Drone-assisted adaptive object detection and privacy-preserving surveillance in smart cities using whale-optimized deep reinforcement learning techniques

Drone/ unmanned aerial vehicles (UAV) surveillance for object/ human detection is familiar in large gatherings in the modern cities era. Artificial intelligence algorithms and computer-aided processing will handle the images extracted from the surveillance videos to reveal the object. This article proposes a novel object detection technique (ODT) that assimilates whale optimization algorithms and deep reinforcement learning. The optimization algorithm detects spreading image features from the origin to the end of x×y pixels. This feature extraction is performed until the complete image pixels are covered to identify their existence in the least position. The forging behaviour of the whales is implied to identify highly overlapping features for object/ human detection/ classification. If the overlapping features increase, the whale's movement is updated from the last-known highest pixel position. The reinforcement learning recommends that new whale agents validate the low overlapping features, such that the forging whale agent moves towards the large overlapping feature. Therefore, the highest overlapping feature-based pixel differentiation is pursued using forging and searching whales to identify objects through collated pixels.

Перевод пока недоступен