Potato Harvesting Prediction Using an Improved ResNet-59 Model

This paper highlights why it is crucial to determine crop production using artificial intelligence for the growth of agriculture. In this paper, an elaborated ResNet-59 model has been developed to estimate potato harvests accurately. The dataset contained a global potato and tomato production data set that began in 1961 and ended in 2021; different deep learning architectures considered were ResNet-59, GoogLeNet, VGG-19, ResNet-50, VGG-16, and MobileNet. Collectively, the outcome of this ResNet-59 model's improvement led to a general superiority with more minor mean squared errors, which were recorded as 0.0083, and a mean absolute error of 0.0762, a median of absolute errors amounted to 0.0750 along with an R2 value equalling 99.05%. According to these results, precision agriculture is another area where ResNet-59 could be effective, thus promoting the rational distribution of resources, minimizing waste and increasing food security. It is epoch-making to deliberate on the capability of artificial intelligence to emancipate sustainable farming and future research.