如何用ObjC++从vector<cv::Point>点列表中提取四个角点？

I see you're trying to pull out four specific corner points from your given vector<cv::Point> list, and your previous attempts didn't work out. Let's break this down with a clear, tailored solution for Objective-C++.

First, let's recap your requirements:

• From the point list [(5,10), (15,25), (40,30), (10,15)]
• Extract topLeft = (5,10), topRight = (10,15), bottomLeft = (15,25), bottomRight = (40,30)

The key here is to leverage the coordinate characteristics of each corner:

• The two top points have smaller y-values, split them by x (smaller x = topLeft, larger x = topRight)
• The two bottom points have larger y-values, split them by x (smaller x = bottomLeft, larger x = bottomRight)

Here's the complete Objective-C++ implementation:

#import <Foundation/Foundation.h>
#include <opencv2/opencv.hpp>
#include <algorithm>

using namespace cv;

int main(int argc, const char * argv[]) {
    @autoreleasepool {
        // Initialize your point list
        std::vector<Point> points = { Point(5,10), Point(15,25), Point(40,30), Point(10,15) };
        
        // Step 1: Sort points by y-coordinate ascending (smallest y first)
        std::sort(points.begin(), points.end(), [](const Point& a, const Point& b) {
            return a.y < b.y;
        });
        
        // Split into top (first 2 points) and bottom (last 2 points) groups
        std::vector<Point> topPoints(points.begin(), points.begin() + 2);
        std::vector<Point> bottomPoints(points.begin() + 2, points.end());
        
        // Step 2: Sort top group by x-coordinate to get topLeft and topRight
        std::sort(topPoints.begin(), topPoints.end(), [](const Point& a, const Point& b) {
            return a.x < b.x;
        });
        Point topLeft = topPoints[0];
        Point topRight = topPoints[1];
        
        // Step 3: Sort bottom group by x-coordinate to get bottomLeft and bottomRight
        std::sort(bottomPoints.begin(), bottomPoints.end(), [](const Point& a, const Point& b) {
            return a.x < b.x;
        });
        Point bottomLeft = bottomPoints[0];
        Point bottomRight = bottomPoints[1];
        
        // Verify the results
        NSLog(@"Extracted Corner Points:");
        NSLog(@"topLeft: (%d, %d)", topLeft.x, topLeft.y);
        NSLog(@"topRight: (%d, %d)", topRight.x, topRight.y);
        NSLog(@"bottomLeft: (%d, %d)", bottomLeft.x, bottomLeft.y);
        NSLog(@"bottomRight: (%d, %d)", bottomRight.x, bottomRight.y);
    }
    return 0;
}

Notes: #

• This approach assumes your input always has exactly 4 points (which matches your scenario). If you need to handle variable point counts (e.g., a contour with more points), you'd need to add checks like filtering outliers or using OpenCV's boundingRect/minAreaRect functions—though those would return axis-aligned or rotated rectangles, which might not match your exact desired points.
• The lambda functions in std::sort work seamlessly in Objective-C++ since it supports mixing C++ and Objective-C syntax.
• Make sure your project is set up for Objective-C++ (rename your .m file to .mm if needed) and has the OpenCV framework linked.

内容的提问来源于stack exchange，提问作者user9507127