HED_edgeDetect_边缘检测_hed_深度学习_

import cv2 as cv import numpy as np class CropLayer(object): def __init__(self, params, blobs): self.xstart = 0 self.xend = 0 self.ystart = 0 self.yend = 0 # Our layer receives two inputs. We need to crop the first input blob # to match a shape of the second one (keeping batch size and number of channels) def getMemoryShapes(self, inputs): inputShape, targetShape = inputs[0], inputs[1] batchSize, numChannels = inputShape[0], inputShape[1] height, width = targetShape[2], targetShape[3] self.ystart = (inputShape[2] - targetShape[2]) // 2 self.xstart = (inputShape[3] - targetShape[3]) // 2 self.yend = self.ystart + height self.xend = self.xstart + width return [[batchSize, numChannels, height, width]] def forward(self, inputs): return [inputs[0][:, :, self.ystart:self.yend, self.xstart:self.xend]] def hed_detect(imagePath): # 加入特殊层，并加载模型; cv.dnn_registerLayer('Crop', CropLayer) # Load the model. net = cv.dnn.readNet('hed_caffe_deploy.prototxt', 'hed_pretrained_bsds.caffemodel') # 读取图像，并做一些预处理 frame = cv.imread(imagePath) # newInput = cv.resize(frame, (640, 480)) # cv.imshow('newImg', newInput) newInput = frame # 使用模型对图像推断，计算结果; inp = cv.dnn.blobFromImage(newInput, scalefactor=1.0, size=(newInput.shape[0], newInput.shape[1]), mean=(104.00698793, 116.66876762, 122.67891434), swapRB=False, crop=False) net.setInput(inp) out = net.forward() out = out[0, 0] out = cv.resize(out, (newInput.shape[1], newInput.shape[0])) out = 255 * out out = out.astype(np.uint8) # thre, out = cv.threshold(out, 15, 0xff, cv.THRESH_BINARY) thre, out = cv.threshold(out, 15, 0xff, cv.THRESH_OTSU) # out = cv.cvtColor(out, cv.COLOR_GRAY2BGR) # con = np.concatenate((newInput, out), axis=1) print(thre) # 显示图像 kWinName = 'Holistically-Nested_Edge_Detection' cv.namedWindow(kWinName, cv.WINDOW_AUTOSIZE) cv.imshow(kWinName, out) cv.waitKey(0) if __name__ == '__main__': imgPath = 'D:/Projects/python/Keras_HED-master/images/000001.jpg' hed_detect(imgPath)