Development And Experiment Of A Hydroponic Cultivation Device Based On OpenCV For Leaf Vegetable Growth Discrimination

Hydroponics is a green planting technology that utilizes nutrient solution for soilless cultivation of plants.It is characterized by modernization,water and fertilizer conservation,high quality and efficiency,and is also the main way to cultivate leafy vegetables and fungus plants in China’s existing plant factories.Compared with traditional soil cultivation,hydroponics can increase crop yield and quality,and reduce the probability of disease and insect damage.However,the control and monitoring of existing hydroponic leafy vegetable devices mainly rely on manual operation,which leads to problems such as high cost and low efficiency.The application of electronic technology and image recognition technology can achieve the control of the hydroponic device and the automation monitoring of leafy vegetables.Therefore,this paper designs a hydroponic carrier device for discerning the growth status of leafy vegetables based on deep liquid flow hydroponic technology.The device consists of PVC drainage pipes and other accessories,including a relatively independent embedded hardware and software hydroponic system,leafy vegetable growth status discernment based on OpenCV machine learning,and an Android-based Bat batch processing script that captures leafy vegetable images and sends them back to Windows.This system effectively helps users to hydroponically cultivate leafy vegetables and assists in discerning their growth status,predicting the relative optimal harvesting date,and providing a certain reference and reference for the development and application of hydroponic technology.The main research content and experimental results of this article are as follows:(1)Design of a Deep Liquid Flow Hydroponic Device.Through the comparative analysis of various hydroponic techniques,and based on the functional goals of the overall requirements,we have completed the hardware and software design and analysis of water pipe cultivation,aluminum profile frame,photovoltaic charging and discharging management decision,light compensation mechanism,rectifier power module,wireless communication module,and other components.We have also completed the modular development,compilation,and debugging of embedded software,achieving the independent operation of the hydroponic system,providing a foundation for the growth and development of leafy vegetables on the device.(2)Algorithm for Recognizing the Growth Status of Leafy Vegetables.The target detection of the inner and outer contours of leafy vegetables based on OpenCV algorithms has been strengthened,and an algorithm for recognizing the growth status of leafy vegetables has been designed.In simulation experiments,it was found that using only HSV color classification for binarization of leafy vegetable images can be greatly affected by the shooting angle and incident angle of light,and there are gaps in the recognition region between the light surface and shadow surface,resulting in the small recognition of connected domains after which edge detection cannot be compensated for.When using HSV color image preprocessing and then switching to RGB dual threshold discrimination,there is an expanding effect on the shadow surface and dark areas,resulting in relatively larger contours detected than their true values.When misjudging the growth regression of leafy vegetable images,a dynamic local expansion of the HSV color space range is used to process the images,and compensation for the leafy vegetable area is made by iterative use of the original recognition image point space.The recognition result of the contour detection of the outer contour of the leafy vegetable shows good effects on the light surface,and there is no obvious expansion of the connected area in the shadow region of the inner contour.(3)Hydroponic System Test for Leaf Vegetables.For coriander,the experiment showed that: 1)the growth assessment algorithm determined that coriander group A reached the relative optimal harvest date on Day 14,with variations in the length of the outer rectangle and outer circle,as well as the radius,of within 3%,and variations in the integral area of the outer contour(excluding the inner contour)of 2.68%,-2.02%,and1.92% respectively;2)the growth assessment algorithm determined that coriander group B reached the relative optimal harvest date on Day 13,with variations in the leaf area of1.59%,-1.03%,and 1.02%,respectively.For lettuce,there is a linear relationship between the integral area of the contour and the quality of the leaves,and the contour integral can be used to assess the growth status and quality.The contour areas of the lettuce image and the lettuce growth status are more closely related when the image is sampled horizontally and above the lettuce.The image sampling period is about 24 hours,which is suitable for short growth cycle leafy vegetables.For the recognition of the growth status of traditional Chinese medicinal herbs with longer growth periods,the growth cycle can be subdivided and the image sampling frequency can be dynamically adjusted to determine the optimal maturity date of the plant image recognition result.