Image Quality Tuning and Benchmarking Services for an AI-Powered Public Surveillance Camera
The camera’s ability to capture an image of a moving vehicle at a very high speed in day/night mode and different frame rates directly affects the image quality. To achieve the best image quality, it is important to tune the image pipeline (and an ISP) by configuring each ISP block optimally to respond to dynamic lighting conditions so that cameras produce accurate or pleasing images/videos in all lighting conditions.
Our customer was building a smart security camera that had a high-resolution image sensor from OmniVision with Qualcomm’s Snapdragon-based processor embedding high-performance Spectra ISP. They were looking to upgrade their camera capabilities to capture vehicles moving at high speed and then automatically detect the vehicle’s license plate, vehicle type and vehicle color using computer vision algorithms. PathPartner with its expert knowledge and credibility in image quality tuning on various ISPs and vision algorithms, helped the customer to achieve the best image and video quality for the intended use case. This case study talks about:
- The key requirements for AI-powered surveillance cameras.
- What were the pre-requisites that were handled before starting the tuning process?
- Which ISP blocks were tuned to achieve the desired output?
- Image quality (IQ) tuning methodology followed including objective and subjective tuning scenarios
- Key factors considered for IQ evaluation and benchmarking.
- What were the major challenges addressed during the process?
- Key customer benefits delivered.
Download the case study to learn more.
