An improved acoustic imaging algorithm combining object detection and beamforming for acoustic camera

Abstract

High computational cost and interference with sidelobes and other mainlobes are the main problems of acoustic imaging algorithms for acoustic cameras. An improved acoustic imaging algorithm that deeply integrates object detection and a beamforming algorithm is proposed to solve the above problems. The YOLO-Fastest algorithm is employed to obtain the area of the specified equipment to reduce the imaging area of the beamforming algorithm. Experimental results show that an acoustic camera that runs the proposed method can reduce sidelobe interference and obtain acoustic images for specified equipment in real-time.

Development of an AI-based image/ultrasonic convergence camera system for accurate gas leak detection in petrochemical plants

“…Finding the local maxima of ultrasound intensity data that consistently last for a few seconds can be a good heuristic method for locating possible leakage points. This is not possible if object detection is performed first, and the ultrasound intensity data do not exist at other locations over the field of view, as in a previous study [ 73 ].…”

“…The state of research has advanced to a stage where we can attempt to detect a leak source, such as a joint, from a distance, by merging ultrasonic gas leak detection technology with artificial intelligence (AI) object detection technology to overcome ambient noise [ 73 ]. However, to-date, there has been a lack of empirical research verifying the performance of developed products.…”

Fusion of Multimodal Imaging and 3D Digitization Using Photogrammetry

“…Nevertheless, the principle is applicable to other kinds of multimodal cameras, such as ultraviolet or polarization-sensitive cameras as well. The application to more exotic cameras like photon-counting (SPAD) [81] or acoustic cameras [82,83] would need further experimental research.…”