American Journal of Business and Operations Research AJBOR 2692-2967 2770-0216 10.54216/AJBOR https://www.americaspg.com/journals/show/826 2018 2018 Department of CSE, Chandigarh University, Punjab, India Vijay Kumar Sinha Department of CSE, Chandigarh University, Punjab, India Shruti Aggarwal With the expanding populace, evaluating swarm thickness is a typical issue for swarm observation in Computer Vision. This issue stays a difficult assignment because of various varieties in scale issues created by various blocked uproars, changing shapes, and point of view variety. To handles these difficulties and to give a decent condition of precision we, in this way, center to gather a tremendous measure of datasets with shifting thickness levels and manufacture an Allied-CNN model. The assortment of the datasets is done from different sources like YouTube and some genuine recordings. The Allied-CNN model is worked in python and prepared on a named dataset of thousand item pictures from different points of view, for deciding thickness levels (as low thickness, medium thickness, and high thickness). Preparing results for thickness estimation show the preparation set precision arrives at 94.8%, the greatest approval exactness of just 88% is accomplished. Along these lines, this model aids in ordering a picture as low thickness, medium thickness, and high thickness. Estimations on this group datasets show that the proposed Allied-CNN performs modest outcomes contrasted with the cutting-edge strategies. 2021 2021 08 20 10.54216/AJBOR.050101 https://www.americaspg.com/articleinfo/1/show/826