机器视觉识别田间成熟番茄的研究

为正确识别自然条件下田间成熟番茄,顺利完成其自动采摘,研究了基于颜色特征的田间番茄识别方法。对采集的100幅自然条件下田间番茄图片进行颜色特征提取和理解的基础上,建立了利用成熟番茄与背景(未成熟番茄、枝叶等)在I2颜色指标上的差异进行番茄识别的颜色模型,并利用Ostu法将成熟番茄从背景中分割出来。通过在顺光条件和逆光条件下进行试验,结果表明该模型可以较好地实现自然条件下田间成熟番茄的识别。

Study on the Machine Vision Recognition of Field Mature Tomatoes

In order to recognize field mature tomatoes under natural condition correctly and finish their automatic picking successfully,the recognition method based on color features of field tomatoes was studied.On the basis of extracting and understanding the color features of 100 field tomato pictures collected under natural condition,the color model that recognized tomatoes by the difference on I2 color index between mature tomatoes and background(premature tomatoes,branches and leaves and etc.) was set up.And the mature tomato pictures were segmented from background by Ostu method.Through conducting experiments under direct light and backlighting conditions,the results showed that the model could realize the recognition of field mature tomatoes under natural condition better.