Camouflage Reconnaissance in Multispectral Images by utilizing Machine Learning algorithms will be the crucial task of identifying concealed camouflaged objects. Essentially this can be applied for various fields including security, military operations, and environmental monitoring. Using Altum-PT multispectral camera, the diverse dataset of multispectral images was collected and processed. This involves employing techniques such as fusion, Numpy false colouring, and pansharpening etc. to enhance image quality and to extract relevant features. The aim of this investigation revolves around detecting concealed camouflaged objects using advanced deep learning models, specifically You Only Look Once (YOLO) V5, V7, V8, V9 and Slicing Aided Hyper Interference (SAHI). These models have been assessed for their effectiveness accurately by identifying concealed objects across different environmental contexts. Furthermore, the work includes a comparative study where the results obtained from each model are analyzed and compared. This provides valuable insights into the strengths and weaknesses of individual models in the context of camouflage reconnaissance. Therefore, this investigation will become an advancement in enhancing surveillance capabilities by contributing to improved situational awareness and security across various domains.