Image processing technology is advancing rapidly and is widely used across multiple domains, including agriculture. While it has been extensively adopted abroad for crop monitoring and produce grading, its success in India remains limited due to high costs, restricted access, and lack of awareness among farmers. Advanced image processing models integrated with proprietary software are often embedded in expensive machinery used by large food processing industries, but small-scale farmers rarely benefit from such innovations. This paper presents a low-cost, automated image processing system for sorting and grading fruits and vegetables based on external appearance. The system employs Convolutional Neural Networks (CNNs) to assess ripeness, defects, and quality, classifying produce into predefined categories. Implemented in Python with deep learning frameworks and integrated into an Android interface, the system follows a two-phase training and testing workflow. Experimental results demonstrate a detection accuracy of 98.96%, highlighting its potential to provide affordable, accessible grading technology for small farmers, enhancing produce value, reducing post-harvest losses, and improving market competitiveness in the Indian agricultural sector.