Consistent, individual-level crop monitoring enhances yields and crop health by providing farmers with relevant insights for each plant, boosting overall productivity and minimizing waste. Traditional methods are time-consuming, labour-intensive, error-prone, and unreliable, making automation necessary. UAVs equipped with cameras are popular for farm monitoring and can capture images over time for further analysis. However, processing these images proves challenging due to varying lighting conditions, changes in scale due to height differences, orientation shifts based on the drone operator's skill, and fluctuating image quality depending on the camera. For effective monitoring, it's crucial to map individual crops across different images taken at various times, achieving a 1:1 crop matching over time. Traditional feature-matching algorithms fail here due to the significant visual changes caused by crop growth, weather, and farm activities. GPS offers a potential solution by tagging each crop with a unique coordinate feature for mapping, but GPS-based systems like Real-Time Kinematic and Post-Processed Kinematic are costly, complex, and struggle on uneven terrains. To address these challenges, we introduce a novel computer vision algorithm that handles variations in image quality, scale, orientation, and terrain by converting crops into 2D points for consistent matching. This method leverages the spatial relationships between crops to create unique geometric descriptors for each crop, enabling precise temporal 1:1 crop matching. Tested with UAV-acquired images, our algorithm achieved 0.84 accuracy in crop matching over time, and by incorporating Lowe’s ratio test, the precision was improved to 0.94, making the method a reliable, cost-effective, robust, and user-friendly solution.