Discrete wavelet transform-based video processing using FPGA applicable for object detection
Monika Sharma, Kuldeep Singh Kaswan, Dileep Kumar Yadav · Infrared Physics & Technology · 2025
Real-time object detection is one of the highest challenges in image processing systems due to the complex environment and the diverse types of objects. The usage of object detection can be made for a variety of purposes, including determining the distances to other items, such as those found in automobiles, and alerting the driver to slow down to avoid crashes. One of the widely used and quick effect techniques for image and video applications, such as object detection, is the discrete wavelet transform (DWT). The best feature of the DWT approach is that it is based on edge detection and compression, and to locate the object we suggest using a variance method to the 2-D DWT outputs of image and video processing. The image is further broken into low- and high-frequency bands using the HAAR wavelet as a reference for this study. In the research work, DWT hardware chip design for image and video object identification is completed, and the design performance is determined in several field programmable gate arrays (FPGAs) based on many comparative indices such as latency, slices, LUTs, flip-flops, and memory. The real-time video is also processed in MATLAB 2020 for the simulation of the real-time distributed images. The novelty of the design is that it provides fast switching and minimum delay in FPGA.