information, leave alone the speed at which data processing happens. On the flip side, computer vision’s range information is usually inaccurate and unreliable because of the time taken to understand the surrounding scene.

  • Efficiency in varied lighting conditions

Operational efficiency of a computer vision system is highly dependent on the surrounding lighting conditions. The sensors on these cars have to adapt and handle these varied lighting conditions. Thus, the lag in time leads to erroneous conclusions. On the contrary, LIDAR systems work well both at night and day time.

  • Cost implications

Whereas most automakers opt for integrating both LIDAR and computer vision systems on self-driving cars for maximum efficiency, they vary significantly in their cost. LIDAR radars have been costlier than computer vision systems. Thus, there is a need for a trade-off in decision making. LIDAR’s capability to provide accurate object distance and velocity information for maximum safety is what makes them very expensive.

The bottom line

LIDAR technology is an advanced form of radar technology where invisible infrared laser light is cast towards a target object. The time taken for that light to bounce back to the source will provide an excellent indicator of the object’s velocity and distance. This technology provides accurate and more extended ranger information, usually between 100m to 300m.

On the other hand, computer vision technology involves collecting, analyzing, and interpreting digital images and videos to understand the world for appropriate action better. Whereas both approaches have considerable differences, they can be integrated into self-driving cars for maximum operational efficiency.

Leave a Reply