Research on Image Enhancement of Crisp-Ripe Stage Red Jujube Based on Particle Swarm Optimization Algorithm

Machine vision technology plays a crucial role in non-destructive detection and online monitoring of deep processing for red dates during the crisp-ripe stage. However, due to factors such as uneven illumination in the collection environment and illumination attenuation on the curved surface of quasi-spherical objects, the collected images of red dates during the crisp-ripe stage often suffer from issues such as low overall image quality, difficulty in recognizing epidermal fruit spots and fine textures, low contrast between fruit surface crack defects and normal fruit surfaces, and blurred fruit body edges, which increase the difficulty of accurately identifying fruit surface features. In response to the common problems of over-enhancement and loss of details in current image enhancement methods when processing images of red dates during the crisp-ripe stage, this paper combines the practical requirements of preserving details of the fruit body and crack edges, enhancing epidermal texture clarity, balancing contrast and noise, and constructs a multi-dimensional image enhancement fitness function. A method for enhancing images of red dates during the crisp-ripe stage based on the Particle Swarm Optimization (PSO) algorithm is proposed. Through comparative experiments on image enhancement of red dates during the crisp-ripe stage, the effectiveness and superiority of the proposed image enhancement method are verified from both subjective visual effects and objective evaluation indicators.