前言

人工智能的浪潮正在席卷全球。一個(gè)已經(jīng)被談?wù)摿藥资甑母拍睿缃襁@幾年，相關(guān)技術(shù)的發(fā)展速度越來(lái)越快。機(jī)器學(xué)習(xí)、深度學(xué)習(xí)、計(jì)算機(jī)視覺(jué)等名詞逐漸走進(jìn)人們的生活，它們同屬于人工智能的范疇之中。

計(jì)算機(jī)視覺(jué)是人工智能領(lǐng)域的一個(gè)分支計(jì)算機(jī)視覺(jué)實(shí)際上是一個(gè)跨領(lǐng)域的交叉學(xué)科，包括計(jì)算機(jī)科學(xué)，數(shù)學(xué)，工程學(xué)，物理學(xué)，生物學(xué)和心理學(xué)等領(lǐng)域。許多科學(xué)家認(rèn)為，計(jì)算機(jī)視覺(jué)為人工智能的發(fā)展開(kāi)拓了道路。

簡(jiǎn)單來(lái)說(shuō)，計(jì)算機(jī)視覺(jué)就是賦予計(jì)算機(jī)一雙觀察世界的眼睛，再使用計(jì)算機(jī)優(yōu)秀的大腦快速的計(jì)算，服務(wù)人類(lèi)。

今天我們將深入淺出，簡(jiǎn)單介紹Python/ target=_blank class=infotextkey>Python計(jì)算機(jī)視覺(jué)中的手勢(shì)識(shí)別方法，識(shí)別手勢(shì)——數(shù)字（一、二、三、四、五和大拇指的贊賞）。

前期準(zhǔn)備

本篇我們將會(huì)用到Python的OpenCV模塊和手部模型模塊mediapipe，在Python的pip安裝方法中，安裝方法如下：

opencv是常用的圖像識(shí)別模塊

mediapipe是谷歌開(kāi)發(fā)并開(kāi)源的多媒體機(jī)器學(xué)習(xí)模型應(yīng)用框架。

pip install opencv-python
pip install mediapipe

如果你的電腦裝有Anaconda，建議選擇在Anaconda的環(huán)境命令行中進(jìn)行相應(yīng)模塊的安裝，以此構(gòu)建更具體的機(jī)器學(xué)習(xí)環(huán)境

當(dāng)你安裝好OpenCV和mediapipe模塊以后，你可以在Python代碼中寫(xiě)入

import cv2
import mediapipe as mp

如果運(yùn)行成功，那么你的Opencv-python模塊即為安裝成功，那么我們現(xiàn)在就開(kāi)始進(jìn)入今天的正題吧！

識(shí)別手部模型

既然要做手勢(shì)識(shí)別，那么就要去找到我們傳入圖像的手部信息。本處我們將使用mediapipe模型去找到手部模型，并完成手部模型的識(shí)別模塊，并命名，我們將在后續(xù)手勢(shì)識(shí)別內(nèi)容中將其作為模塊引入

HandTrackingModule.py

# -*- coding:utf-8 -*-

"""

CODE >>> SINCE IN CAIXYPROMISE.
MOTTO >>> STRIVE FOR ExcelLENT.
CONSTANTLY STRIVING FOR SELF-IMPROVEMENT.

@ By: CaixyPromise
@ Date: 2021-10-17

"""
import cv2
import mediapipe as mp

class HandDetector:
    """
    使用mediapipe庫(kù)查找手。導(dǎo)出地標(biāo)像素格式。添加了額外的功能。
    如查找方式，許多手指向上或兩個(gè)手指之間的距離。而且提供找到的手的邊界框信息。
    """
    def __init__(self, mode=False, maxHands=2, detectionCon=0.5, minTrackCon=0.5):
        """
        :param mode: 在靜態(tài)模式下，對(duì)每個(gè)圖像進(jìn)行檢測(cè)
        :param maxHands: 要檢測(cè)的最大手?jǐn)?shù)
        :param detectionCon: 最小檢測(cè)置信度
        :param minTrackCon: 最小跟蹤置信度
        """
        self.mode = mode
        self.maxHands = maxHands
        self.detectionCon = detectionCon
        self.minTrackCon = minTrackCon

        self.mpHands = mp.solutions.hands
        self.hands = self.mpHands.Hands(self.mode, self.maxHands,
                                        self.detectionCon, self.minTrackCon)
        self.mpDraw = mp.solutions.drawing_utils
        self.tipIds = [4, 8, 12, 16, 20]
        self.fingers = []
        self.lmList = []

    def findHands(self, img, draw=True):
        """
        從圖像(BRG)中找到手部。
        :param img: 用于查找手的圖像。
        :param draw: 在圖像上繪制輸出的標(biāo)志。
        :return: 帶或不帶圖形的圖像
        """
        imgRGB = cv2.cvtColor(img, cv2.COLOR_BGR2RGB) # 將傳入的圖像由BGR模式轉(zhuǎn)標(biāo)準(zhǔn)的Opencv模式——RGB模式，
        self.results = self.hands.process(imgRGB)

        if self.results.multi_hand_landmarks:
            for handLms in self.results.multi_hand_landmarks:
                if draw:
                    self.mpDraw.draw_landmarks(img, handLms,
                                               self.mpHands.HAND_CONNECTIONS)
        return img

    def findPosition(self, img, handNo=0, draw=True):
        """
        查找單手的地標(biāo)并將其放入列表中像素格式。還可以返回手部周?chē)倪吔缈颉?        :param img: 要查找的主圖像
        :param handNo: 如果檢測(cè)到多只手，則為手部id
        :param draw: 在圖像上繪制輸出的標(biāo)志。(默認(rèn)繪制矩形框)
        :return: 像素格式的手部關(guān)節(jié)位置列表；手部邊界框
        """

        xList = []
        yList = []
        bbox = []
        bboxInfo =[]
        self.lmList = []
        if self.results.multi_hand_landmarks:
            myHand = self.results.multi_hand_landmarks[handNo]
            for id, lm in enumerate(myHand.landmark):
                h, w, c = img.shape
                px, py = int(lm.x * w), int(lm.y * h)
                xList.Append(px)
                yList.append(py)
                self.lmList.append([px, py])
                if draw:
                    cv2.circle(img, (px, py), 5, (255, 0, 255), cv2.FILLED)
            xmin, xmax = min(xList), max(xList)
            ymin, ymax = min(yList), max(yList)
            boxW, boxH = xmax - xmin, ymax - ymin
            bbox = xmin, ymin, boxW, boxH
            cx, cy = bbox[0] + (bbox[2] // 2), 
                     bbox[1] + (bbox[3] // 2)
            bboxInfo = {"id": id, "bbox": bbox,"center": (cx, cy)}

            if draw:
                cv2.rectangle(img, (bbox[0] - 20, bbox[1] - 20),
                              (bbox[0] + bbox[2] + 20, bbox[1] + bbox[3] + 20),
                              (0, 255, 0), 2)

        return self.lmList, bboxInfo

    def fingersUp(self):
        """
        查找列表中打開(kāi)并返回的手指數(shù)。會(huì)分別考慮左手和右手
        ：return：豎起手指的數(shù)組(列表)，數(shù)組長(zhǎng)度為5，
        其中，由大拇指開(kāi)始數(shù)，立起標(biāo)為1，放下為0。
        """
        if self.results.multi_hand_landmarks:
            myHandType = self.handType()
            fingers = []
            # Thumb
            if myHandType == "Right":
                if self.lmList[self.tipIds[0]][0] > self.lmList[self.tipIds[0] - 1][0]:
                    fingers.append(1)
                else:
                    fingers.append(0)
            else:
                if self.lmList[self.tipIds[0]][0] < self.lmList[self.tipIds[0] - 1][0]:
                    fingers.append(1)
                else:
                    fingers.append(0)

            # 4 Fingers
            for id in range(1, 5):
                if self.lmList[self.tipIds[id]][1] < self.lmList[self.tipIds[id] - 2][1]:
                    fingers.append(1)
                else:
                    fingers.append(0)
        return fingers

    def handType(self):
        """
        檢查傳入的手部是左還是右
        ：return: "Right" 或 "Left"
        """
        if self.results.multi_hand_landmarks:
            if self.lmList[17][0] < self.lmList[5][0]:
                return "Right"
            else:
                return "Left"

識(shí)別視頻輸入方法

完成手部模型的獲取與識(shí)別，現(xiàn)在我們就要將內(nèi)容傳入到計(jì)算機(jī)當(dāng)中，使其能進(jìn)行手部的識(shí)別以及手勢(shì)的識(shí)別。本處我們將使用OpenCV進(jìn)行內(nèi)容的輸入流，開(kāi)啟計(jì)算機(jī)的攝像頭獲取內(nèi)容，并使用剛剛我們寫(xiě)的HandTrackingModule模塊作為手部的識(shí)別模塊。

Main.py

# -*- coding:utf-8 -*-

"""

CODE >>> SINCE IN CAIXYPROMISE.
MOTTO >>> STRIVE FOR EXCELLENT.
CONSTANTLY STRIVING FOR SELF-IMPROVEMENT.

@ By: CaixyPromise
@ Date: 2021-10-17

"""
import cv2
from HandTrackingModule import HandDetector

class Main:
    def __init__(self):
        self.camera = cv2.VideoCapture(0,cv2.CAP_DSHOW) # 以視頻流傳入
        self.camera.set(3, 1280) # 設(shè)置分辨率
        self.camera.set(4, 720)
        
    def Gesture_recognition(self):
        while True:
            self.detector = HandDetector()
            frame, img = self.camera.read()
            img = self.detector.findHands(img) # 找到你的手部
            lmList, bbox = self.detector.findPosition(img) # 獲取你手部的方位
             
            cv2.imshow("camera", img)
            if cv2.getWindowProperty('camera', cv2.WND_PROP_VISIBLE) < 1:
                break
            # 通過(guò)關(guān)閉按鈕退出程序
            cv2.waitKey(1)   
            # if cv2.waitKey(1) & 0xFF == ord("q"):
            #     break # 按下q退出

現(xiàn)在，當(dāng)我們運(yùn)行程序后，程序會(huì)運(yùn)行你的計(jì)算機(jī)默認(rèn)攝像頭，當(dāng)你露出你的手時(shí)，會(huì)傳出圖像圈住你的手部，并且繪制出你的手部主要關(guān)節(jié)點(diǎn)。

其中，你的手部主要關(guān)節(jié)點(diǎn)已經(jīng)標(biāo)好序號(hào)，你的手部分為了21個(gè)關(guān)節(jié)點(diǎn)，指尖分別為4 8 12 16 20

具體關(guān)節(jié)分為:

手勢(shì)識(shí)別方法

通過(guò)前面的講解，我們完成了手部獲取與識(shí)別、識(shí)別內(nèi)容的輸入，那么我們現(xiàn)在就來(lái)開(kāi)始寫(xiě)我們的手勢(shì)識(shí)別方法。這里，我們用到識(shí)別模塊里的fingersUp()方法。

找到我們剛剛寫(xiě)的Main.py文件(識(shí)別內(nèi)容輸入方法)，當(dāng)我們找到并繪制出我們的手部位置以后，此時(shí)的findPosition()方法會(huì)得到你的手部具體方位，其中l(wèi)mList是關(guān)節(jié)位置方位(type:list)，bbox是邊框方位(type:dict)，當(dāng)未識(shí)別到內(nèi)容時(shí)兩者均為空。所以，我們只需要寫(xiě)當(dāng)數(shù)組中存在數(shù)據(jù)（非空），進(jìn)行手指判斷即可，那么我們可以寫(xiě)成

# -*- coding:utf-8 -*-

"""

CODE >>> SINCE IN CAIXYPROMISE.
MOTTO >>> STRIVE FOR EXCELLENT.
CONSTANTLY STRIVING FOR SELF-IMPROVEMENT.

@ By: CaixyPromise
@ Date: 2021-10-17

"""
def Gesture_recognition(self):
    while True:
        self.detector = HandDetector()
        frame, img = self.camera.read()
        img = self.detector.findHands(img)
        lmList, bbox = self.detector.findPosition(img)
        
        if lmList:
          x1, x2, x3, x4, x5 = self.detector.fingersUp()

上面我們fingersUp()方法談到，fingersUp()方法會(huì)傳回從大拇指開(kāi)始數(shù)的長(zhǎng)度為5的數(shù)組，立起的手指標(biāo)記為1，放下標(biāo)記為0。

本次我們的目的是寫(xiě)一個(gè)識(shí)別我們生活常見(jiàn)的數(shù)字手勢(shì)以及一個(gè)贊揚(yáng)大拇指的手勢(shì)。結(jié)合我們生活，識(shí)別你的手勢(shì)可以寫(xiě)為

# -*- coding:utf-8 -*-

"""

CODE >>> SINCE IN CAIXYPROMISE.
MOTTO >>> STRIVE FOR EXCELLENT.
CONSTANTLY STRIVING FOR SELF-IMPROVEMENT.

@ By: CaixyPromise
@ Date: 2021-10-17

"""
def Gesture_recognition(self):
    while True:
        self.detector = HandDetector()
        frame, img = self.camera.read()
        img = self.detector.findHands(img)
        lmList, bbox = self.detector.findPosition(img)
        
        if lmList:
          x1, x2, x3, x4, x5 = self.detector.fingersUp()
          if (x2 == 1 and x3 == 1) and (x4 == 0 and x5 == 0 and x1 == 0):
            # TWO
          elif (x2 == 1 and x3 == 1 and x4 == 1) and (x1 == 0 and x5 == 0):
            # THREE
          elif (x2 == 1 and x3 == 1 and x4 == 1 and x5 == 1) and (x1 == 0):
            # FOUR
          elif x1 == 1 and x2 == 1 and x3 == 1 and x4 == 1 and x5 == 1:
            # FIVE
          elif x2 == 1 and (x1 == 0, x3 == 0, x4 == 0, x5 == 0):
            # ONE
          elif x1 and (x2 == 0, x3 == 0, x4 == 0, x5 == 0):
            # NICE_GOOD

完成基本的識(shí)別以后，我們要把內(nèi)容表達(dá)出來(lái)。這里我們結(jié)合bbox返回的手部方框方位，再使用opencv里的putText方法，實(shí)現(xiàn)識(shí)別結(jié)果的輸出。

# -*- coding:utf-8 -*-

"""

CODE >>> SINCE IN CAIXYPROMISE.
MOTTO >>> STRIVE FOR EXCELLENT.
CONSTANTLY STRIVING FOR SELF-IMPROVEMENT.

@ By: CaixyPromise
@ Date: 2021-10-17

"""
def Gesture_recognition(self):
    while True:
        self.detector = HandDetector()
        frame, img = self.camera.read()
        img = self.detector.findHands(img)
        lmList, bbox = self.detector.findPosition(img)

        if lmList:
            x_1, y_1 = bbox["bbox"][0], bbox["bbox"][1]
            x1, x2, x3, x4, x5 = self.detector.fingersUp()

            if (x2 == 1 and x3 == 1) and (x4 == 0 and x5 == 0 and x1 == 0):
                cv2.putText(img, "2_TWO", (x_1, y_1), cv2.FONT_HERSHEY_PLAIN, 3,
                            (0, 0, 255), 3)
            elif (x2 == 1 and x3 == 1 and x4 == 1) and (x1 == 0 and x5 == 0):
                cv2.putText(img, "3_THREE", (x_1, y_1), cv2.FONT_HERSHEY_PLAIN, 3,
                            (0, 0, 255), 3)
            elif (x2 == 1 and x3 == 1 and x4 == 1 and x5 == 1) and (x1 == 0):
                cv2.putText(img, "4_FOUR", (x_1, y_1), cv2.FONT_HERSHEY_PLAIN, 3,
                            (0, 0, 255), 3)
            elif x1 == 1 and x2 == 1 and x3 == 1 and x4 == 1 and x5 == 1:
                cv2.putText(img, "5_FIVE", (x_1, y_1), cv2.FONT_HERSHEY_PLAIN, 3,
                            (0, 0, 255), 3)
            elif x2 == 1 and (x1 == 0, x3 == 0, x4 == 0, x5 == 0):
                cv2.putText(img, "1_ONE", (x_1, y_1), cv2.FONT_HERSHEY_PLAIN, 3,
                            (0, 0, 255), 3)
            elif x1 and (x2 == 0, x3 == 0, x4 == 0, x5 == 0):
                cv2.putText(img, "GOOD!", (x_1, y_1), cv2.FONT_HERSHEY_PLAIN, 3,
                            (0, 0, 255), 3)
        cv2.imshow("camera", img)
        if cv2.getWindowProperty('camera', cv2.WND_PROP_VISIBLE) < 1:
            break
        cv2.waitKey(1)

現(xiàn)在，我們已經(jīng)完成手勢(shì)的識(shí)別與結(jié)果輸出，我們把完整代碼運(yùn)行一下，即可驗(yàn)證出我們的代碼效果。

完整代碼

完整代碼如下

# -*- coding:utf-8 -*-

"""

CODE >>> SINCE IN CAIXYPROMISE.
STRIVE FOR EXCELLENT.
CONSTANTLY STRIVING FOR SELF-IMPROVEMENT.
@ by: caixy
@ date: 2021-10-1

"""

import cv2
from HandTrackingModule import HandDetector

class Main:
    def __init__(self):
        self.camera = cv2.VideoCapture(0,cv2.CAP_DSHOW)
        self.camera.set(3, 1280)
        self.camera.set(4, 720)

    def Gesture_recognition(self):
        while True:
            self.detector = HandDetector()
            frame, img = self.camera.read()
            img = self.detector.findHands(img)
            lmList, bbox = self.detector.findPosition(img)

            if lmList:
                x_1, y_1 = bbox["bbox"][0], bbox["bbox"][1]
                x1, x2, x3, x4, x5 = self.detector.fingersUp()

                if (x2 == 1 and x3 == 1) and (x4 == 0 and x5 == 0 and x1 == 0):
                    cv2.putText(img, "2_TWO", (x_1, y_1), cv2.FONT_HERSHEY_PLAIN, 3,
                                (0, 0, 255), 3)
                elif (x2 == 1 and x3 == 1 and x4 == 1) and (x1 == 0 and x5 == 0):
                    cv2.putText(img, "3_THREE", (x_1, y_1), cv2.FONT_HERSHEY_PLAIN, 3,
                                (0, 0, 255), 3)
                elif (x2 == 1 and x3 == 1 and x4 == 1 and x5 == 1) and (x1 == 0):
                    cv2.putText(img, "4_FOUR", (x_1, y_1), cv2.FONT_HERSHEY_PLAIN, 3,
                                (0, 0, 255), 3)
                elif x1 == 1 and x2 == 1 and x3 == 1 and x4 == 1 and x5 == 1:
                    cv2.putText(img, "5_FIVE", (x_1, y_1), cv2.FONT_HERSHEY_PLAIN, 3,
                                (0, 0, 255), 3)
                elif x2 == 1 and (x1 == 0, x3 == 0, x4 == 0, x5 == 0):
                    cv2.putText(img, "1_ONE", (x_1, y_1), cv2.FONT_HERSHEY_PLAIN, 3,
                                (0, 0, 255), 3)
                elif x1 and (x2 == 0, x3 == 0, x4 == 0, x5 == 0):
                    cv2.putText(img, "GOOD!", (x_1, y_1), cv2.FONT_HERSHEY_PLAIN, 3,
                                (0, 0, 255), 3)
            cv2.imshow("camera", img)
            if cv2.getWindowProperty('camera', cv2.WND_PROP_VISIBLE) < 1:
                break
            cv2.waitKey(1)
            # if cv2.waitKey(1) & 0xFF == ord("q"):
            #     break

if __name__ == '__main__':
    Solution = Main()
    Solution.Gesture_recognition()

效果一目了然，計(jì)算機(jī)成功識(shí)別了你的手勢(shì)并把內(nèi)容輸出。快去試試吧！