cloud.opencode.base.image.transform.PerspectiveTransformOp

public final class PerspectiveTransformOp extends Object

Perspective (Homography) Transform Operations for Images 图像透视（单应性）变换操作

Applies perspective (projective) transformations to images using a homography matrix computed from 4 point correspondences via the Direct Linear Transform (DLT) algorithm. Unlike affine transforms, perspective transforms can model foreshortening and vanishing-point effects.

使用通过直接线性变换（DLT）算法从 4 个点对应关系计算的单应性矩阵，对图像进行透视（射影）变换。与仿射变换不同，透视变换可以模拟透视缩短和消失点效果。

Features | 主要功能:

4-point perspective transform via DLT homography - 通过 DLT 单应性的 4 点透视变换
Bilinear interpolation for sub-pixel accuracy - 双线性插值实现亚像素精度
Document rectification (trapezoid to rectangle) - 文档矫正（梯形转矩形）

Usage Examples | 使用示例:

// Rectify a trapezoid region to a rectangle
double[] src = {50,50, 350,30, 380,280, 20,300}; // trapezoid corners
double[] dst = {0,0, 400,0, 400,300, 0,300};     // rectangle corners
BufferedImage rectified = PerspectiveTransformOp.apply(image, src, dst, 400, 300);

Security | 安全性:

Thread-safe: Yes (stateless) - 线程安全: 是（无状态）
Null-safe: No (validates input, throws on null) - 空值安全: 否（验证输入，null 时抛出异常）

Since:

JDK 25, opencode-base-image V2.0.0

Author:

Leon Soo www.LeonSoo.com

See Also:

Method Summary

Modifier and Type

Method

Description

static BufferedImage

apply(BufferedImage image, double[] srcPoints, double[] dstPoints, int outputWidth, int outputHeight)

Apply a perspective (homography) transform from 4 source points to 4 destination points.

Methods inherited from class Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Method Details
- apply
  
  public static BufferedImage apply(BufferedImage image, double[] srcPoints, double[] dstPoints, int outputWidth, int outputHeight)
  
  Apply a perspective (homography) transform from 4 source points to 4 destination points. 应用从 4 个源点到 4 个目标点的透视（单应性）变换。
  Points are specified as flat arrays: [x0, y0, x1, y1, x2, y2, x3, y3]. The homography is computed using the Direct Linear Transform (DLT) algorithm, then the inverse mapping is applied with bilinear interpolation.
  
  Parameters:
  
  image - the source image | 源图像
  
  srcPoints - source points [x0,y0, x1,y1, x2,y2, x3,y3] | 源点坐标
  
  dstPoints - destination points [x0,y0, x1,y1, x2,y2, x3,y3] | 目标点坐标
  
  outputWidth - the output image width | 输出图像宽度
  
  outputHeight - the output image height | 输出图像高度
  
  Returns:
  
  the transformed image | 变换后的图像
  
  Throws:
  
  ImageOperationException - if any input is null, points arrays have wrong length, output dimensions are not positive, or the homography is degenerate 当输入为 null、点数组长度不对、输出尺寸非正数或单应性退化时抛出

Class PerspectiveTransformOp

Method Summary

Methods inherited from class Object

Method Details

apply