AVs rely on the detection and recognition of objects within images to successfully navigate. Design of camera systems is non-trivial and involves trading system specifications across many parameters to optimize performance, such as f-number, focal length, CFA choice, pixel and sensor size. As such, tools are needed to evaluate and predict the performance of such cameras for object detection. Of critical importance is an ability to estimate the capability of a camera system to detect objects at distance across a wide array of illumination conditions. Apart from obvious safety considerations, how else will be it possible to ascertain the performance that a DNN should be capable of delivering if the basic objective performance of the input imaging device cannot be determined?
CDP is a relatively new objective image quality metric proposed to rank the performance of camera systems intended for use in autonomous vehicles. Detectability index is derived from signal detection theory as applied to imaging systems and is used to estimate the ability of a system to statistically distinguish objects, most notably in the medical imaging and defence fields. A brief overview of CDP and detectability index is given after which an imaging model is developed to compare and explore the behavior of each with respect to camera parameters. Behavior is compared to matched filter detection performance and conclusion drawn regarding the performance of both metrics.