This research project aims to implement deep learning models for large-scale agricultural crop detection. The study investigates the performance of YOLOv8, Detection Transformer (DETR), Neural network, and Support Vector Machine (SVM) models in terms of mean average precision (mAP), precision, recall, F1-score, and detection time. The dataset used for training and evaluation comprises 684 images with a total of 27,852 bounding boxes, including 2542 instances of Capsicum Annum L. Var Kulai (Chilli Kulai) crops. The dataset is split into training, validation, and test sets. All machine learning and deep learning models are trained with 200 iterations. The results indicate that YOLOv8 achieves the highest mean average precision of 96.2% and an average recall of 92.6%. The neural network shows good performance, with a mean average precision of 90.5% and a recall of 90.5%. The SVM model also performs well, with a mean average precision of 89.0% and a recall of 89.1%, providing a good balance between accuracy and detection speed. DETR has the lowest mean average precision of 71.9% and a recall of 45.0%. In conclusion, this research shows that YOLOv8 is the leader in performance. It was found that the model for yolov8 is the best. The study demonstrates the potential for further development in the automation of the local agricultural sector, highlighting the benefits of using advanced machine learning models to enhance crop management techniques.