The accelerated progression of computer applications, transitioning from basic data processing to the realm of machine learning, can be attributed to the increased accessibility and abundance of extensive data sets obtained from various sensors and the internet. This study provides evidence that robots have the capacity to acquire self-awareness and evolve through appropriate teaching. The effectiveness of human categorization of photos is contingent upon the utilization of image analysis techniques and the implementation of diverse approaches, which provide inherent challenges. Due to the intricate nature of the task at hand, it has become imperative to implement automation techniques in order to achieve a significant level of precision. The proposed study aims to provide a comparative analysis on the efficacy and accuracy of a number of machine learning algorithms in the context of image categorization. The logistic regression approach, Naive Bayes classifier approach, Support Vector approach, and Random Forest algorithms for classifiers were employed to evaluate the UC Merced dataset. The experimentation began by preprocessing and training of the dataset, which was then followed by the testing phase. At this juncture, the inquiry shifts its focus towards determining the optimal selection algorithm by analyzing the computed accuracies of the evaluated methods.