使用|使用 OpenCV-Python 识别答题卡判卷功能使用OpenCV-Python识别答题卡判卷

任务识别用相机拍下来的答题卡，并判断最终得分（假设正确答案是B, E, A, D, B）

文章图片

主要步骤

轮廓识别——答题卡边缘识别
透视变换——提取答题卡主体
轮廓识别——识别出所有圆形选项，剔除无关轮廓
检测每一行选择的是哪一项，并将结果储存起来，记录正确的个数
计算最终得分并在图中标注

分步实现轮廓识别——答题卡边缘识别
输入图像

import cv2 as cvimport numpy as np # 正确答案right_key = {0: 1, 1: 4, 2: 0, 3: 3, 4: 1} # 输入图像img = cv.imread('./images/test_01.jpg')img_copy = img.copy()img_gray = cv.cvtColor(img, cv.COLOR_BGR2GRAY)cvshow('img-gray', img_gray)

文章图片

图像预处理

# 图像预处理# 高斯降噪img_gaussian = cv.GaussianBlur(img_gray, (5, 5), 1)cvshow('gaussianblur', img_gaussian)# canny边缘检测img_canny = cv.Canny(img_gaussian, 80, 150)cvshow('canny', img_canny)

文章图片

文章图片

轮廓识别——答题卡边缘识别

# 轮廓识别——答题卡边缘识别cnts, hierarchy = cv.findContours(img_canny, cv.RETR_EXTERNAL, cv.CHAIN_APPROX_SIMPLE)cv.drawContours(img_copy, cnts, -1, (0, 0, 255), 3)cvshow('contours-show', img_copy)

文章图片

透视变换——提取答题卡主体
对每个轮廓进行拟合，将多边形轮廓变为四边形

docCnt = None # 确保检测到了if len(cnts) > 0:# 根据轮廓大小进行排序cnts = sorted(cnts, key=cv.contourArea, reverse=True) # 遍历每一个轮廓for c in cnts:# 近似peri = cv.arcLength(c, True)# arclength 计算一段曲线的长度或者闭合曲线的周长；# 第一个参数输入一个二维向量，第二个参数表示计算曲线是否闭合 approx = cv.approxPolyDP(c, 0.02 * peri, True)# 用一条顶点较少的曲线/多边形来近似曲线/多边形，以使它们之间的距离<=指定的精度；# c是需要近似的曲线，0.02*peri是精度的最大值，True表示曲线是闭合的 # 准备做透视变换if len(approx) == 4:docCnt = approxbreak

透视变换——提取答题卡主体

# 透视变换——提取答题卡主体docCnt = docCnt.reshape(4, 2)warped = four_point_transform(img_gray, docCnt)cvshow('warped', warped)

def four_point_transform(img, four_points):rect = order_points(four_points)(tl, tr, br, bl) = rect # 计算输入的w和h的值widthA = np.sqrt((tr[0] - tl[0]) ** 2 + (tr[1] - tl[1]) ** 2)widthB = np.sqrt((br[0] - bl[0]) ** 2 + (br[1] - bl[1]) ** 2)maxWidth = max(int(widthA), int(widthB)) heightA = np.sqrt((tl[0] - bl[0]) ** 2 + (tl[1] - bl[1]) ** 2)heightB = np.sqrt((tr[0] - br[0]) ** 2 + (tr[1] - br[1]) ** 2)maxHeight = max(int(heightA), int(heightB)) # 变换后对应的坐标位置dst = np.array([[0, 0],[maxWidth - 1, 0],[maxWidth - 1, maxHeight - 1],[0, maxHeight - 1]], dtype='float32') # 最主要的函数就是 cv2.getPerspectiveTransform(rect, dst) 和 cv2.warpPerspective(image, M, (maxWidth, maxHeight))M = cv.getPerspectiveTransform(rect, dst)warped = cv.warpPerspective(img, M, (maxWidth, maxHeight))return warped def order_points(points):res = np.zeros((4, 2), dtype='float32')# 按照从前往后0，1，2，3分别表示左上、右上、右下、左下的顺序将points中的数填入res中 # 将四个坐标x与y相加，和最大的那个是右下角的坐标，最小的那个是左上角的坐标sum_hang = points.sum(axis=1)res[0] = points[np.argmin(sum_hang)]res[2] = points[np.argmax(sum_hang)] # 计算坐标x与y的离散插值np.diff()diff = np.diff(points, axis=1)res[1] = points[np.argmin(diff)]res[3] = points[np.argmax(diff)] # 返回resultreturn res

文章图片

轮廓识别——识别出选项

# 轮廓识别——识别出选项thresh = cv.threshold(warped, 0, 255, cv.THRESH_BINARY_INV | cv.THRESH_OTSU)[1]cvshow('thresh', thresh)thresh_cnts, _ = cv.findContours(thresh, cv.RETR_EXTERNAL, cv.CHAIN_APPROX_SIMPLE)w_copy = warped.copy()cv.drawContours(w_copy, thresh_cnts, -1, (0, 0, 255), 2)cvshow('warped_contours', w_copy) questionCnts = []# 遍历，挑出选项的cntsfor c in thresh_cnts:(x, y, w, h) = cv.boundingRect(c)ar = w / float(h)# 根据实际情况指定标准if w >= 20 and h >= 20 and ar >= 0.9 and ar <= 1.1:questionCnts.append(c) # 检查是否挑出了选项w_copy2 = warped.copy()cv.drawContours(w_copy2, questionCnts, -1, (0, 0, 255), 2)cvshow('questionCnts', w_copy2)

文章图片

文章图片

成功将无关轮廓剔除
检测每一行选择的是哪一项，并将结果储存起来，记录正确的个数

# 检测每一行选择的是哪一项，并将结果储存在元组bubble中，记录正确的个数correct# 按照从上到下t2b对轮廓进行排序questionCnts = sort_contours(questionCnts, method="t2b")[0]correct = 0# 每行有5个选项for (i, q) in enumerate(np.arange(0, len(questionCnts), 5)):# 排序cnts = sort_contours(questionCnts[q:q+5])[0] bubble = None# 得到每一个选项的mask并填充，与正确答案进行按位与操作获得重合点数for (j, c) in enumerate(cnts):mask = np.zeros(thresh.shape, dtype='uint8')cv.drawContours(mask, [c], -1, 255, -1)# cvshow('mask', mask) # 通过按位与操作得到thresh与mask重合部分的像素数量bitand = cv.bitwise_and(thresh, thresh, mask=mask)totalPixel = cv.countNonZero(bitand) if bubble is None or bubble[0] < totalPixel:bubble = (totalPixel, j) k = bubble[1]color = (0, 0, 255)if k == right_key[i]:correct += 1color = (0, 255, 0) # 绘图cv.drawContours(warped, [cnts[right_key[i]]], -1, color, 3)cvshow('final', warped)

def sort_contours(contours, method="l2r"):# 用于给轮廓排序，l2r, r2l, t2b, b2treverse = Falsei = 0if method == "r2l" or method == "b2t":reverse = Trueif method == "t2b" or method == "b2t":i = 1 boundingBoxes = [cv.boundingRect(c) for c in contours](contours, boundingBoxes) = zip(*sorted(zip(contours, boundingBoxes), key=lambda a: a[1][i], reverse=reverse))return contours, boundingBoxes

文章图片

用透过mask的像素的个数来判断考生选择的是哪个选项

文章图片

计算最终得分并在图中标注

# 计算最终得分并在图中标注score = (correct / 5.0) * 100print(f"Score: {score}%")cv.putText(warped, f"Score: {score}%", (10, 30), cv.FONT_HERSHEY_SIMPLEX, 0.9, (0, 0, 255), 2)cv.imshow("Original", img)cv.imshow("Exam", warped)cv.waitKey(0)

文章图片

完整代码

import cv2 as cvimport numpy as np def cvshow(name, img):cv.imshow(name, img)cv.waitKey(0)cv.destroyAllWindows() def four_point_transform(img, four_points):rect = order_points(four_points)(tl, tr, br, bl) = rect # 计算输入的w和h的值widthA = np.sqrt((tr[0] - tl[0]) ** 2 + (tr[1] - tl[1]) ** 2)widthB = np.sqrt((br[0] - bl[0]) ** 2 + (br[1] - bl[1]) ** 2)maxWidth = max(int(widthA), int(widthB)) heightA = np.sqrt((tl[0] - bl[0]) ** 2 + (tl[1] - bl[1]) ** 2)heightB = np.sqrt((tr[0] - br[0]) ** 2 + (tr[1] - br[1]) ** 2)maxHeight = max(int(heightA), int(heightB)) # 变换后对应的坐标位置dst = np.array([[0, 0],[maxWidth - 1, 0],[maxWidth - 1, maxHeight - 1],[0, maxHeight - 1]], dtype='float32') # 最主要的函数就是 cv2.getPerspectiveTransform(rect, dst) 和 cv2.warpPerspective(image, M, (maxWidth, maxHeight))M = cv.getPerspectiveTransform(rect, dst)warped = cv.warpPerspective(img, M, (maxWidth, maxHeight))return warped def order_points(points):res = np.zeros((4, 2), dtype='float32')# 按照从前往后0，1，2，3分别表示左上、右上、右下、左下的顺序将points中的数填入res中 # 将四个坐标x与y相加，和最大的那个是右下角的坐标，最小的那个是左上角的坐标sum_hang = points.sum(axis=1)res[0] = points[np.argmin(sum_hang)]res[2] = points[np.argmax(sum_hang)] # 计算坐标x与y的离散插值np.diff()diff = np.diff(points, axis=1)res[1] = points[np.argmin(diff)]res[3] = points[np.argmax(diff)] # 返回resultreturn res def sort_contours(contours, method="l2r"):# 用于给轮廓排序，l2r, r2l, t2b, b2treverse = Falsei = 0if method == "r2l" or method == "b2t":reverse = Trueif method == "t2b" or method == "b2t":i = 1 boundingBoxes = [cv.boundingRect(c) for c in contours](contours, boundingBoxes) = zip(*sorted(zip(contours, boundingBoxes), key=lambda a: a[1][i], reverse=reverse))return contours, boundingBoxes # 正确答案right_key = {0: 1, 1: 4, 2: 0, 3: 3, 4: 1} # 输入图像img = cv.imread('./images/test_01.jpg')img_copy = img.copy()img_gray = cv.cvtColor(img, cv.COLOR_BGR2GRAY)cvshow('img-gray', img_gray) # 图像预处理# 高斯降噪img_gaussian = cv.GaussianBlur(img_gray, (5, 5), 1)cvshow('gaussianblur', img_gaussian)# canny边缘检测img_canny = cv.Canny(img_gaussian, 80, 150)cvshow('canny', img_canny) # 轮廓识别——答题卡边缘识别cnts, hierarchy = cv.findContours(img_canny, cv.RETR_EXTERNAL, cv.CHAIN_APPROX_SIMPLE)cv.drawContours(img_copy, cnts, -1, (0, 0, 255), 3)cvshow('contours-show', img_copy) docCnt = None # 确保检测到了if len(cnts) > 0:# 根据轮廓大小进行排序cnts = sorted(cnts, key=cv.contourArea, reverse=True) # 遍历每一个轮廓for c in cnts:# 近似peri = cv.arcLength(c, True)# arclength 计算一段曲线的长度或者闭合曲线的周长；# 第一个参数输入一个二维向量，第二个参数表示计算曲线是否闭合 approx = cv.approxPolyDP(c, 0.02 * peri, True)# 用一条顶点较少的曲线/多边形来近似曲线/多边形，以使它们之间的距离<=指定的精度；# c是需要近似的曲线，0.02*peri是精度的最大值，True表示曲线是闭合的 # 准备做透视变换if len(approx) == 4:docCnt = approxbreak # 透视变换——提取答题卡主体docCnt = docCnt.reshape(4, 2)warped = four_point_transform(img_gray, docCnt)cvshow('warped', warped) # 轮廓识别——识别出选项thresh = cv.threshold(warped, 0, 255, cv.THRESH_BINARY_INV | cv.THRESH_OTSU)[1]cvshow('thresh', thresh)thresh_cnts, _ = cv.findContours(thresh, cv.RETR_EXTERNAL, cv.CHAIN_APPROX_SIMPLE)w_copy = warped.copy()cv.drawContours(w_copy, thresh_cnts, -1, (0, 0, 255), 2)cvshow('warped_contours', w_copy) questionCnts = []# 遍历，挑出选项的cntsfor c in thresh_cnts:(x, y, w, h) = cv.boundingRect(c)ar = w / float(h)# 根据实际情况指定标准if w >= 20 and h >= 20 and ar >= 0.9 and ar <= 1.1:questionCnts.append(c) # 检查是否挑出了选项w_copy2 = warped.copy()cv.drawContours(w_copy2, questionCnts, -1, (0, 0, 255), 2)cvshow('questionCnts', w_copy2) # 检测每一行选择的是哪一项，并将结果储存在元组bubble中，记录正确的个数correct# 按照从上到下t2b对轮廓进行排序questionCnts = sort_contours(questionCnts, method="t2b")[0]correct = 0# 每行有5个选项for (i, q) in enumerate(np.arange(0, len(questionCnts), 5)):# 排序cnts = sort_contours(questionCnts[q:q+5])[0] bubble = None# 得到每一个选项的mask并填充，与正确答案进行按位与操作获得重合点数for (j, c) in enumerate(cnts):mask = np.zeros(thresh.shape, dtype='uint8')cv.drawContours(mask, [c], -1, 255, -1)cvshow('mask', mask) # 通过按位与操作得到thresh与mask重合部分的像素数量bitand = cv.bitwise_and(thresh, thresh, mask=mask)totalPixel = cv.countNonZero(bitand) if bubble is None or bubble[0] < totalPixel:bubble = (totalPixel, j) k = bubble[1]color = (0, 0, 255)if k == right_key[i]:correct += 1color = (0, 255, 0) # 绘图cv.drawContours(warped, [cnts[right_key[i]]], -1, color, 3)cvshow('final', warped) # 计算最终得分并在图中标注score = (correct / 5.0) * 100print(f"Score: {score}%")cv.putText(warped, f"Score: {score}%", (10, 30), cv.FONT_HERSHEY_SIMPLEX, 0.9, (0, 0, 255), 2)cv.imshow("Original", img)cv.imshow("Exam", warped)cv.waitKey(0)

【使用|使用 OpenCV-Python 识别答题卡判卷功能】到此这篇关于使用 OpenCV-Python 识别答题卡判卷的文章就介绍到这了,更多相关OpenCV Python 识别答题卡判卷内容请搜索脚本之家以前的文章或继续浏览下面的相关文章希望大家以后多多支持脚本之家！