Simp!a nhianto$-$

Page $1$ of $2$

An example of an input image:

The output image with the corresponding text: An example of an input image:

The output image with the corresponding text:

Circle, Magenta, $1$

Circle, Blue, $1$

Circle, Cyan, $1$

Circle, Black, $2$

Circle, Yellow, $2$

Triangle, Green, $1$

Triangle, Red $1$

Triangle, Blue, $2$

Triangle, Cyan, $1$

Rectangle, Yellow, $2$

Rectangle, Cyan, $2$

Rectangle, Black, $1$

Rectangle, Red, $1$

Rectangle, Magenta, $1$

Square, Green, $1$Please implement a program $to$ recognize simple objects and basic shapes $inan$ input

image $as$ shown $in$ the figures below. Consider the following regulations:

The input: A colored image.

The output:

The same input image with a label above each object $to$ identify $it$ from the

list below

House and its type $(Class1or$ Class $2),$ Tree and its type $(Class1or$

Class $2),$ Car and Truck

Print the name, the color and the number $of$ basic shapes $in$ the image

$(Circle,$ triangle, rectangle $or$ square$)on$ the command prompt $as$ a list

$(Don\text{'}t$ display them $on$ the image$).$

Each student must complete the assignment individually.

Only $C,C++,or$ Python programming languages can $be$ used, $in$ conjunction

with OpenCV.

Simple objects: import cv$2$

import numpy as np

# Load the image

image$\_$path $=$ r$\text{'}$C:$\backslash$Users$\backslash$Hp$\backslash$OneDrive$\backslash$Desktop$\backslash$input$1-$ Copy.png$\text{'}$ # Replace with the actual path to your image

image $=$ cv$2.$imread$($image$\_$path$)$

# Convert to grayscale

gray $=$ cv$2.$cvtColor$($image$,$ cv$2.$COLOR$\_$BGR$2$GRAY$)$

# Apply binary thresholding

$\_,$ thresh $=$ cv$2.$threshold$($gray$,240,255,$ cv$2.$THRESH$\_$BINARY$\_$INV$)$

# Find contours

contours, $\_=$ cv$2.$findContours$($thresh$,$ cv$2.$RETR$\_$EXTERNAL, cv$2.$CHAIN$\_$APPROX$\_$SIMPLE$)$

# Function to detect shape

def detect$\_$shape$($contour$)$:

shape $=$ "unidentified"

peri $=$ cv$2.$arcLength$($contour$,$ True$)$

approx $=$ cv$2.$approxPolyDP$($contour$,0\mathrm{.}04*$ peri, True$)$

if len$($approx$)==3$:

shape $=$ "triangle"

elif len$($approx$)==4$:

x$,$ y$,$ w$,$ h $=$ cv$2.$boundingRect$($approx$)$

aspect$\_$ratio $=$ w $/$ float$($h$)$

shape $=$ "square" if aspect$\_$ratio else "rectangle"

else:

shape $=$ "circle"

return shape

# Function to detect color

def detect$\_$color$($image$,$ contour$)$:

mask $=$ np$.$zeros$($image$.$shape$[$:$2],$ dtype$=$"uint$8")$

cv$2.$drawContours$($mask$,[$contour$],-1,255,-1)$

mean $=$ cv$2.$mean$($image$,$ mask$=$mask$)[$:$3]$

color $=$ "unknown"

if mean$[2]>150$ and mean$[1]<mtext\; id="EditorElement\_57889"></mtext><mn\; id="EditorElement\_57890">1</mn><mn\; id="EditorElement\_57891">0</mn><mn\; id="EditorElement\_57892">0</mn>$ and mean$[0]<mtext\; id="EditorElement\_57907"></mtext><mn\; id="EditorElement\_57908">1</mn><mn\; id="EditorElement\_57909">0</mn><mn\; id="EditorElement\_57910">0</mn>$:

color $=$ "red"

elif mean$[2]<mtext\; id="EditorElement\_57953"></mtext><mn\; id="EditorElement\_57954">1</mn><mn\; id="EditorElement\_57955">0</mn><mn\; id="EditorElement\_57956">0</mn>$ and mean$[1]>150$ and mean$[0]<mtext\; id="EditorElement\_57989"></mtext><mn\; id="EditorElement\_57990">1</mn><mn\; id="EditorElement\_57991">0</mn><mn\; id="EditorElement\_57992">0</mn>$:

color $=$ "green"

elif mean$[2]<mtext\; id="EditorElement\_58037"></mtext><mn\; id="EditorElement\_58038">1</mn><mn\; id="EditorElement\_58039">0</mn><mn\; id="EditorElement\_58040">0</mn>$ and mean$[1]<mtext\; id="EditorElement\_58055"></mtext><mn\; id="EditorElement\_58056">1</mn><mn\; id="EditorElement\_58057">0</mn><mn\; id="EditorElement\_58058">0</mn>$ and mean$[0]>150$:

color $=$ "blue"

elif mean$[2]>200$ and mean$[1]>200$ and mean$[0]<mtext\; id="EditorElement\_58156"></mtext><mn\; id="EditorElement\_58157">1</mn><mn\; id="EditorElement\_58158">0</mn><mn\; id="EditorElement\_58159">0</mn>$:

color $=$ "yellow"

elif mean$[2]>150$ and mean$[1]<mtext\; id="EditorElement\_58223"></mtext><mn\; id="EditorElement\_58224">1</mn><mn\; id="EditorElement\_58225">0</mn><mn\; id="EditorElement\_58226">0</mn>$ and mean$[0]>150$:

color $=$ "magenta"

elif mean$[2]<mtext\; id="EditorElement\_58291"></mtext><mn\; id="EditorElement\_58292">1</mn><mn\; id="EditorElement\_58293">0</mn><mn\; id="EditorElement\_58294">0</mn>$ and mean$[1]>150$ and mean$[0]>150$:

color $=$ "cyan"

elif mean$[2]>200$ and mean$[1]>200$ and mean$[0]>200$:

color $=$ "white"

elif mean$[2]<mtext\; id="EditorElement\_58458"></mtext><mn\; id="EditorElement\_58459">5</mn><mn\; id="EditorElement\_58460">0</mn>$ and mean$[1]<mtext\; id="EditorElement\_58475"></mtext><mn\; id="EditorElement\_58476">5</mn><mn\; id="EditorElement\_58477">0</mn>$ and mean$[0]<mtext\; id="EditorElement\_58492"></mtext><mn\; id="EditorElement\_58493">5</mn><mn\; id="EditorElement\_58494">0</mn>$:

color $=$ "black"

return color

# Dictionary to store the number of each shape

shapes$\_$count $=\{"$circle$"$: $0,$ "triangle": $0,$ "square": $0,$ "rectangle": $0\}$

# Iterate through contours and classify shapes

detected$\_$shapes $=[]$

for contour in contours:

shape $=$ detect$\_$shape$($contour$)$

color $=$ detect$\_$color$($image$,$ contour$)$

# Increase the shape count

if shape in shapes$\_$count:

shapes$\_$count$[$shape$]+=1$

# Find the center of the contour for labeling

M $=$ cv$2.$moments$($contour$)$

if M$["$m$00"]!=0$:

cX $=$ int$($M$["$m$10"]/$ M$["$m$00"])$

cY $=$ int$($M$["$m$01"]/$ M$["$m$00"])$

else:

cX$,$ cY $=0,0$

detected$\_$shapes.append$(($shape$,$ color, contour, cX$,$ cY$))$

# Helper function to check if contours are close

def are$\_$contours$\_$close$($cnt$1,$ cnt$2,$ max$\_$dist$=50)$:

for pt$1$ in cnt$1$:

for pt$2$ in cnt$2$:

dist $=$ np$.$linalg.norm$($pt$1-$ pt$2)$

if dist $$ max$\_$dist:

return True

return False

def is$\_$equilateral$($contour$)$:

peri $=$ cv$2.$arcLength$($contour$,$ True$)$

approx $=$ cv$2.$approxPolyDP$($contour$,0\mathrm{.}04*$ peri, True$)$

if len$($approx$)==3$:

# Calculate the lengths of each side

side$1=$ np$.$linalg.norm$($approx$[0]-$ approx$[1])$

side$2=$ np$.$linalg.norm$($approx$[1]-$ approx$[2])$

side$3=$ np$.$linalg.norm$($approx$[2]-$ approx$[0])$

# Check if all sides are nearly equal

return abs$($side$1-$ side$2)<mtext\; id="EditorElement\_59933"></mtext><mn\; id="EditorElement\_59934">0</mn><mn\; id="EditorElement\_59935">.</mn><mn\; id="EditorElement\_59936">0</mn><mn\; id="EditorElement\_59937">5</mn><mtext\; id="EditorElement\_59938"></mtext><mi\; id="EditorElement\_59939">*</mi>$ side$1$ and abs$($side$2-$ side$3)<mtext\; id="EditorElement\_59971"></mtext><mn\; id="EditorElement\_59972">0</mn><mn\; id="EditorElement\_59973">.</mn><mn\; id="EditorElement\_59974">0</mn><mn\; id="EditorElement\_59975">5</mn><mtext\; id="EditorElement\_59976"></mtext><mi\; id="EditorElement\_59977">*</mi>$ side$2$ and abs$($side$3-$ side$1)<mtext\; id="EditorElement\_60009"></mtext><mn\; id="EditorElement\_60010">0</mn><mn\; id="EditorElement\_60011">.</mn><mn\; id="EditorElement\_60012">0</mn><mn\; id="EditorElement\_60013">5</mn><mtext\; id="EditorElement\_60014"></mtext><mi\; id="EditorElement\_60015">*</mi>$ side$3$

return False

def is$\_$inside$($contour$1,$ contour$2)$:

x$,$ y$,$ w$,$ h $=$ cv$2.$boundingRect$($contour$2)$

return cv$2.$pointPolygonTest$($contour$1,($x $+$ w $/2,$ y $+$ h $/2),$ False$)>$