Other SoC solutions integrate essential functions in the camera as IP blocks on a single chip. These task-specific blocks can be developed in-house by a supplier, or licensed from companies that have proven and reputable solutions. The blocks can be dedicated to image processing, video compression, peripherals and more.
Vendors who do not have the resources or market share to develop everything from scratch may purchase SoC solutions to use in their product.
"However, cameras using the same chip do not necessarily translate to the same performance or quality. Vendors with technical knowledge can tweak the provided algorithms to meet their specific needs or build their own applications, said Cyril Clocher, Video Business Manager at Texas Instruments. “This allows them to add value to their products and differentiate their offerings when they use a programmable and flexible SoC.”
For example, manufacturers may have their own compression algorithms. Video compression is a component that can greatly affect performance. “The difference between a good and mediocre H.264 encoder can be more than two times in bit rate,” Christensson said. “Some codecs can even handle poorly moving objects or camera movement, which cause blur or blockiness.” Compression of the video is necessary because bandwidth and storage is not infinite, and will not be in the foreseeable future.
VCA on Edge
Megapixel cameras enable practical applications for video analytics, due to the increase in image size and clarity. There is a trend to move analytics to the edge, since the entire system can benefit from increased responsiveness when the camera is more intelligent. “For some analytics, the camera can do some processing and transmit metadata of the video to the backend servers, which can use a database to do face and behavior matching at a higher level,” Christensson said. “Another reason is to have better control over bandwidth. Instead of sending the video, the edge device can send the number of a license plate, as well as the color and speed of a car.”
"There are many benefits for analytics on edge devices; one of them is that analytics reduces the demand for bandwidth and storage," Clocher said. “The IP camera SoC can also couple a vision co-processor capable of running smart analytics, enabling the customer to reduce the cost of video analytics-enabled IP cameras.”
Video encoding The encoder inside a camera takes information passed into it and outputs the video in various formats for viewing, streaming and storing. One important factor of image quality is the information fed into the codec. Whether the image is upscaled or downscaled greatly affects the image quality.
đang được dịch, vui lòng đợi..
