Agriculture is significant in human life. World population is directly reliant on cultivation. The agriculture industry has several challenges as out-migration of agriculture labors has profound effects on rural economic evolution, food security, nutrition, and poverty, affecting agricultural presentation, rural households and in general the rural economy, causing declination in the amount of agricultural labors and growing price of yield harvesting. Maintaining the labor and scrambling up in crop growing is essential in resolving these difficulties. In current scenario, the mechanization has been evolving for maintaining the labor and large-scale agriculture. There was never been encounter an experimental approach in detection of cotton crop. However, performance is being a greater opportunity to research. Therefore, in this research we propose an architectural frame work in detecting the cotton crop using faster R-CNN and Yolov3. The experiment carried out to identifying and detecting the cotton crop efficiency using faster R-CNN and YOLOv3. A dataset from the cotton crop field in Karnataka, south India was considered. The proposed framework performance is measured using mAP (mean Average Precision) and confidence-score. Experiment resulting 85% and 97.73% confidence-score for faster R-CNN and Yolov3 for 10000 iterations respectively. It is observed that Confidence score is sensitive to the number of iterations.