点云的统计异常值删除算法

点云pcl库中的异常值删除，高斯分布异常值，使用接口StatisticalOutlierRemoval。

使用效果如下：

 使用均值10，标准差1

 取反

实现klib平台的插件算子如下:

struct FilterStatisticalOutlierRemoval
{
	//FilterIndices
	bool extract_removed_indices = false;
	bool negative = false;
	bool keep_organized = false;
	float userFilterValue = std::numeric_limits<float>::quiet_NaN();
	//StatisticalOutlierRemoval
	int meanK = 10;
	double stddevMulThresh = 1.0;

	//
	template<class PT>
	KArbit operator()(const pcl::shared_ptr<PointCloud<PT> >& cloud)
	{
		pcl::shared_ptr<PointCloud<PT> > res(new PointCloud<PT>);
		StatisticalOutlierRemoval<PT> f(extract_removed_indices);
		f.setNegative(negative);
		f.setKeepOrganized(keep_organized);
		f.setUserFilterValue(userFilterValue);
		f.setMeanK(meanK);
		f.setStddevMulThresh(stddevMulThresh);
		f.setInputCloud(cloud);
		f.filter(*res);
		return res;
	}

};


struct KPclFilterRemovalStatisticalOutlier : KPlugBase
{
	FilterStatisticalOutlierRemoval p;
	
	KPclFilterRemovalStatisticalOutlier()
	{
		addProp("extract_removed_indices").setType(PropTypeBool).setValueAddr(&p.extract_removed_indices);
		addProp("negative").setType(PropTypeBool).setValueAddr(&p.negative);
		addProp("keep_organized").setType(PropTypeBool).setValueAddr(&p.keep_organized);
		addProp("userFilterValue").setType(PropTypeFloat).setValueAddr(&p.userFilterValue);
		addProp("meanK").setType(PropTypeInt32).setValueAddr(&p.meanK);
		addProp("stddevMulThresh").setType(PropTypeDouble).setValueAddr(&p.stddevMulThresh);
	}

	int __exec(const KArbit& src, KArbit& dst) override
	{
		dst = cld_op(src, p);
		return true;
	}
};