Image Recognition Method Of Tomato Appearance Characteristics

Tomato grading detection is one of the important parts of post-harvest processing.However,in recent years,most of the tomato sorting in China are still selected by artificial selection.This method mainly relies on artificial vision,with low efficiency but lack of objectivity.In addition,the size of tomato is different and the surface condition is more complex,so the grading effect of tomato will be affected directly.Using computer vision technology to classify tomato has the advantages of no damage to tomatoes with high classification accuracy.In order to solve the problems in artificial detection,the maturity,size and shape of tomato were examined in different grades.(1)The test device of detecting tomato image is established,including hardware and software systems.The hardware system consists of Panasonic WV-CF344 color hemisphere camera,Daheng DH-CG320 image acquisition card,computer and lighting equipment.The software system includes Windows7 operating system,Microsoft Visual Studio development tool and OpenCV(computer open source visual library)technology.The experimental light source and the background of collection are also studied.The experiment results show that the images collected on the white background do not have shadows and the quality is high.(2)Taking tomato as the research object,the images of tomato are disposed by grayscale and binarization.In order to remove the noise,the median filter method is adopted to realize fast denoising.For the classification of tomato maturity,the RGB color model,the HSI color model and the reciprocal transformation between them are introduced,and the HSI model is selected for it is more in line with the human visual characteristics.Findcontous method is used to find and extract tomato contour.The tomato color ratio is calculated,and the maturity is classified(3)The size and shape of tomato are classified using computer vision and digital image processing technology.First,the size of tomato is discriminated by equivalent diameter.Then,the shape of tomato is rapidly classified using the circular degree method.Finally,the quality level of tomato is judged.(4)A software system for automatic detection and classification of tomatos is developed based on OpenCV technology.The main functions of the software system comprise image preprocessing,the feature extraction and classification detection of color,size and shape.This system is proved by experiment to be feasible,simple and stable.