np.dot(),np.outer(),np.multiply(),*

本文详细探讨了一维数组、二维数组及矩阵的各种乘法运算,包括np.dot(), np.multiply(), np.outer()等函数的应用场景及其区别,并通过具体示例进行说明。

这一个博客主要是整理所有的关于矩阵还是数组的乘法的细节。

代码:

你需要在下面的代码里面主要到这些点:

(1):np.dot()如果碰到的是秩为1的数组,那么执行的是对应位置的元素相乘再相加;如果遇到的是秩不为1的数组,那么执行的是矩阵相乘。但是需要注意的是矩阵与矩阵相乘是秩为2,矩阵和向量相乘秩为1。

(2):np.multiply()表示的是数组和矩阵对应位置相乘,输出和输出的结果shape一致。

(3):np.outer()表示的是两个向量相乘,拿第一个向量的元素分别与第二个向量所有元素相乘得到结果的一行。

(4):*对数组执行的是对应位置相乘,对矩阵执行的是矩阵相乘。

import numpy as np

print("下面我们将讨论一些关于一维数组的乘法的问题")
A=np.array([1,2,3])
B=np.array([2,3,4])
c=[1,2,3]
print("*:",A*B)#对数组执行的是对应位置元素相乘
print("np.dot():",np.dot(A,B))#当dot遇到佚为1,执行按位乘并相加
print("np.multiply():",np.multiply(A,B))#对数组执行的是对应位置的元素相乘
print("np.outer():",np.outer(A,B))#A的一个元素和B的元素相乘的到结果的一行

print("下面我们将讨论一些关于二维数组和二位数组的乘法的问题")
a=np.array([[1,2,3],[3,4,5]])
b=np.array([[1,1],[2,2],[3,3]])
c=np.array([[2,2,2],[3,3,3]])
#出错:维度不对应:print("*:",a*b)
print("*:",a*c)#*对数组执行的是对应位置相乘
print("np.dot():",np.dot(a,b))#当dot遇到佚不为1执行矩阵的乘法(2,3)×(3,2)=(2,2)
#出错,维度不对应:print("np.multiply():",np.multiply(a,b))
print("np.multiply():",np.multiply(a,c))#数组或者矩阵对应位置元素相乘,返回的是与原数组或者矩阵的大小一致

print("下面我们将讨论一些关于矩阵的乘法的问题")
A=np.mat([[1,2,3],[3,4,5]])
B=np.mat([[1,1],[2,2],[3,3]])
C=np.mat([[2,2,2],[3,3,3]])
D=[1,2,3]
print("*:",A*B)#*对矩阵执行的是矩阵相乘
print("np.dot():",np.dot(A,B))#dot对矩阵执行的是矩阵相乘
print("np.dot():",np.dot(A,D))
#这里可以看出矩阵和矩阵的相相乘是轶为2的,所以是执行的矩阵乘法,但是矩阵和向量相乘是轶为1的,执行的是对应相乘加和
print("np.multiply():",np.multiply(A,C))#multiply执行的是矩阵对应元素相乘

结果:

下面我们将讨论一些关于一维数组的乘法的问题
*: [ 2  6 12]
np.dot(): 20
np.multiply(): [ 2  6 12]
np.outer(): [[ 2  3  4]
 [ 4  6  8]
 [ 6  9 12]]
下面我们将讨论一些关于二维数组和二位数组的乘法的问题
*: [[ 2  4  6]
 [ 9 12 15]]
np.dot(): [[14 14]
 [26 26]]
np.multiply(): [[ 2  4  6]
 [ 9 12 15]]
下面我们将讨论一些关于矩阵的乘法的问题
*: [[14 14]
 [26 26]]
np.dot(): [[14 14]
 [26 26]]
np.dot(): [[14 26]]
np.multiply(): [[ 2  4  6]
 [ 9 12 15]]

 

对于该段代码:# Import necessary libraries for GUI and image processing import tkinter as tk from tkinter import filedialog, ttk, messagebox import numpy as np from PIL import Image, ImageTk, ImageDraw import math import time # Class to display algorithm information in a separate window class AlgorithmInfoWindow: def __init__(self, parent): # Create a new window with fade-in effect self.window = tk.Toplevel(parent) self.window.title("Canny Edge Detection - Algorithm Information") self.window.geometry("800x600") self.window.configure(bg='#f0f0f0') # Make the window float on top of the parent window self.window.transient(parent) self.window.grab_set() # Set initial transparency for fade-in effect self.window.attributes('-alpha', 0.0) # Configure styles for labels and frames style = ttk.Style() style.configure('Info.TLabel', font=('Helvetica', 11), background='#f0f0f0', wraplength=700) style.configure('InfoTitle.TLabel', font=('Helvetica', 14, 'bold'), background='#f0f0f0', foreground='#2c3e50') style.configure('InfoSection.TFrame', background='#ffffff', relief='solid') style.configure('Hover.TFrame', background='#e8f0fe') # Create main frame with custom canvas for smooth scrolling main_frame = ttk.Frame(self.window, style='InfoSection.TFrame') main_frame.pack(fill=tk.BOTH, expand=True, padx=20, pady=20) # Create canvas with custom scrolling self.canvas = tk.Canvas(main_frame, bg='#ffffff', highlightthickness=0, relief='flat') scrollbar = ttk.Scrollbar(main_frame, orient="vertical", command=self.smooth_scroll) self.content_frame = ttk.Frame(self.canvas, style='InfoSection.TFrame') # Configure scrolling self.canvas.configure(yscrollcommand=scrollbar.set) # Pack scrollbar and canvas scrollbar.pack(side="right", fill="y") self.canvas.pack(side="left", fill="both", expand=True) # Create window in canvas self.canvas_frame = self.canvas.create_window( (0, 0), window=self.content_frame, anchor="nw", width=self.canvas.winfo_reqwidth() ) # Add sections with animation delays self.sections = [ ("Canny Edge Detection Algorithm", """The Canny edge detection algorithm is a multi-stage algorithm developed by John F. Canny in 1986. It is considered one of the most robust edge detection algorithms."""), ("1. Grayscale Conversion", """Convert the image to grayscale using weighted sum: gray = 0.2989 * R + 0.5870 * G + 0.1140 * B This weights are based on human perception of color."""), ("2. Gaussian Blur", """Apply Gaussian blur to reduce noise: - Create 5x5 Gaussian kernel using the formula: G(x,y) = (1/2πσ²)e^(-(x²+y²)/2σ²) - Convolve image with kernel - Reduces noise while preserving edges"""), ("3. Gradient Calculation", """Calculate intensity gradients: - Apply Sobel operators in x and y directions - Find gradient magnitude: √(Gx² + Gy²) - Find gradient direction: θ = arctan(Gy/Gx) Sobel operators: X = [[-1, 0, 1], [-2, 0, 2], [-1, 0, 1]] Y = [[-1, -2, -1], [ 0, 0, 0], [ 1, 2, 1]]"""), ("4. Non-Maximum Suppression", """Thin edges by suppressing non-maximum values: 1. Round gradient direction to nearest 45° 2. Compare with pixels in gradient direction 3. Suppress if not local maximum This creates thin, precise edges."""), ("5. Double Thresholding", """Identify strong and weak edges: - High threshold (strong): typically 0.15 * max - Low threshold (weak): typically 0.05 * max Creates three categories: - Strong edges (keep) - Weak edges (evaluate) - Non-edges (discard)"""), ("6. Edge Tracking by Hysteresis", """Connect edges using hysteresis: 1. Start with strong edges 2. Recursively add connected weak edges 3. Remove isolated weak edges This creates continuous edge lines.""") ] # Add sections with animation self.section_frames = [] for i, (title, content) in enumerate(self.sections): self.window.after(i * 100, lambda t=title, c=content: self.add_section_with_animation(t, c)) # Configure canvas scrolling self.content_frame.bind('<Configure>', self.on_frame_configure) self.canvas.bind('<Configure>', self.on_canvas_configure) # Bind mouse wheel for smooth scrolling self.canvas.bind_all('<MouseWheel>', self.on_mousewheel) # Close button with hover effect self.close_btn = ttk.Button(self.window, text="Close", command=self.close_with_animation, style='Custom.TButton') self.close_btn.pack(pady=10) # Start fade-in animation self.fade_in() def fade_in(self, alpha=0.0): """Animate window fade in""" if alpha < 1.0: alpha += 0.1 self.window.attributes('-alpha', alpha) self.window.after(20, lambda: self.fade_in(alpha)) def fade_out(self, alpha=1.0): """Animate window fade out""" if alpha > 0: alpha -= 0.1 self.window.attributes('-alpha', alpha) self.window.after(20, lambda: self.fade_out(alpha)) else: self.window.destroy() def close_with_animation(self): """Close window with fade-out animation""" self.fade_out() def add_section_with_animation(self, title, content): """Add a section with slide-in animation""" frame = ttk.Frame(self.content_frame, style='InfoSection.TFrame') frame.pack(fill=tk.X, pady=10, padx=10) frame.pack_propagate(False) # Prevent size changes # Create content title_label = ttk.Label(frame, text=title, style='InfoTitle.TLabel') title_label.pack(anchor='w', pady=(5, 0)) content_label = ttk.Label(frame, text=content, style='Info.TLabel') content_label.pack(anchor='w', pady=(5, 10)) # Add hover effect frame.bind('<Enter>', lambda e: self.on_section_hover(frame, True)) frame.bind('<Leave>', lambda e: self.on_section_hover(frame, False)) # Animate frame height frame.update() required_height = title_label.winfo_reqheight() + content_label.winfo_reqheight() + 20 frame.configure(height=1) self.animate_frame_height(frame, required_height) self.section_frames.append(frame) def animate_frame_height(self, frame, target_height, current_height=1): """Animate frame height smoothly""" if current_height < target_height: current_height += (target_height - current_height) * 0.2 if current_height < target_height - 1: frame.configure(height=int(current_height)) self.window.after(10, lambda: self.animate_frame_height(frame, target_height, current_height)) else: frame.configure(height=target_height) frame.pack_propagate(True) def on_section_hover(self, frame, entering): """Handle section hover effect""" frame.configure(style='Hover.TFrame' if entering else 'InfoSection.TFrame') def smooth_scroll(self, *args): """Implement smooth scrolling""" if len(args) > 1: self.canvas.yview_moveto(args[1]) else: # Use smoother scrolling with acceleration amount = int(args[0]) # Apply scrolling with acceleration effect if amount != 0: for i in range(3): factor = 0.7 ** i # Decreasing factor for deceleration scroll_amount = int(amount * factor) if amount * factor >= 1 or amount * factor <= -1 else amount self.window.after(i * 5, lambda a=scroll_amount: self.canvas.yview_scroll(a, 'units')) def on_mousewheel(self, event): """Handle smooth mousewheel scrolling with improved animation""" # Get the delta value and normalize it delta = -1 * (event.delta // 120) # Use more steps with smaller increments for smoother animation steps = 15 # Increased steps for smoother animation # Apply scrolling with cubic deceleration curve for i in range(steps): factor = 1 - (i / steps) ** 3 # Cubic deceleration for smoother stop scroll_amount = int(delta * 2 * factor) # Multiply by 2 for better initial momentum scroll_amount = max(1, scroll_amount) if scroll_amount > 0 else min(-1, scroll_amount) # Apply with increasing delay for natural deceleration self.window.after(i * 5, lambda a=scroll_amount: self.canvas.yview_scroll(a, 'units')) def on_frame_configure(self, event=None): """Reset scroll region when content frame size changes""" self.canvas.configure(scrollregion=self.canvas.bbox("all")) def on_canvas_configure(self, event): """Update canvas window size when canvas is resized""" self.canvas.itemconfig(self.canvas_frame, width=event.width) # Main class for the Canny Edge Detection tool class CannyEdgeDetector: def __init__(self): # Initialize the main window self.window = tk.Tk() self.window.title("Canny Edge Detection Tool") self.window.geometry("1200x800") self.window.configure(bg='#f0f0f0') # Set theme for the application style = ttk.Style() style.theme_use('clam') # Configure styles for frames and buttons style.configure('Custom.TFrame', background='#f0f0f0') style.configure('Custom.TButton', padding=10, font=('Helvetica', 10, 'bold')) style.configure('Title.TLabel', font=('Helvetica', 28, 'bold'), # Increased font size background='#f0f0f0', foreground='#2c3e50') style.configure('Subtitle.TLabel', font=('Helvetica', 12), background='#f0f0f0', foreground='#34495e') style.configure('Progress.Horizontal.TProgressbar', background='#2ecc71', troughcolor='#ecf0f1', bordercolor='#bdc3c7') # Create main canvas for scrolling self.main_canvas = tk.Canvas(self.window, bg='#f0f0f0', highlightthickness=0) self.scrollbar = ttk.Scrollbar(self.window, orient="vertical", command=self.smooth_scroll) self.main_canvas.configure(yscrollcommand=self.scrollbar.set) # Pack scrollbar and canvas self.scrollbar.pack(side="right", fill="y") self.main_canvas.pack(side="left", fill="both", expand=True) # Create main frame inside canvas self.main_frame = ttk.Frame(self.main_canvas, padding="20", style='Custom.TFrame') self.canvas_frame = self.main_canvas.create_window( (0, 0), window=self.main_frame, anchor="nw", width=self.main_canvas.winfo_reqwidth() ) # Title and buttons frame title_frame = ttk.Frame(self.main_frame, style='Custom.TFrame') title_frame.grid(row=0, column=0, columnspan=2, sticky='ew') title_frame.grid_columnconfigure(0, weight=1) # Make middle column expandable # Title (centered) self.title_label = ttk.Label(title_frame, text="Canny Edge Detection", style='Title.TLabel', anchor='center') self.title_label.grid(row=0, column=0, pady=(0, 20), sticky='ew') # Centered title # Button frame self.button_frame = ttk.Frame(self.main_frame, style='Custom.TFrame') self.button_frame.grid(row=1, column=0, columnspan=2, pady=(0, 20)) # Buttons (all in one line) self.choose_btn = ttk.Button(self.button_frame, text="Choose Image", command=self.load_image, style='Custom.TButton') self.choose_btn.grid(row=0, column=0, padx=10) self.process_btn = ttk.Button(self.button_frame, text="Process", command=self.process_image_with_progress, style='Custom.TButton') self.process_btn.grid(row=0, column=1, padx=10) self.reset_btn = ttk.Button(self.button_frame, text="Reset", command=self.reset_images, style='Custom.TButton') self.reset_btn.grid(row=0, column=2, padx=10) # Info button (next to reset button) self.info_btn = ttk.Button(self.button_frame, text="ℹ️ Algorithm Info", command=self.show_algorithm_info, style='Custom.TButton') self.info_btn.grid(row=0, column=3, padx=10) # Progress frame self.progress_frame = ttk.Frame(self.main_frame, style='Custom.TFrame') self.progress_frame.grid(row=2, column=0, columnspan=2, pady=(0, 20)) # Progress bar self.progress_var = tk.DoubleVar() self.progress_bar = ttk.Progressbar(self.progress_frame, variable=self.progress_var, maximum=100, mode='determinate', length=400, style='Progress.Horizontal.TProgressbar') self.progress_bar.grid(row=0, column=0, padx=(0, 10)) # Progress percentage label self.progress_label = ttk.Label(self.progress_frame, text="0%", style='Subtitle.TLabel') self.progress_label.grid(row=0, column=1) # Hide progress frame initially self.progress_frame.grid_remove() # Status label self.status_label = ttk.Label(self.main_frame, text="", style='Subtitle.TLabel') self.status_label.grid(row=3, column=0, columnspan=2, pady=(0, 20)) # Image frames self.create_image_frame("Original Image", 4, 0) self.create_image_frame("Edge Detected Image", 4, 1) # Initialize variables self.current_image = None self.processed_image = None # Load default square image self.create_default_square_image() # Configure grid weights for responsive layout self.window.grid_rowconfigure(0, weight=1) self.window.grid_columnconfigure(0, weight=1) self.main_frame.grid_rowconfigure(4, weight=1) # Make image frames expandable self.main_frame.grid_columnconfigure(0, weight=1) self.main_frame.grid_columnconfigure(1, weight=1) # Bind events for smooth scrolling with improved responsiveness self.main_canvas.bind('<Configure>', self.on_canvas_configure) self.main_frame.bind('<Configure>', self.on_frame_configure) self.main_canvas.bind_all('<MouseWheel>', self.on_mousewheel) # Ensure buttons stay visible during processing self.button_frame.lift() self.progress_frame.lift() def create_default_square_image(self): """Create default square image as shown in screenshot""" # Create larger image to better fill the frame size = 400 # Increased size img = Image.new('RGB', (size, size), 'black') draw = ImageDraw.Draw(img) # Calculate sizes for squares outer_size = int(size * 0.6) # 60% of total size inner_size = int(outer_size * 0.3) # 30% of outer square # Calculate positions outer_offset = (size - outer_size) // 2 outer_box = [(outer_offset, outer_offset), (outer_offset + outer_size, outer_offset + outer_size)] # Draw outer white square draw.rectangle(outer_box, fill='white') # Calculate inner square position inner_offset = (size - inner_size) // 2 inner_box = [(inner_offset, inner_offset), (inner_offset + inner_size, inner_offset + inner_size)] # Draw inner black square draw.rectangle(inner_box, fill='black') self.current_image = np.array(img) self.display_image(img, self.original_image_label) def show_algorithm_info(self): """Show algorithm information window""" AlgorithmInfoWindow(self.window) def reset_images(self): """Reset to default square image""" self.create_default_square_image() if self.processed_image_label: self.processed_image_label.configure(image='') self.hide_progress() self.status_label.config(text="Reset complete") def update_progress(self, value, status_text): """Update progress bar, percentage and status text""" self.progress_var.set(value) self.progress_label.config(text=f"{int(value)}%") self.status_label.config(text=status_text) self.window.update() def hide_progress(self): """Hide progress elements""" self.progress_frame.grid_remove() self.status_label.config(text="") self.progress_var.set(0) self.progress_label.config(text="0%") def create_image_frame(self, title, row, column): """Create a frame for displaying images with a title and border""" # Create main frame with fixed size and gray background frame = ttk.Frame(self.main_frame, padding="10", relief="solid", borderwidth=1) frame.grid(row=row, column=column, padx=5, pady=(0, 10), sticky="nsew") # Force both frames to have identical width self.main_frame.grid_columnconfigure(0, weight=1, uniform="equal") # Use uniform to ensure equal size self.main_frame.grid_columnconfigure(1, weight=1, uniform="equal") # Use uniform to ensure equal size # Configure frame size with optimal dimensions for small screens frame.grid_propagate(False) # Prevent frame from resizing to content frame.configure(width=350, height=350) # Reduced dimensions for better small screen compatibility # Title with enhanced styling style = ttk.Style() style.configure('FrameTitle.TLabel', font=('Helvetica', 14, 'bold'), foreground='#2c3e50', background='#e0e0e0', padding=5) # Title container frame with background title_container = ttk.Frame(frame, style='Custom.TFrame') title_container.pack(fill="x", pady=(0, 10)) title_container.configure(height=40) title_label = ttk.Label(title_container, text=title, style='FrameTitle.TLabel', anchor='center') title_label.pack(fill="x", expand=True) # Create image container frame with gray background image_container = ttk.Frame(frame, style='Custom.TFrame') image_container.pack(expand=True, fill="both", padx=5, pady=5) image_container.configure(width=480, height=430) # Image label with gray background image_label = ttk.Label(image_container) image_label.pack(expand=True, fill="both") # Save button container save_container = ttk.Frame(frame, style='Custom.TFrame') save_container.pack(fill="x", pady=(5, 0)) # Save button with enhanced style style.configure('Save.TButton', font=('Helvetica', 10), padding=5) save_btn = ttk.Button(save_container, text="Save Image", style='Save.TButton', command=lambda: self.save_image(column)) save_btn.pack(pady=5) if column == 0: self.original_image_label = image_label else: self.processed_image_label = image_label def display_image(self, image, label): """ Display numpy array image in GUI label Args: image: numpy array of image data label: target label widget for display Maintains aspect ratio while fitting to display area """ if image: # Use fixed dimensions for both frames to ensure they're identical frame_width = 480 frame_height = 430 # Calculate scaling ratio while preserving aspect ratio img_width, img_height = image.size width_ratio = frame_width / img_width height_ratio = frame_height / img_height scale_ratio = min(width_ratio, height_ratio) # Calculate new size new_width = int(img_width * scale_ratio) new_height = int(img_height * scale_ratio) # Resize image resized_image = image.resize((new_width, new_height), Image.Resampling.LANCZOS) # Create new image with gray background final_image = Image.new('RGB', (frame_width, frame_height), '#f0f0f0') # Calculate position to center the image x_offset = (frame_width - new_width) // 2 y_offset = (frame_height - new_height) // 2 # Paste resized image onto background final_image.paste(resized_image, (x_offset, y_offset)) # Convert to PhotoImage and display photo = ImageTk.PhotoImage(final_image) label.configure(image=photo) label.image = photo # Keep reference def save_image(self, image_type): """Save the image to disk""" if image_type == 0 and self.current_image is not None: image_to_save = Image.fromarray(self.current_image) title = "Save Original Image" elif image_type == 1 and hasattr(self, 'processed_image'): image_to_save = Image.fromarray((self.processed_image * 255).astype(np.uint8)) title = "Save Edge Detected Image" else: messagebox.showwarning("Warning", "No image to save!") return # Ask for save location file_path = filedialog.asksaveasfilename( title=title, defaultextension=".png", filetypes=[ ("PNG files", "*.png"), ("JPEG files", "*.jpg"), ("All files", "*.*") ] ) if file_path: try: image_to_save.save(file_path) messagebox.showinfo("Success", "Image saved successfully!") except Exception as e: messagebox.showerror("Error", f"Failed to save image: {str(e)}") def process_image_with_progress(self): """ Execute edge detection pipeline with progress tracking Steps: 1. Input validation 2. Grayscale conversion 3. Gaussian blur 4. Gradient calculation 5. Non-max suppression 6. Double thresholding 7. Hysteresis edge tracking """ if self.current_image is None: messagebox.showwarning("Warning", "Please select an image first!") return self.progress_frame.grid() # Show progress frame self.status_label.grid() # Show status label try: # Convert to grayscale self.update_progress(20, "Converting to grayscale...") time.sleep(0.3) # Simulate processing time gray_image = self.to_grayscale(self.current_image) # Apply Gaussian blur self.update_progress(40, "Applying Gaussian blur...") time.sleep(0.3) blurred = self.apply_gaussian_blur(gray_image) # Calculate gradients self.update_progress(60, "Calculating gradients...") time.sleep(0.3) gradient_magnitude, gradient_direction = self.sobel_filters(blurred) # Apply non-maximum suppression self.update_progress(70, "Applying non-maximum suppression...") time.sleep(0.3) suppressed = self.non_maximum_suppression(gradient_magnitude, gradient_direction) # Apply double threshold self.update_progress(80, "Applying double threshold...") time.sleep(0.3) strong_edges, weak_edges = self.double_threshold(suppressed) # Apply hysteresis self.update_progress(90, "Applying hysteresis...") time.sleep(0.3) final_edges = self.hysteresis(strong_edges, weak_edges) # Store processed image self.processed_image = final_edges # Display result self.update_progress(100, "Complete!") display_image = Image.fromarray((final_edges * 255).astype(np.uint8)) self.display_image(display_image, self.processed_image_label) # Hide progress elements after a delay self.window.after(1000, self.hide_progress) except Exception as e: messagebox.showerror("Error", f"An error occurred: {str(e)}") self.hide_progress() def load_image(self): """ Load an image through file dialog Supported formats: JPEG, PNG, BMP Updates original image display Handles common file errors """ file_path = filedialog.askopenfilename( filetypes=[ ("Image files", "*.jpg;*.jpeg;*.png;*.bmp;*.gif"), ("All files", "*.*") ] ) if file_path: try: # Load image image = Image.open(file_path) # Convert to RGB if necessary if image.mode != 'RGB': image = image.convert('RGB') # Store original image for processing self.current_image = np.array(image) # Display image with proper scaling self.display_image(image, self.original_image_label) # Clear processed image if self.processed_image_label: self.processed_image_label.configure(image='') self.status_label.config(text="Image loaded successfully") except Exception as e: messagebox.showerror("Error", f"Failed to load image: {str(e)}") self.status_label.config(text="Failed to load image") def to_grayscale(self, image): """Convert RGB image to grayscale using manual implementation""" if len(image.shape) == 3: return np.dot(image[..., :3], [0.2989, 0.5870, 0.1140]).astype(np.float32) return image def gaussian_kernel(self, size, sigma=1.4): """Generate Gaussian kernel manually""" kernel = np.zeros((size, size)) center = size // 2 for x in range(size): for y in range(size): x_dist = x - center y_dist = y - center kernel[x, y] = (1 / (2 * np.pi * sigma ** 2)) * np.exp(-(x_dist ** 2 + y_dist ** 2) / (2 * sigma ** 2)) return kernel / np.sum(kernel) def apply_gaussian_blur(self, image, kernel_size=5): """Apply Gaussian blur manually""" kernel = self.gaussian_kernel(kernel_size) padding = kernel_size // 2 padded = np.pad(image, padding, mode='edge') output = np.zeros_like(image) for i in range(image.shape[0]): for j in range(image.shape[1]): output[i, j] = np.sum( padded[i:i + kernel_size, j:j + kernel_size] * kernel ) return output def sobel_filters(self, image): """Apply Sobel filters manually""" Gx = np.array([[-1, 0, 1], [-2, 0, 2], [-1, 0, 1]]) Gy = np.array([[-1, -2, -1], [0, 0, 0], [1, 2, 1]]) padding = 1 padded = np.pad(image, padding, mode='edge') gradient_x = np.zeros_like(image) gradient_y = np.zeros_like(image) for i in range(image.shape[0]): for j in range(image.shape[1]): gradient_x[i, j] = np.sum( padded[i:i + 3, j:j + 3] * Gx ) gradient_y[i, j] = np.sum( padded[i:i + 3, j:j + 3] * Gy ) gradient_magnitude = np.sqrt(gradient_x ** 2 + gradient_y ** 2) gradient_direction = np.arctan2(gradient_y, gradient_x) return gradient_magnitude, gradient_direction def non_maximum_suppression(self, gradient_magnitude, gradient_direction): """Apply non-maximum suppression""" height, width = gradient_magnitude.shape output = np.zeros_like(gradient_magnitude) # Convert angles from radians to degrees angle = gradient_direction * 180 / np.pi angle[angle < 0] += 180 for i in range(1, height - 1): for j in range(1, width - 1): q = 255 r = 255 # Angle 0 if (0 <= angle[i, j] < 22.5) or (157.5 <= angle[i, j] <= 180): q = gradient_magnitude[i, j + 1] r = gradient_magnitude[i, j - 1] # Angle 45 elif (22.5 <= angle[i, j] < 67.5): q = gradient_magnitude[i + 1, j - 1] r = gradient_magnitude[i - 1, j + 1] # Angle 90 elif (67.5 <= angle[i, j] < 112.5): q = gradient_magnitude[i + 1, j] r = gradient_magnitude[i - 1, j] # Angle 135 elif (112.5 <= angle[i, j] < 157.5): q = gradient_magnitude[i - 1, j - 1] r = gradient_magnitude[i + 1, j + 1] if (gradient_magnitude[i, j] >= q) and (gradient_magnitude[i, j] >= r): output[i, j] = gradient_magnitude[i, j] else: output[i, j] = 0 return output def double_threshold(self, image, low_ratio=0.05, high_ratio=0.15): """Apply double threshold""" high_threshold = image.max() * high_ratio low_threshold = high_threshold * low_ratio strong_edges = (image >= high_threshold) weak_edges = (image >= low_threshold) & (image < high_threshold) return strong_edges, weak_edges def hysteresis(self, strong_edges, weak_edges): """Apply hysteresis to connect edges""" height, width = strong_edges.shape output = np.copy(strong_edges) dx = [-1, -1, -1, 0, 0, 1, 1, 1] dy = [-1, 0, 1, -1, 1, -1, 0, 1] # Iterate until no more changes while True: previous = np.copy(output) for i in range(1, height - 1): for j in range(1, width - 1): if weak_edges[i, j]: # Check if any neighbor is a strong edge for k in range(8): if output[i + dx[k], j + dy[k]]: output[i, j] = True break if np.array_equal(previous, output): break return output def on_mousewheel(self, event): """Handle smooth mousewheel scrolling with improved animation""" # Get the delta value and normalize it delta = -1 * (event.delta // 120) # Use more steps with smaller increments for smoother animation steps = 15 # Increased number of steps for even smoother scrolling # Apply scrolling with acceleration and deceleration for i in range(steps): # Calculate a smooth deceleration curve factor = 1 - (i / steps) ** 2 # Quadratic deceleration for natural feel scroll_amount = max(1, int(delta * factor)) if delta > 0 else min(-1, int(delta * factor)) # Apply with increasing delay for natural deceleration self.window.after(i * 4, lambda a=scroll_amount: self.main_canvas.yview_scroll(a, 'units')) def smooth_scroll(self, *args): """Implement smooth scrolling""" if len(args) > 1: self.main_canvas.yview_moveto(args[1]) else: # Use smoother scrolling with acceleration amount = int(args[0]) # Apply scrolling with acceleration effect if amount != 0: for i in range(5): # Increased range for smoother scrolling factor = 0.8 ** i # Adjusted factor for better deceleration scroll_amount = int(amount * factor) if amount * factor >= 1 or amount * factor <= -1 else amount self.window.after(i * 4, lambda a=scroll_amount: self.main_canvas.yview_scroll(a, 'units')) def on_frame_configure(self, event=None): """Reset scroll region when content frame size changes""" self.main_canvas.configure(scrollregion=self.main_canvas.bbox("all")) # Ensure the canvas is large enough to accommodate all content self.main_frame.update_idletasks() def on_canvas_configure(self, event): """Update canvas window size when canvas is resized""" self.main_canvas.itemconfig(self.canvas_frame, width=event.width) # Ensure the canvas window is properly sized for future content additions self.main_canvas.configure(width=event.width) def run(self): self.window.mainloop() if __name__ == "__main__": app = CannyEdgeDetector() app.run()如何在训练阶段,使用农田边缘区域的掩码标注数据,限定模型仅在边缘区域进行目标检测,忽略农田内部区域。
07-31
评论 3
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值