目标检测与识别

1. 颜色检测

采集大量敌方机器人的图片数据，并进行训练，得到对方机器人的颜色区间，

并以此为阈值对整幅图像进行颜色检测，找到疑似敌方机器人的区域，量化

成二值图。

2. 滤除噪声点

对得到的二值图像进行开运算处理，滤除颜色检测结果中的噪声点。

3. 连通区域检测

对图像中的疑似区域进行连通区域检测，计算出每个疑似区域的外部轮廓，

用矩形近似表示。

4. 连通区域合并

根据连通区域之间的距离和颜色相似性，将距离接近且相似性高的连通区域

进行合并。

5. 形状和大小过滤

对大量敌方机器人图片进行训练，得到对方机器人的形状信息（例如长宽比）

和大小信息（面积），并以此为依据将不符合的区域过滤掉。

经过以上五步的处理，可以初步得到敌方机器人的位置、大小和形状信息。-Target detection and recognition

1 color detection

Collect a lot of image data of the enemy robot, and training, get the color range of the other robot,

And on the threshold of the color detection on the image, to find the suspected enemy robot area quantization

Cheng two values.

2 filter out noise points

The two valued image is processed to remove the noise points in the color detection results.

3 connected region detection

In the image of the suspected region of the connected region detection, calculation of each suspected region of the external contour,

Representation of a rectangular approximation.

4 connected region merging

According to the distance between the connected regions and the color similarity, the distance is close to and the similarity is high.

To merge.

5 shape and size filtering

For a large number of images of the enemy robot training, get the other side of the robot shape information (such as the length and width ratio)

And the size of information (ar