![\"Template](\"https:\/\/data-flair.training\/blogs\/wp-content\/uploads\/sites\/2\/2023\/11\/Template-Matching-formula.webp\")
<\/a><\/p>\nIn the following sections, we’ll delve into the mechanisms behind template matching, explore the suite of tools OpenCV offers, and discover how to tackle challenges like scale variation and noise.<\/p>\n
By the end of this journey, you’ll be armed with the knowledge to unlock the potential of template matching and add it to your arsenal of computer vision techniques. So, let’s embark on this visual adventure and uncover the secrets of template matching together.<\/p>\n
What is Template Matching in OpenCV?<\/h3>\n
Template matching in OpenCV is a technique used for finding a specific template image within a larger target image. This method is widely employed in computer vision to identify patterns, objects, or specific regions of interest within an image. OpenCV provides a set of functions and methods that allow you to perform template matching efficiently and accurately.<\/p>\n
The basic idea behind template matching in OpenCV involves sliding the template image over the target image and calculating a similarity score at each position. This similarity score indicates how closely the template matches the content of the target image at that particular position. The position with the highest similarity score corresponds to the location where the template best matches the target image.<\/p>\n
OpenCV’s cv2.matchTemplate() function is the cornerstone of template matching. This function takes the template image and the target image as inputs and returns a result matrix. Each element of this result matrix corresponds to a position in the target image and contains a value that indicates the degree of similarity between the template and the content of the target image at that position.<\/p>\n
Different methods can be used to calculate the similarity score, and OpenCV offers several options such as:<\/strong><\/p>\n 1. Normalized Cross-Correlation (`cv2.TM_CCORR_NORMED`):<\/strong> This method computes the cross-correlation between the template and the target image while normalizing the pixel values. The resulting matrix will have peaks at positions where the template best matches the target image.<\/p>\n 2. Sum of Squared Differences (`cv2.TM_SQDIFF`):<\/strong> This method calculates the sum of squared differences between the template and the target image. The lower the value, the better the match. This method is often used with the minimum value being the best match.<\/p>\n 3. Correlation Coefficient (`cv2.TM_CCOEFF`):<\/strong> This method calculates the correlation coefficient between the template and the target image. A higher value indicates a better match, with 1 being a perfect match.<\/p>\n These methods offer flexibility in choosing the most suitable approach based on the specific application and the nature of the images involved.<\/p>\n Overall, template matching in OpenCV provides a powerful tool for various tasks, including object detection, pattern recognition, and image analysis. However, it’s important to note that template matching might have limitations when dealing with scale variation, rotation, and complex scenes. In such cases, additional techniques or algorithms might be required to enhance the accuracy of the results.<\/p>\n The best match is determined using this mode’s sum of squared differences. The match will be better if the score is lower.<\/p>\n TM_SQDIFF:<\/strong><\/p>\n This mode calculates the cross-correlation between the template and the target image. The higher the score, the better the match.<\/p>\n Use the same Libraries and path from above code:<\/strong><\/p>\n This mode calculates the correlation coefficient between the template and the target image. The higher the score, the better the match.<\/p>\n Use the same Libraries and path from above code:<\/strong><\/p>\n These examples demonstrate how to perform template matching using different match modes in OpenCV.<\/p>\n Certainly, performing template matching with multiple objects involves iterating over regions of the target image and applying the template matching process to each region.<\/p>\n Here’s an example code that demonstrates how to perform template matching with multiple objects using OpenCV:<\/strong><\/p>\n Original Image:<\/strong><\/p>\n Template Image:\u00a0<\/strong><\/p>\n Matched Image:<\/strong> Here, multiple similar shaped Spots are marked up,<\/p>\n In this code, the cv2.matchTemplate function is used with cv2.TM_CCOEFF_NORMED to perform template matching. The code then finds all locations where the result exceeds a predefined threshold (0.7 in this case) to consider them as matches. It iterates over these matched locations and draws rectangles around each match on the copy of the original image for visualization.<\/p>\n Applying template matching with multiple scales involves resizing the template and target image at different scales and performing template matching on each scale.<\/p>\n Here’s an example code that demonstrates how to perform multi-scale template matching using OpenCV:<\/strong><\/p>\n Template Image:<\/strong><\/p>\nUnlocking the Art of Visual Recognition: Exploring Template Matching with Varied Match Modes<\/h3>\n
1. cv2.TM_SQDIFF:<\/h4>\n
import cv2\r\nimport numpy as np\r\n\r\n# Load the original image, template image, and convert to grayscale\r\noriginal_image = cv2.imread(r\"C:\\Users\\satchit\\Desktop\\DataFlair\\OpenCv Template Matching\\traffic-street-img.jpg\",cv2.IMREAD_GRAYSCALE)\r\ntemplate_image = cv2.imread(r\"C:\\Users\\satchit\\Desktop\\DataFlair\\OpenCv Template Matching\\Template_image.png\",cv2.IMREAD_GRAYSCALE)\r\n\r\n# Perform template matching using cv2.TM_SQDIFF\r\nresult = cv2.matchTemplate(original_image, template_image, cv2.TM_SQDIFF)\r\nmin_val, max_val, min_loc, max_loc = cv2.minMaxLoc(result)\r\ntop_left = min_loc\r\nbottom_right = (top_left[0] + template_image.shape[1], top_left[1] + template_image.shape[0])\r\n\r\n# Create a copy of the original image for visualization\r\nmatched_image = original_image.copy()\r\ncv2.rectangle(matched_image, top_left, bottom_right, (0, 255, 0), 2)\r\n# Display the images\r\ncv2.imshow('Original Image', original_image)\r\ncv2.imshow('Template Image', template_image)\r\ncv2.imshow('Matched Image (TM_SQDIFF)', matched_image)\r\ncv2.waitKey(0)\r\ncv2.destroyAllWindows()<\/pre>\n<\/a><\/p>\n2. cv2.TM_CCORR:<\/h4>\n
# Perform template matching using cv2.TM_CCORR\r\nresult = cv2.matchTemplate(original_image, template_image, cv2.TM_CCORR)\r\nmin_val, max_val, min_loc, max_loc = cv2.minMaxLoc(result)\r\ntop_left = max_loc\r\nbottom_right = (top_left[0] + template_image.shape[1], top_left[1] + template_image.shape[0])\r\n\r\n# Create a copy of the original image for visualization\r\nma\r\ntched_image = original_image.copy()\r\ncv2.rectangle(matched_image, top_left, bottom_right, (0, 255, 0), 2)\r\n\r\n# Display the images\r\ncv2.imshow('Original Image', original_image)\r\ncv2.imshow('Template Image', template_image)\r\ncv2.imshow('Matched Image (TM_CCORR)', matched_image)\r\ncv2.waitKey(0)\r\ncv2.destroyAllWindows()<\/pre>\n3. cv2.TM_CCOEFF:<\/h4>\n
# Perform template matching using cv2.TM_CCOEFF\r\nresult = cv2.matchTemplate(original_image, template_image, cv2.TM_CCOEFF)\r\nmin_val, max_val, min_loc, max_loc = cv2.minMaxLoc(result)\r\ntop_left = max_loc\r\nbottom_right = (top_left[0] + template_image.shape[1], top_left[1] + template_image.shape[0])\r\n\r\n# Create a copy of the original image for visualization\r\nmatched_image = original_image.copy()\r\ncv2.rectangle(matched_image, top_left, bottom_right, (0, 255, 0), 2)\r\n\r\n# Display the images\r\ncv2.imshow('Original Image', original_image)\r\ncv2.imshow('Template Image', template_image)\r\ncv2.imshow('Matched Image (TM_CCOEFF)', matched_image)\r\ncv2.waitKey(0)\r\ncv2.destroyAllWindows()<\/pre>\nMulti-Object Discovery: Unveiling the Power of OpenCV Template Matching with Multiple Targets<\/h3>\n
import cv2\r\nimport numpy as np\r\n\r\n# Load the original image, template image, and convert to grayscale\r\noriginal_image = cv2.imread(r\"C:\/Users\/sanji\/Desktop\/DataFlair\/OpenCv Template Matching\/wildlife-leopard.jpg\",cv2.IMREAD_GRAYSCALE)\r\ntemplate_image = cv2.imread(r\"C:\\Users\\sanji\\Desktop\\DataFlair\\OpenCv Template Matching\\leopard templeate.png\",cv2.IMREAD_GRAYSCALE)\r\n\r\n# Get the dimensions of the template\r\ntemplate_height, template_width = template_image.shape\r\n\r\n# Perform template matching using TM_CCOEFF_NORMED\r\nresult = cv2.matchTemplate(original_image, template_image, cv2.TM_CCOEFF_NORMED)\r\n\r\n# Define a threshold for considering matches\r\nthreshold = 0.7\r\n\r\n# Find all locations where the result exceeds the threshold\r\nlocations = np.where(result >= threshold)\r\n\r\n# Create a copy of the original image for visualization\r\nmatched_image = original_image.copy()\r\n\r\n# Iterate over matched locations\r\nfor loc in zip(*locations[::-1]):\r\n top_left = loc\r\n bottom_right = (top_left[0] + template_width, top_left[1] + template_height)\r\n cv2.rectangle(matched_image, top_left, bottom_right, (0, 255, 0), 2)\r\n\r\n# Display the images\r\ncv2.imshow('Original Image', original_image)\r\ncv2.imshow('Template Image', template_image)\r\ncv2.imshow('Matched Image', matched_image)\r\ncv2.waitKey(0)\r\ncv2.destroyAllWindows()<\/pre>\n<\/a><\/p>\n<\/a><\/p>\n<\/a><\/p>\nScaling New Heights: Unveiling the Magic of Multi-Scale Template Matching with OpenCV<\/h3>\n
import cv2\r\nimport numpy as np\r\n\r\n# Load the original image, template image, and convert to grayscale\r\noriginal_image = cv2.imread(r\"C:\\Users\\satchit\\Desktop\\DataFlair\\OpenCv Template Matching\\walking-dog.jpg\",cv2.IMREAD_GRAYSCALE)\r\ntemplate_image = cv2.imread(r\"C:\\Users\\satchit\\Desktop\\DataFlair\\OpenCv Template Matching\\walking_dog_template.png\",cv2.IMREAD_GRAYSCALE)\r\n\r\n# Define a list of scales to consider\r\nscales = [0.5, 1.0, 1.5]\r\n\r\n# Create a copy of the original image for visualization\r\nmatched_image = original_image.copy()\r\n\r\n# Iterate over each scale\r\nfor scale in scales:\r\n # Resize the template image based on the current scale\r\n scaled_template = cv2.resize(template_image, None, fx=scale, fy=scale)\r\n\r\n # Perform template matching using TM_CCOEFF_NORMED\r\n result = cv2.matchTemplate(original_image, scaled_template, cv2.TM_CCOEFF_NORMED)\r\n\r\n # Define a threshold for considering matches\r\n threshold = 0.7\r\n\r\n # Find all locations where the result exceeds the threshold\r\n locations = np.where(result >= threshold)\r\n\r\n # Iterate over matched locations\r\n for loc in zip(*locations[::-1]):\r\n top_left = loc\r\n bottom_right = (top_left[0] + scaled_template.shape[1], top_left[1] + scaled_template.shape[0])\r\n cv2.rectangle(matched_image, top_left, bottom_right, (0, 255, 0), 2)\r\n\r\n# Display the images\r\ncv2.imshow('Original Image', original_image)\r\ncv2.imshow('Template Image', template_image)\r\ncv2.imshow('Matched Image (Multi-Scale)', matched_image)\r\ncv2.waitKey(0)\r\ncv2.destroyAllWindows()<\/pre>\n