试图改变一个像素的 HSV 颜色，结果得到了意想不到的颜色。为什么？

Question

以下两个图像中的第一个，rgbaMat.bmp，是RGBA 格式的5px*5px .bmp 图像。这张图片是使用 Highgui.imread 从 sdcard 中读取的，然后使用 Imgproc.cvtColor(rgbaMat, hsvMat, Imgproc.COLOR_BGR2HSV); 转换为 HSV 得到 hsvMat.bmp（第二张图片）：

以下是在 MS Paint 中放大到 100px*100px 的相同图像，以供查看。

然后执行下面的代码后，我得到的changedMat.bmp的图片如下，后面是放大版。

问题是我在编写代码时所期望的是交替像素应该有 white (H=0, S=0, V=255) 和 black (``)色调分别，可以检查in the HSV section on this website。但我得到的是红色和黑色。

问题是为什么？我哪里错了？

public void doProcessing(View view) {
    Mat rgbaMat = Highgui.imread("/mnt/sdcard/DCIM/rgbaMat.bmp");

    Mat hsvMat = new Mat();
    Imgproc.cvtColor(rgbaMat, hsvMat, Imgproc.COLOR_BGR2HSV);
    Highgui.imwrite("/mnt/sdcard/DCIM/hsvMat.bmp", hsvMat);//check

    int counter=1;
    for (int firstCoordinate = 0; firstCoordinate<hsvMat.rows(); firstCoordinate++) {
        for (int secondCoordinate = 0; secondCoordinate<hsvMat.cols(); secondCoordinate++) {
            Log.i(TAG, "HAPPY " + Arrays.toString(hsvMat.get(firstCoordinate, secondCoordinate)));//check

            double[] pixelChannels = hsvMat.get(firstCoordinate, secondCoordinate);

            if (counter%2 != 0) {
                pixelChannels[0]=0;
                pixelChannels[1]=0;
                pixelChannels[2]=255;
            } else {
                pixelChannels[0]=0;
                pixelChannels[1]=0;
                pixelChannels[2]=0;
            }

            hsvMat.put(firstCoordinate, secondCoordinate, pixelChannels);

            counter++;
            Log.i(TAG, "HAPPY PAPPY" + Arrays.toString(hsvMat.get(firstCoordinate, secondCoordinate)));//check
        }
    }
    Highgui.imwrite("/mnt/sdcard/DCIM/matChanged.bmp", hsvMat);//check
}

.................................................. ..................................................... ..................................................... ......

---

编辑：

Line#11 和 Line#28 上的 Log.i 语句打印出以下输出，这很好。

01-12 13:46:44.445: I/MainActivity(29690): HAPPY [0.0, 0.0, 164.0]
01-12 13:46:44.445: I/MainActivity(29690): HAPPY PAPPY[0.0, 0.0, 255.0]
01-12 13:46:44.445: I/MainActivity(29690): HAPPY [29.0, 252.0, 255.0]
01-12 13:46:44.445: I/MainActivity(29690): HAPPY PAPPY[0.0, 0.0, 0.0]
01-12 13:46:44.445: I/MainActivity(29690): HAPPY [0.0, 0.0, 164.0]
01-12 13:46:44.446: I/MainActivity(29690): HAPPY PAPPY[0.0, 0.0, 255.0]
01-12 13:46:44.446: I/MainActivity(29690): HAPPY [29.0, 252.0, 255.0]
01-12 13:46:44.446: I/MainActivity(29690): HAPPY PAPPY[0.0, 0.0, 0.0]
01-12 13:46:44.446: I/MainActivity(29690): HAPPY [0.0, 0.0, 164.0]
01-12 13:46:44.446: I/MainActivity(29690): HAPPY PAPPY[0.0, 0.0, 255.0]
01-12 13:46:44.446: I/MainActivity(29690): HAPPY [29.0, 252.0, 255.0]
01-12 13:46:44.447: I/MainActivity(29690): HAPPY PAPPY[0.0, 0.0, 0.0]
01-12 13:46:44.447: I/MainActivity(29690): HAPPY [0.0, 0.0, 164.0]
01-12 13:46:44.447: I/MainActivity(29690): HAPPY PAPPY[0.0, 0.0, 255.0]
01-12 13:46:44.447: I/MainActivity(29690): HAPPY [29.0, 252.0, 255.0]
01-12 13:46:44.447: I/MainActivity(29690): HAPPY PAPPY[0.0, 0.0, 0.0]
01-12 13:46:44.448: I/MainActivity(29690): HAPPY [0.0, 0.0, 164.0]
01-12 13:46:44.448: I/MainActivity(29690): HAPPY PAPPY[0.0, 0.0, 255.0]
01-12 13:46:44.448: I/MainActivity(29690): HAPPY [29.0, 252.0, 255.0]
01-12 13:46:44.448: I/MainActivity(29690): HAPPY PAPPY[0.0, 0.0, 0.0]
01-12 13:46:44.448: I/MainActivity(29690): HAPPY [0.0, 0.0, 164.0]
01-12 13:46:44.448: I/MainActivity(29690): HAPPY PAPPY[0.0, 0.0, 255.0]
01-12 13:46:44.448: I/MainActivity(29690): HAPPY [29.0, 252.0, 255.0]
01-12 13:46:44.448: I/MainActivity(29690): HAPPY PAPPY[0.0, 0.0, 0.0]
01-12 13:46:44.448: I/MainActivity(29690): HAPPY [0.0, 0.0, 164.0]
01-12 13:46:44.448: I/MainActivity(29690): HAPPY PAPPY[0.0, 0.0, 255.0]
01-12 13:46:44.449: I/MainActivity(29690): HAPPY [29.0, 252.0, 255.0]
01-12 13:46:44.449: I/MainActivity(29690): HAPPY PAPPY[0.0, 0.0, 0.0]
01-12 13:46:44.449: I/MainActivity(29690): HAPPY [0.0, 0.0, 164.0]
01-12 13:46:44.450: I/MainActivity(29690): HAPPY PAPPY[0.0, 0.0, 255.0]
01-12 13:46:44.450: I/MainActivity(29690): HAPPY [29.0, 252.0, 255.0]
01-12 13:46:44.450: I/MainActivity(29690): HAPPY PAPPY[0.0, 0.0, 0.0]
01-12 13:46:44.450: I/MainActivity(29690): HAPPY [0.0, 0.0, 164.0]
01-12 13:46:44.450: I/MainActivity(29690): HAPPY PAPPY[0.0, 0.0, 255.0]
01-12 13:46:44.450: I/MainActivity(29690): HAPPY [29.0, 252.0, 255.0]
01-12 13:46:44.451: I/MainActivity(29690): HAPPY PAPPY[0.0, 0.0, 0.0]
01-12 13:46:44.451: I/MainActivity(29690): HAPPY [0.0, 0.0, 164.0]
01-12 13:46:44.451: I/MainActivity(29690): HAPPY PAPPY[0.0, 0.0, 255.0]
01-12 13:46:44.451: I/MainActivity(29690): HAPPY [29.0, 252.0, 255.0]
01-12 13:46:44.451: I/MainActivity(29690): HAPPY PAPPY[0.0, 0.0, 0.0]
01-12 13:46:44.451: I/MainActivity(29690): HAPPY [0.0, 0.0, 164.0]
01-12 13:46:44.452: I/MainActivity(29690): HAPPY PAPPY[0.0, 0.0, 255.0]
01-12 13:46:44.452: I/MainActivity(29690): HAPPY [29.0, 252.0, 255.0]
01-12 13:46:44.452: I/MainActivity(29690): HAPPY PAPPY[0.0, 0.0, 0.0]
01-12 13:46:44.452: I/MainActivity(29690): HAPPY [0.0, 0.0, 164.0]
01-12 13:46:44.452: I/MainActivity(29690): HAPPY PAPPY[0.0, 0.0, 255.0]
01-12 13:46:44.452: I/MainActivity(29690): HAPPY [29.0, 252.0, 255.0]
01-12 13:46:44.452: I/MainActivity(29690): HAPPY PAPPY[0.0, 0.0, 0.0]
01-12 13:46:44.452: I/MainActivity(29690): HAPPY [0.0, 0.0, 164.0]
01-12 13:46:44.452: I/MainActivity(29690): HAPPY PAPPY[0.0, 0.0, 255.0]

Answer 1

为了检查，我将像素的 HSV 转换为 Android Color 并使用

记录 Android Color 的值

//CONVERTING HSV COLOR TO ANDROID COLOR
//Step 1: Convert double[] to float[]. Step 2: Scale the float[]. STep 3: Use Android's HSVToColor method
float[] channelsFloatArrayScaled = new float[3];
for (int i = 0; i < hsvMat.get(firstCoordinate, secondCoordinate).length; i++) {
    if (i == 0) {
        channelsFloatArrayScaled[i] = ((float) hsvMat.get(firstCoordinate, secondCoordinate)[i]) * 2;// TODO Wrap an ArrayIndexOutOfBoundsException wrapper
    } else if (i == 1 || i == 2) {
        channelsFloatArrayScaled[i] = ((float) hsvMat.get(firstCoordinate, secondCoordinate)[i]) / 255;// TODO Wrap an ArrayIndexOutOfBoundsException wrapper
    }
}
int androidColor = Color.HSVToColor(channelsFloatArrayScaled);
Log.i(TAG, "HAPPY PAPPY FAPPY" + androidColor);//check

结果是 -1 和 -1777216 交替像素，代表白色和黑色。

所以一切都很好，除了我不知道 OpenCV 在将图像写入 sdcard 时在做什么（最后一个 imwrite）。