TW202123152A - Fruits and vegetables monitoring system and method thereof - Google Patents
Fruits and vegetables monitoring system and method thereof Download PDFInfo
- Publication number
- TW202123152A TW202123152A TW108145637A TW108145637A TW202123152A TW 202123152 A TW202123152 A TW 202123152A TW 108145637 A TW108145637 A TW 108145637A TW 108145637 A TW108145637 A TW 108145637A TW 202123152 A TW202123152 A TW 202123152A
- Authority
- TW
- Taiwan
- Prior art keywords
- fruit
- vegetable
- stalk
- fruits
- vegetables
- Prior art date
Links
Images
Abstract
Description
本發明是有關於一種智慧農業技術,且特別是有關於一種蔬果監控系統及其方法。The present invention relates to a smart agricultural technology, and particularly relates to a vegetable and fruit monitoring system and method.
近年來,政府輔導協助青年返鄉從農。然而,初入門的青農通常沒有任何農場經驗。即便投入2至3年的學習成本,也可能沒有好的結果(例如,收成欠佳、成本過高等)。In recent years, the government has assisted young people to return to their hometowns to work in agriculture. However, young farmers who are beginning to enter usually do not have any farm experience. Even if the cost of learning for 2 to 3 years is invested, there may not be good results (for example, poor harvest, high cost, etc.).
另一方面,不同類型的蔬果有各自合適的生長環境。然而,環境因素的變異很大,諸如晝夜溫差、乾旱、病蟲害等因素通常難以單憑藉著人力來即時改善。On the other hand, different types of fruits and vegetables have their own suitable growth environment. However, environmental factors vary greatly. Factors such as temperature difference between day and night, drought, pests and diseases, etc. are usually difficult to improve immediately with manpower alone.
有鑑於此,本發明實施例提供一種蔬果監控系統及其方法,以物聯網(IoT)架構及人工智慧(AI)技術監控蔬果的生長環境,讓農民可輕鬆管理農場。In view of this, the embodiments of the present invention provide a vegetable and fruit monitoring system and a method thereof, which use the Internet of Things (IoT) architecture and artificial intelligence (AI) technology to monitor the growth environment of the vegetable and fruit, so that farmers can easily manage the farm.
本發明實施例的蔬果監控系統,其適用於監控農場中種植的蔬果。蔬果監控系統包括但不僅限於影像擷取裝置及控制裝置。影像擷取裝置用以拍攝蔬果以取得影像。控制裝置依據影像判斷蔬果的成熟程度及位置資訊。位置資訊包括果實位置及蒂頭採收位置,且果實位置包括對應蔬果的果實的頂部位置及底部位置。控制裝置依據相對位置關係估測蒂頭採收位置。此相對位置關係是基於對應頂部位置及底部位置所得出,且蒂頭採收位置是供外部物件剪摘的位置。The vegetable and fruit monitoring system of the embodiment of the present invention is suitable for monitoring the vegetable and fruit grown in the farm. The vegetable and fruit monitoring system includes but is not limited to image capture devices and control devices. The image capturing device is used to capture fruits and vegetables to obtain images. The control device judges the maturity and location information of the fruits and vegetables based on the images. The position information includes the position of the fruit and the harvest position of the stalk, and the position of the fruit includes the top position and the bottom position of the fruit corresponding to the vegetable and fruit. The control device estimates the picking position of the stalk based on the relative position relationship. This relative positional relationship is obtained based on the corresponding top position and bottom position, and the stalk harvesting position is the position for cutting external objects.
另一方面,本發明實施例的蔬果監控方法,其適用於監控農場中種植的蔬果,並包括下列步驟:拍攝蔬果,以取得影像。依據影像判斷蔬果的成熟程度及位置資訊。依據相對位置關係估測蒂頭採收位置。此位置資訊包括果實位置及蒂頭採收位置,果實位置包括對應蔬果的果實的頂部位置及底部位置,相對位置關係是基於對應頂部位置及底部位置所得出,且蒂頭採收位置是供外部物件剪摘的位置。On the other hand, the method for monitoring fruits and vegetables according to embodiments of the present invention is suitable for monitoring fruits and vegetables grown in farms, and includes the following steps: photographing fruits and vegetables to obtain images. Judge the maturity and location information of fruits and vegetables based on images. Estimate the picking position of the stalk based on the relative position relationship. The position information includes the position of the fruit and the harvesting position of the stalk. The position of the fruit includes the top and bottom positions of the fruit corresponding to the fruits and vegetables. The relative position relationship is based on the corresponding top and bottom positions. The harvesting position of the stalk is for the outside The location where the object is cut.
基於上述,本發明實施例的蔬果監控系統及其方法,透過影像辨識技術判斷蔬果的成熟程度及所在位置,並基於果實與蒂頭的相對位置關係估測蒂頭採收位置。藉此,可方便農民在遠端即可了解蔬果的生長情形,並可進一步規劃採收作業。Based on the above, the vegetable and fruit monitoring system and method of the embodiments of the present invention judge the maturity and location of the vegetable and fruit through image recognition technology, and estimate the stalk harvesting position based on the relative positional relationship between the fruit and the stalk. In this way, it is convenient for farmers to understand the growth situation of fruits and vegetables at the remote end, and can further plan harvesting operations.
為讓本發明的上述特徵和優點能更明顯易懂,下文特舉實施例,並配合所附圖式作詳細說明如下。In order to make the above-mentioned features and advantages of the present invention more comprehensible, the following specific embodiments are described in detail in conjunction with the accompanying drawings.
圖1是依據本發明一實施例蔬果監控系統100的方塊圖。請參照圖1,蔬果監控系統100包括但不僅限於感測裝置110、影像擷取裝置130、環境調節裝置150、採收裝置170及控制裝置190。蔬果監控系統100適用於監控至少一個農場(例如,網室、溫室或開放農地)中種植的至少一種或至少一個蔬果。需說明的是,蔬果監控系統100中各裝置的數量可依據實際需求而自行調整。FIG. 1 is a block diagram of a vegetable and
感測裝置110可以是溫度計、濕度計、可見光計(或照度計,並用於量測陽光照度)、其他類型的環境感測器、或其組合。溫度計可能包括熱電偶(Thermocouple)、熱敏電阻(thermistor)與熱電阻(thermal resistor)等感測元件。在一實施例中,在農業環境內(例如,溫度約0-50℃),電阻式溫度計所用的金屬材料的電阻與溫度變化線性關係較好,因此可透過電阻式溫度感測器來取得農場(例如,溫室)的溫度資料。在另一實施例中,由於臺灣所處亞熱帶地區的濕度較高,因此可電阻式相對濕度計來取得溫室的溼度資料。The
影像擷取裝置130可以是各類型相機、錄影機、閉路攝影機、智慧型手機、或平板電腦。影像擷取裝置130可能固設於農場或移動裝置(例如,機器人、無人飛機、遙控機、或軌道車等)上,並可對特定區域(受限於視野(FOV))、或受控於指示而改變視野來拍攝。The image capturing
環境調節裝置150可以是風扇、除濕機、灑水器、空調設備、補光燈、或其組合。The
採收裝置170可以是機器手臂、或機器人。在一實施例中,採收裝置170包括可抓取、剪摘、及/或推拉等行為的自動化機械部件。在本發明實施例中,採收裝置170用於對採收蔬果。而由於不同類型蔬果的採收行為不同,因此應用者可依據需求自行調整採收裝置170的機械部件。The
控制裝置190可以是桌上型電腦、筆記型電腦、伺服器、工作站、智慧型手機、平板電腦、智能助理等裝置,並至少具有諸如CPU、或晶片等處理器,從而進行影像處理、影像分析、影像辨識、感測資料收集及分析、遙控指示等作業。The
需說明的是,前述裝置110~190更分別包括相容通訊技術(例如,Wi-Fi、藍芽、紅外線等)的通訊收發器,使彼此之間可相互通訊。在一些實施例中,部分裝置更可能整合成單一裝置。例如,影像擷取裝置130與採收裝置170及控制裝置190整合在一起。It should be noted that the
為了方便理解本發明實施例的操作流程,以下將舉諸多實施例詳細說明本發明實施例中農場所種植蔬果的監控流程。下文中,將搭配蔬果監控系統1中的各項裝置、元件及模組說明本發明實施例所述之方法。本方法的各個流程可依照實施情形而隨之調整,且並不僅限於此。In order to facilitate the understanding of the operation process of the embodiment of the present invention, a number of embodiments will be given below to describe in detail the monitoring process of the fruits and vegetables grown on the farm in the embodiment of the present invention. Hereinafter, various devices, components, and modules in the vegetable and fruit monitoring system 1 will be used to illustrate the method described in the embodiment of the present invention. Each process of the method can be adjusted accordingly according to the implementation situation, and it is not limited to this.
圖2是依據本發明一實施例蔬果監控方法的流程圖。請參照圖2,影像擷取裝置130拍攝蔬果,以取得影像(步驟S210)。農地以番茄園為例(即,蔬果以番茄為例),控制裝置190可驅動移動裝置,使影像擷取裝置130移動到指定位置(例如,與番茄相距大約30~50公分處),並進行拍攝作業,從而取得影像。又例如,影像擷取裝置130固設於番茄種植處四周,並對蕃茄錄製或拍攝影像。需說明的是,本發明實施例不加以限制影像取得方式。此外,蔬果的類型也不限於番茄。例如,香蕉、草莓等蔬果。Fig. 2 is a flowchart of a method for monitoring fruits and vegetables according to an embodiment of the present invention. Please refer to FIG. 2, the image capturing
接著,控制裝置190自影像擷取裝置130取得影像,並依據一張或更多張影像判斷蔬果的成熟程度及位置資訊(步驟S230)。具體而言,控制裝置190是基於影像辨識技術來判斷成熟程度及位置資訊。在一實施例中,此影像辨識技術是基於機器學習演算法。機器學習演算法可能是區域基礎卷積神經網路(Region-based Convolutional Neural Network,RCNN)、快速區域基礎卷積神經網路(Fast Region-based Convolutional Neural Network,Fast R-CNN)、或較快速區域基礎卷積神經網路(Faster Region-based Convolutional Neural Network,Faster R-CNN)等架構。這些類神經網路訓練架構在目標檢測方法受廣泛使用。Then, the
在另一實施例中,機器學習演算法是遮罩區域基礎卷積神經網路(Mask Region-based Convolutional Neural Network,Mask RCNN)架構。簡單而言,Mask RCNN是在Faster R-CNN的基礎上增加遮罩的回歸,以輸出語意分割的結果(例如,分割遮罩(segmentation mask))。Mask R-CNN 包括兩階段程序,第一個階段掃描影像並生成目標的區域,第二階段分類並生成邊界框和遮罩。In another embodiment, the machine learning algorithm is a Mask Region-based Convolutional Neural Network (Mask RCNN) architecture. Simply put, Mask RCNN is based on Faster R-CNN to increase the regression of the mask to output the result of semantic segmentation (for example, segmentation mask). Mask R-CNN includes a two-stage program. The first stage scans the image and generates the target area, and the second stage classifies and generates bounding boxes and masks.
圖3是一範例說明果實辨識結果。請參照圖3,以番茄為例,控制裝置190不僅可偵測到果實,更能進一步決定果實在影像中的輪廓(圖中以較深色遮罩示意物件偵測結果)。Figure 3 is an example illustrating the results of fruit identification. Referring to FIG. 3, taking tomato as an example, the
需說明的是,本發明實施例不限制機器學習技術的類型。而值得注意的是,在CNN的架構中,卷積層(Convolution layer)及池化層(Pooling layer)增強了圖形辨識及資料相鄰之間的關係,使CNN應用在影像、聲音等訊號類型的資料型態能得到很好的效果。與傳統的多層感知(multilayer perceptron,MLP)網路最大的差異在於,CNN增加卷積層及池化層兩層,以維持形狀資訊並避免參數大幅增加。此外,架構更包括全連接層(Full connection),最後為分類器輸出結果。而此分類器是透過輸入訓練樣本所產生,且這些訓練樣本例如是包括不同類型蔬果的影像。It should be noted that the embodiment of the present invention does not limit the type of machine learning technology. It is worth noting that in the CNN architecture, the convolution layer and the pooling layer enhance the relationship between image recognition and data neighbors, so that CNN can be used in image, sound and other signal types. Data types can get very good results. The biggest difference from the traditional multilayer perceptron (MLP) network is that CNN adds two layers: convolutional layer and pooling layer to maintain shape information and avoid significant increase in parameters. In addition, the architecture includes a full connection layer (Full connection), and finally outputs the result for the classifier. The classifier is generated by inputting training samples, and these training samples are, for example, images including different types of fruits and vegetables.
在一些實施例中,激活函數(activation function)可以是ReLU、Leaky ReLU、或Swish,但不以此為限。In some embodiments, the activation function can be ReLU, Leaky ReLU, or Swish, but it is not limited to this.
除了可自影像中偵測指定蔬果,控制裝置190亦可基於機器學習演算法預測蔬果的邊界(例如,特定形狀的框、或蔬果的輪廓),從而得出蔬果的位置資訊。此位置資訊可能是在某一空間座標系統的座標、或與其他參考物的相對位置。若能辨識蔬果的輪廓,控制裝置190可進一步決定蔬果的輪廓/外表的位置資訊。In addition to detecting the specified fruits and vegetables from the images, the
在一實施例中,蔬果的位置資訊包括果實位置及蒂頭採收位置,且果實位置包括對應蔬果的果實的頂部位置及底部位置。具體而言,果實位置可以是其邊界上任一點或更多點的位置、中心位置、或坐落於果實上的任何位置。而果實的頂部位置是果實與蒂頭交界處,且底部位置是代表果實最底部位置。另一方面,蒂頭採收位置是供外部物件(例如,手、採收裝置170、剪刀等)剪摘的位置。值得注意的是,對於玉女番茄這種農產品,現今消費者通常認為「有蒂頭的番茄才是新鮮的」。而相較於一般自動採收農場直接將果實扭下的作法,本發明實施例是針對番茄蒂頭進行位置判斷,且採收時針對蒂頭摘剪,以保留蒂頭,並讓消費者了解此等採收方式所得的產品可被稱為「新鮮的番茄」。In one embodiment, the position information of the fruits and vegetables includes the position of the fruit and the harvesting position of the stalk, and the position of the fruits includes the top position and the bottom position of the fruit corresponding to the fruits and vegetables. Specifically, the position of the fruit can be any point or more points on the boundary, the center position, or any position on the fruit. The top position of the fruit is the junction of the fruit and the pedicle, and the bottom position represents the bottom position of the fruit. On the other hand, the stalk harvesting position is a position for cutting by external objects (for example, hands,
在一實施例中,控制裝置190依據相對位置關係估測蒂頭採收位置(步驟S231)。具體而言,相對位置關係是基於對應果實的頂部位置及底部位置所得出。基於分析結果可得出,蒂頭採收位置位於頂部位置及底部位置的延伸處,且蒂頭採收位置至頂部位置的距離與頂部位置至底部位置的距離有特定比例關係(例如,1:5、或1:3等)。In one embodiment, the
在一實施例中,相對位置關係包括長度資訊及由底部位置至頂部位置之延伸線。控制裝置190決定對應蔬果的頂部位置及底部位置(例如,基於影像辨識技術)之後,可依據此頂部位置及底部位置取得對應果實的長度資訊。即,長度資訊相關於頂部位置至底部位置的距離。接著,控制裝置190可依據長度資訊及前述延伸線估測蒂頭採收位置。蒂頭保留長度可基於前述比例關係及長度資訊得出。也就是說,蒂頭採收位置是在此延伸線上並與頂部位置相距此蒂頭保留長度的位置。In one embodiment, the relative position relationship includes length information and an extension line from the bottom position to the top position. After the
舉例而言,圖4是依據本發明一實施例的相對位置關係的示意圖。請參照圖4,以番茄10為例,假設其果實11的頂部位置TP的座標為(Xm, Ym),且其底部位置BP的座標為(Xd, Yd)。長度資訊Red_d可由公式(1)得出:…(1)
假設比例關係為1/5,則蒂頭保留長度Green_d可由公式(2)得出:…(2)
假設蒂頭13上的蒂頭採收位置CP位於由底部位置BP至頂部位置TP之延伸線,且設有比例關係。因此,蒂頭採收位置CP至頂部位置TP的距離與頂部位置TP至底部位置BP的距離之間的比例關係將相同於蒂頭採收位置CP至虛擬點VP2(蒂頭採收位置CP垂直延伸線與頂部位置TP水平延伸線之交點)的距離與頂部位置TP至虛擬點VP(頂部位置TP垂直延伸線與底部位置BP水平延伸線之交點)的距離之間的比例關係、或頂部位置TP至虛擬點VP2的距離與底部位置BP至虛擬點VP的距離之間的比例關係。此外,蒂頭採收位置CP對應夾角θ相同於頂部位置TP對應夾角θ。For example, FIG. 4 is a schematic diagram of the relative position relationship according to an embodiment of the present invention. 4, taking
夾角θ可由公式(3)得出:…(3) ,且蒂頭採收位置的座標(Xt, Yt)可由公式(4)得出: …(4)The angle θ can be obtained by formula (3): …(3), and the coordinates (Xt, Yt) of the stalk harvest position can be obtained by formula (4): …(4)
需說明的是,在其他實施例中,控制裝置170也可直接基於影像辨識技術得出蒂頭採收位置,或者基於機器學習演算法估測蒂頭採收位置,且不以此為限。It should be noted that in other embodiments, the
此外,針對成熟程度,在一實施例中,控制裝置190可依據位置資訊及影像中對應果實的顏色分佈決定成熟程度(步驟S233)。以圖5為例,圖5是依據本發明一實施例的成熟程度辨識的示意圖。請參照圖5,控制裝置190由位置資訊得出影像中對應蔬果(圖中為番茄10)的果實11的邊界/輪廓,並將邊界作為興趣區域ROI。接著,控制裝置190可決定蔬果對應的成熟顏色(例如,番茄對應於紅色、或香蕉對應於黃色等),並判斷此成熟顏色的分布情形。此分布情形可相關於像素所占比例(即,成熟顏色對應的像素數占所有果實的像素數的比例)。此比例即可對應於此蔬果的成熟程度。In addition, with regard to the degree of maturity, in one embodiment, the
需說明的是,分布情形不限於比例,在其他實施例中,中位數對應顏色、眾數對應顏色等都可用於判斷成熟程度,且不以此為限。在一些實施例中,控制裝置190可辨識果實大小,並據以決定成熟程度。例如,果實越大越成熟,越小則尚未成熟。It should be noted that the distribution situation is not limited to the ratio. In other embodiments, the color corresponding to the median and the color corresponding to the mode can be used to determine the degree of maturity, and it is not limited thereto. In some embodiments, the
在一實施例中,控制裝置190可將農場區分成數個區域,並依據蔬果的成熟程度估測各區域的產量資訊(步驟S235)。具體而言,控制裝置190可依據位置、種類、種植時間或其他規則來設定區域。此外,控制裝置190可分別統計各區域中不同成熟程度的對應數量。例如,區域A包括50顆三天後成熟的蔬果、及5顆已成熟的蔬果。值得注意的是,控制裝置190也可基於成熟程度預估蔬果的成熟時間。例如,前述成熟顏色的像素數所占比例對應特定預計成熟時間。控制裝置190可基於預估的成熟時間及已成熟數量來估測各區域的產量資訊。此產量資訊相關於當前已成熟數量、未來特定天數將成熟的數量或其組合。而此產量資訊可供農民作為擬訂採收、銷售、或配送策略所用。例如,農民可基於成熟數量或預估成熟時間來決定是否當前進行採收作業、預定未來特定天數後進行採收作業、或指定部分區域進行採收作業。此外,控制裝置190可記錄各蔬果的生長日誌(例如,時間對應成熟度、區域內成熟數量等),以掌握生產過程,並可供與預期情況比較差異。In one embodiment, the
在一實施例中,控制裝置190可控制採收裝置170依據位置資訊移動,並依據蒂頭採收位置採收對應蔬果。除了自影像中得出二維的位置資訊,若影像擷取裝置130配置有深度感測器(例如,基於紅外線、雷達、或立體相機等),則控制裝置190可進一步決定三維的位置資訊。接著,採收裝置170可依據控制裝置190的指令移動至可供採收的蔬果附近(即,蔬果的位置資訊周圍特定範圍內),並透過剪摘部件(例如,機械手臂、剪刀、或夾爪等)在蒂頭採收位置進行剪摘,從而採收此蔬果。In one embodiment, the
在一實施例中,控制裝置190可透過感測裝置110感測農場中的溫度、濕度及照度,並依據溫度、濕度及照度而自規則資料庫得出轉速及除濕開關指示。具體而言,不同類型蔬果的合適生長環境不同。例如,玉女番茄性喜溫暖,其適合生長的溫度為攝氏19-24。若溫度過高(尤其白天溫度過高),則容易引起嚴重落花現象,進而造成結果不良,甚至不完全結果或結的果實不符合預期。而若溫度過低,則果實生育緩慢,且對霜害無抵抗力。玉女番茄果實紅色『茄紅素』的發育適溫在攝氏19-24左右。若溫度超過攝氏30,則會造成番茄發育不良,且果實著色不佳(例如,不呈紅色而呈橙黃色)。另一方面,雖然日照長短對於番茄花芽分化並無顯著關係,但光照強弱對於花芽發育則有影響。例如,若光照強度低弱(低於設定門檻值),則易引起徒長、或落花現象,也容易發生病害,且果實不易肥大。由此可知,溫度、濕度及照度是蔬果成長的重要因素。In one embodiment, the
舉例而言,玉女番茄生長過程中,生長環境對於後續的銷售極為重要,若未記錄玉女番茄生長環境與其成熟度產量之關聯,可能導致訂單供不應求,也容易惹怒消費者。此外,若未針對場域中每一株玉女番茄分析,也可能導致部分區域生長差異過大。For example, during the growth process of Yunv tomato, the growth environment is extremely important for subsequent sales. If the relationship between the growth environment of Yunv tomato and its maturity and yield is not recorded, the supply of orders may fall short of demand and it is easy to anger consumers. In addition, if there is no analysis for each of the virgin tomatoes in the field, it may also lead to excessive growth differences in some areas.
在一實施例中,規則資料庫是基於模糊邏輯(fuzzy logic)。模糊控制系統的構架如圖6所示,圖6是依據本發明一實施例的環境監控的示意圖。請參照圖6,控制裝置190將三種感測裝置130所得之溫度、濕度及照度資料作為模糊系統之輸入,且將輸出設為控制風扇轉速與除濕機開關(假設環境調節裝置150是風扇及除濕機),再針對各輸入定義對應歸屬函數,從而建立規則資料庫。例如,溫度為低、濕度為乾且照度為低,對應歸屬函數為風扇轉速低且除濕開關指示為關;溫度為中、濕度為乾且照度為低,對應歸屬函數為風扇轉速中且除濕開關指示為關。In one embodiment, the rule database is based on fuzzy logic. The structure of the fuzzy control system is shown in FIG. 6, which is a schematic diagram of environmental monitoring according to an embodiment of the present invention. Please refer to FIG. 6, the
透過本發明實施例之模糊推論,可將農民之經驗作為規則資料庫建立之參考,以達到精準的環境控制。此等方式不再只是單純地開關風扇,而是能夠透過模糊規則來控制風扇,以推論出實際情況所需之最佳風量大小,且除濕機的控制亦同。此外,在其他實施例中,控制裝置190還能針對灌溉行程、或其他生長條件進行決策,且不以此為限。Through the fuzzy inference of the embodiment of the present invention, farmers’ experience can be used as a reference for the establishment of a rule database to achieve precise environmental control. These methods are no longer simply switching the fan on and off, but can control the fan through fuzzy rules to infer the optimal air volume required by the actual situation, and the control of the dehumidifier is also the same. In addition, in other embodiments, the
在一實施例中,控制裝置190是透過資料探勘技術找出隱藏的特殊關聯性及特徵。舉例而言,在每株玉女番茄的成長過程,對玉女番茄採用大數據探勘法則,並記錄每株玉女番茄開花數量、已結番茄數、已採收番茄數等結果率。控制裝置190可透過一維碼、二維碼或其他編碼來記錄蔬果的位置資訊、成熟程度、以及各區域的可採收數量(例如,已成熟數量)等,並進一步傳送到其他伺服器以進行大數據分析(亦可能控制裝置190自行執行),從而對成熟時間、產量等資訊進行估測。In one embodiment, the
在一些實施例中,資料進行探勘之前可進行前置處理(Preprocessing),以排除錯誤、遺失、或不完整的資料。接著,可利用諸如C4.5、決策樹(decision tree)、頻繁模式樹(Frequent Pattern Tree,FP tree)、ID3、或分類及回歸樹(Classification And Regression Tree,CART)等演算法建構決策樹模型,最後再代入測試資料進行驗證。此決策樹模型即可供產量及/或成熟時間之評估作業所用。In some embodiments, preprocessing (preprocessing) may be performed before data exploration to eliminate errors, missing, or incomplete data. Then, algorithms such as C4.5, decision tree (decision tree), frequent pattern tree (Frequent Pattern Tree, FP tree), ID3, or classification and regression tree (Classification And Regression Tree, CART) can be used to construct a decision tree model. , And finally substitute the test data for verification. This decision tree model can be used for the evaluation of yield and/or maturity time.
在一實施例中,控制裝置190可提供使用者介面,以供用戶觀看特定區域的環境感測資訊(例如,溫度、濕度及/或照度等)、環境調節裝置150當前運作情形(例如,轉速、開關情形等)、蔬果辨識結果(例如,果實位置、蒂頭採收位置、成熟程度等)及預計採收時間等資訊。In one embodiment, the
在一實施例中,環境調節裝置150更包括氫水氣泡機,以透過噴灑氫水來對抑制蟲害發生,進而減少或避免農藥噴灑。In one embodiment, the
綜上所述,本發明實施例的蔬果監控系統及其方法,偵測果實及蒂頭採收位置,並辨識蔬果的成熟程度。此外,本發明實施例將經驗法則作為環境調節的控制原則。藉此,可讓沒有農場經驗的農民快速上手,也能幫助所有農民預測產量及記錄生長情形,以提早訂定銷售策略並快速建立生產履歷。此外,藉由應用氫水氣泡機在農場,可減少或避免農藥噴灑。In summary, the fruit and vegetable monitoring system and method of the embodiments of the present invention detect the harvesting position of the fruit and the stalk, and identify the ripeness of the fruit and vegetable. In addition, the embodiment of the present invention uses the rule of thumb as the control principle of environmental regulation. In this way, farmers without farm experience can get started quickly, and it can also help all farmers predict yields and record growth conditions, so as to formulate sales strategies early and quickly build production resumes. In addition, by using the hydrogen water bubble machine in the farm, pesticide spraying can be reduced or avoided.
雖然本發明已以實施例揭露如上,然其並非用以限定本發明,任何所屬技術領域中具有通常知識者,在不脫離本發明的精神和範圍內,當可作些許的更動與潤飾,故本發明的保護範圍當視後附的申請專利範圍所界定者為準。Although the present invention has been disclosed in the above embodiments, it is not intended to limit the present invention. Anyone with ordinary knowledge in the relevant technical field can make some changes and modifications without departing from the spirit and scope of the present invention. The protection scope of the present invention shall be subject to those defined by the attached patent application scope.
100:蔬果監控系統 110:感測裝置 130:影像擷取裝置 150:環境調節裝置 170:採收裝置 190:控制裝置 S210~S235:步驟 10:番茄 11:果實 13:蒂頭 TP:頂部位置 BP:底部位置 CP:蒂頭採收位置 VP、VP2:虛擬點 θ:夾角 ROI:興趣區域100: Vegetable and Fruit Monitoring System 110: sensing device 130: Image capture device 150: Environmental regulation device 170: Harvesting device 190: control device S210~S235: steps 10: Tomato 11: Fruit 13: pedicle TP: Top position BP: bottom position CP: Picking position of stalk VP, VP2: virtual point θ: included angle ROI: region of interest
圖1是依據本發明一實施例蔬果監控系統的方塊圖。 圖2是依據本發明一實施例蔬果監控方法的流程圖。 圖3是一範例說明果實辨識結果。 圖4是依據本發明一實施例的相對位置關係的示意圖。 圖5是依據本發明一實施例的成熟程度辨識的示意圖。 圖6是依據本發明一實施例的環境監控的示意圖。Fig. 1 is a block diagram of a vegetable and fruit monitoring system according to an embodiment of the present invention. Fig. 2 is a flowchart of a method for monitoring fruits and vegetables according to an embodiment of the present invention. Figure 3 is an example illustrating the results of fruit identification. Fig. 4 is a schematic diagram of a relative position relationship according to an embodiment of the present invention. FIG. 5 is a schematic diagram of maturity recognition according to an embodiment of the present invention. Fig. 6 is a schematic diagram of environmental monitoring according to an embodiment of the present invention.
S210~S235:步驟S210~S235: steps
Claims (10)
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
TW108145637A TW202123152A (en) | 2019-12-13 | 2019-12-13 | Fruits and vegetables monitoring system and method thereof |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
TW108145637A TW202123152A (en) | 2019-12-13 | 2019-12-13 | Fruits and vegetables monitoring system and method thereof |
Publications (1)
Publication Number | Publication Date |
---|---|
TW202123152A true TW202123152A (en) | 2021-06-16 |
Family
ID=77516552
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
TW108145637A TW202123152A (en) | 2019-12-13 | 2019-12-13 | Fruits and vegetables monitoring system and method thereof |
Country Status (1)
Country | Link |
---|---|
TW (1) | TW202123152A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114253322A (en) * | 2021-12-08 | 2022-03-29 | 黑龙江省天鹏翔羽数据科技有限公司 | Intelligent agricultural data acquisition and analysis system and method |
TWI793051B (en) * | 2022-07-28 | 2023-02-11 | 國立虎尾科技大學 | Monitoring System and Method of Internet of Things for Mushroom Cultivation |
-
2019
- 2019-12-13 TW TW108145637A patent/TW202123152A/en unknown
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114253322A (en) * | 2021-12-08 | 2022-03-29 | 黑龙江省天鹏翔羽数据科技有限公司 | Intelligent agricultural data acquisition and analysis system and method |
TWI793051B (en) * | 2022-07-28 | 2023-02-11 | 國立虎尾科技大學 | Monitoring System and Method of Internet of Things for Mushroom Cultivation |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11744189B2 (en) | Plant treatment based on morphological and physiological measurements | |
Narvaez et al. | A survey of ranging and imaging techniques for precision agriculture phenotyping | |
Montoya-Cavero et al. | Vision systems for harvesting robots: Produce detection and localization | |
CN111418349B (en) | Intelligent fruit picking robot and method for realizing fruit picking | |
WO2023050783A1 (en) | Weeding robot and method and apparatus for planning weeding path thereof, and medium | |
TWI661770B (en) | Intelligent deep learning agricultural and fishery training system | |
Lin et al. | A review on computer vision technologies applied in greenhouse plant stress detection | |
Miao et al. | Efficient tomato harvesting robot based on image processing and deep learning | |
TW202123152A (en) | Fruits and vegetables monitoring system and method thereof | |
KR102265860B1 (en) | Artificial Intelligence Robot System for Unmanned Crop Cultivation and Harvesting | |
Yoshida et al. | Fruit recognition method for a harvesting robot with RGB-D cameras | |
Dharmasena et al. | Autonomous cloud robotic system for smart agriculture | |
KR102291827B1 (en) | System and method for automatic crop growth measurement in smart farm | |
Cho et al. | Plant growth information measurement based on object detection and image fusion using a smart farm robot | |
WO2022091092A1 (en) | System and method for indoor crop management | |
You et al. | Semiautonomous Precision Pruning of Upright Fruiting Offshoot Orchard Systems: An Integrated Approach | |
KR102371909B1 (en) | Smart farm crop monitoring method and system using drones | |
US20230389474A1 (en) | Method for determining a fruit to be harvested and a device for harvesting a fruit | |
US20230133026A1 (en) | Sparse and/or dense depth estimation from stereoscopic imaging | |
Silwal | Machine vision system for robotic apple harvesting in fruiting wall orchards | |
Kabilesh et al. | Research on artificial intelligence based fruit disease identification system (AI-FDIS) with the internet of things (IoT) | |
Barhate et al. | Comprehensive study on automated image detection by using robotics for agriculture applications | |
KR20230061863A (en) | Apparatus for predicting fruit development stage using ensemble model of convolutional neural network and multi layer perceptron and method thereof | |
JP2021185758A (en) | Crop harvesting system and crop harvesting apparatus | |
Hughes et al. | Field robotics for harvesting: A review of field robotics approaches for harvesting |