Non destructive detection of egg freshness based on computer vision

Non-destructive detection of egg freshness and storage time has great significance in the fields of egg production, transportation and processing. The experiment of egg freshness (Haff value) changes with storage time under the certain temperature was carried out. Digital images were achieved by high-resolution digital cameras with cold light illumination in reverse direction (intensity of illumination is approximate 10 000 Lx), and the index of egg yolk and the index of egg air room from image characters were extracted. The relational models of egg freshness with index of egg yolk, storage time with egg freshness, storage time with index of egg yolk and index of egg air room were established. Egg freshness with index of egg yolk showed linear relationship and the average relative error between actual test data and predicted data was 6% after detection. Storage time with egg freshness and storage time with index of egg yolk and index of egg air room showed quadratic function relationship and the average relative error was less than two days. Experimental results show that image information of egg yolk and air room from egg transillumination images generated with reverse direction lighting method on the computer vision can predict egg freshness and storage time.