The article examines a methodology for object recognition in images and video recordings using a combination of Haar and Daubechies wavelet transforms. The proposed method enables efficient processing of images in real time, ensuring high recognition accuracy. The use of a combined approach leads to improved results compared to other methods such as Gabor filters or neural networks.