Joint target detection method and device based on video frames and pulse array signals

Abstract

The invention discloses a joint target detection method and device based on a video frame and a pulse array signal. The method comprises the following steps: taking the video frame and the pulse arraysignal as input of target detection; performing self-adaptive division on the continuous pulse array signals; performing synchronous fusion or asynchronous fusion detection according to the relationship between the video frame frequency and the pulse array signal division frequency; fusing the video frame and the pulse array signal in a feature expression form; outputting a detection result. According to the invention, the high time resolution and high dynamic capability of the pulse array signal can be effectively utilized to provide the target detection precision of the traditional camera,and especially, the detection problems of high-speed motion blur, overexposure or low illumination and other scenes are solved. The method has wide application potential in the fields of unmanned visual inspection and navigation, unmanned aerial vehicle cruising and positioning, robot visual navigation, video monitoring and the like which relate to high-speed motion.

Description

technical field [0001] The invention relates to the technical field of computer vision, in particular to a joint target detection method and device based on video frames and pulse array signals. Background technique [0002] The task of target detection is to find out the target objects of interest in the visual scene, determine their position and size, which is the core issue in the field of machine vision. And video surveillance and other fields are widely used. [0003] Compared with lidar and ultrasonic, visual sensors have high spatial resolution and capture sufficient video data, which plays an important role in target detection tasks. Traditional cameras are based on fixed exposure time sampling, and there are video frames (generally 30-120 frames per second) composed of light intensity integration at fixed time intervals. In slow scenes and normal lighting scenes, clear texture details of the scene can be captured. However, The low temporal resolution and low dynam...

AI Technical Summary

Problems solved by technology

Traditional frame rate vision sensors can collect high-definition images, but there are shortcomings in data collection such as large redundancy, small dynamic range of light sensitivity, low temporal resolution of fixed frame rate, and high-speed motion blur.

Images